Preparing Today for Linux Tomorrow
Linux Questions and Answers
A Linux White Paper
Preface
For someone new (and even not so new) to Linux® , there are hundreds of questions,
problems and concerns that arise during the learning process, especially for someone coming
from the world of Microsoft® Windows® . Other Linux white papers deal with individual topics,
often in lengthy fashion. This paper addresses a number of miscellaneous questions, both
frequently- and rarely-asked, grouped by topic. The focus will be on helping Windows users
make the transition to Linux, but those converting from other operating systems should find
useful information here as well.
Note: Because of the differences between Linux “distributions” from various vendors, some of
the information below may apply to one distribution but not another. This will be pointed out
where known; however, with all the available distributions it is impossible to identify all such
situations here. If the problem is that a command or program described doesn’t exist in a
given distribution, it is generally possible to download a copy of that program from a Web site,
if needed. On the other hand, there may be a functionally similar utility already provided with
that distribution. To find out, consult the user’s manual or contact the distribution vendor. If all
else fails, there is a list of Web sites in the Miscellaneous section, below, that may be of
assistance.
For general terminology questions, please refer to the white paper entitled A Brief Linux
Glossary for Windows Users, available from the same sources as this paper.
Special Note: If you are reading this document online with Adobe Acrobat Reader, simply
click on the Web addresses highlighted in blue to go to those Web sites via your Web
browser.
Preparing Today for Linux Tomorrow
2
Contents
Preface.................................................................................................................................................... 1
Installation.............................................................................................................................................. 5
Q. During installation Linux creates a swap space partition. Why do I need this and how is it
different from a Windows swap file? ................................................................................................. 5
Q. During installation, I have a choice of creating swap space in a Primary or an Extended
partition. Which should I choose?..................................................................................................... 5
Q. How large a swap partition should I create? ................................................................................ 5
Q. How can I speed up performance by using multiple swap partitions?......................................... 6
Q. How can I create multiple swap partitions on one or more drives? ............................................. 7
Q. How do I create a swap file in an existing Linux data partition? .................................................. 9
Q. I already have a large swap file in my Windows partition. Is there a way for Linux to use that
swap space instead of creating another file?.................................................................................. 10
Q. If I install Linux on an IBM® ThinkPad® with a TrackPoint® II or III pointer, what kind of mouse
should I select? ............................................................................................................................... 11
Migration from Windows..................................................................................................................... 11
Q. How do I get Netscape for Linux to recognize my Netscape for Windows bookmark file?........ 11
Q. Are there any Linux programs that will read and write Microsoft Office files? ........................... 12
Configuration & Setup ........................................................................................................................ 12
Q. How do I get my “winmodem” to work with Linux?..................................................................... 12
Q. How can I tell how much memory Linux is using? ..................................................................... 13
Q. If Linux is not using all the memory I have installed, how do I make it use the rest? ................ 13
Q. I have a printer attached to the parallel port. What is this port called in Linux? ........................ 13
Q. I have a device attached to a serial port. What is this port called in Linux? .............................. 13
Q. What is my floppy drive called in Linux? .................................................................................... 14
Q. Wait a minute! Are you saying that there are no drive letters in Linux?..................................... 14
Q. What about disk drive partitions? Don’t they have drive letters either?..................................... 15
Q. Wow, that’s confusing! Isn’t there a simpler way to access drives? .......................................... 15
Q. Is there an easy way to see what devices, such as drives, have been mounted? .................... 15
Q. Is there any way to change the default language used by the operating system and
applications? ................................................................................................................................... 15
General Usage ..................................................................................................................................... 16
Q. How do I format a floppy (diskette)? .......................................................................................... 16
Q. Is there online documentation for most programs? ................................................................... 17
Q. How do I kill a program that has locked up? .............................................................................. 17
Q. Is there a way to increase the priority of a program to make it run faster?................................ 18
Q. Is there a spelling checker that I can use outside of a word processor? ................................... 18
GUI/Desktop ......................................................................................................................................... 19
Preparing Today for Linux Tomorrow
3
Q. My copy of Linux came with KDE and Gnome. Why do I need both? ....................................... 19
Q. I have both KDE and Gnome installed. How do I switch between them?.................................. 19
Q. Can I add more desktop themes to use with Gnome and KDE? ............................................... 19
Q. How can I add programs to the Panel?...................................................................................... 20
Q. What is a virtual desktop? .......................................................................................................... 20
Q. You said I could have eight virtual desktops, but I see only four. Where are the rest?............. 20
Q. Is it possible to share a program across multiple virtual desktops?........................................... 21
Q. How can I change my desktop background color? .................................................................... 21
Q. Is there a faster way to copy/cut & paste text than to use the pull-down menu?....................... 21
Command Shell ................................................................................................................................... 21
Q. If I want to issue a command, how do I open a command prompt? .......................................... 21
Q. What are the Linux equivalents to DOS/Windows commands? ................................................ 22
Q. How do I get help for command parameters?............................................................................ 23
Q. How do I repeat a command? .................................................................................................... 23
Q. What can I do if I don’t remember the name of a command I need?......................................... 24
Q. I just used a command and got completely different results from the last time I used it. Why? 25
Q. Long pathnames are a pain to type in. Is there any sort of shortcut I can use? ........................ 25
Q. Is there a way to use one command to start more than one program at a time? ...................... 25
Q. Is there a way to stack commands and have them execute concurrently in other command
sessions? ........................................................................................................................................ 26
Logging In/Logging Out/Shutting Down ........................................................................................... 27
Q. I can’t seem to log in, even though I’m using the correct password. What’s wrong? ................ 27
Q. Why do I need to use Shutdown? Can’t I just turn the power off?............................................. 27
Q. I can find a Logout option, but where is Shutdown? .................................................................. 27
Q. Is there a faster way to “warm-boot” the system than using Shutdown?................................... 28
Q. Do I have to use Shutdown if I am just turning over the system to another user? .................... 28
Q. I clicked on Logout and now the session is locked up. How do I log out?................................. 28
Q. How can I enable Linux to automatically restart applications that are running when I use
Shutdown or Logout?...................................................................................................................... 28
Files and Directories ........................................................................................................................... 28
Q. What is the root directory? ......................................................................................................... 28
Q. Can you explain more how the directory structure of Linux works? .......................................... 29
Q. There are a number of directories I don’t recognize in the root directory. What are they?........ 30
Q. Is there a command line shortcut to the login directory? ........................................................... 30
Q. Some filenames are preceded by a dot. What does this mean? .............................................. 30
Q. When I look at a list of files and directories, the names are followed by something like -rw-rw-r--
or lrwxrwxrwx. What does this mean? ........................................................................................... 31
Q. The ls command doesn’t show everything in the directory. What’s wrong? ............................ 31
Preparing Today for Linux Tomorrow
4
Q. Can I use wildcard searches with the ls command, as I can with Dir in DOS/Windows? ........ 31
Q. I want to do a wildcard search on files with ? or * in the names. How do I do that?................. 32
Q. How do I create or rename a file with special characters in the name? .................................... 32
Q. How long can a filename be?..................................................................................................... 32
Q. Is there a way to find files from a command shell?.................................................................... 33
Q. Is there a way to browse through a text file from the command line?........................................ 33
Q. Is there a command to search text files for a specific character string?.................................... 34
Q. What are “symbolic” and hard links, and how can I identify them?............................................ 35
Q. How do I create links (symbolic or hard)?.................................................................................. 36
Q. Can you summarize the differences between symbolic and hard links and differentiate between
using links and simply duplicating a file? ........................................................................................ 37
Q. I recognize a number of file types, such as .JPG, .GIF, .WAV, .TXT, .HTM and .ZIP from
Windows, but I see many other file types I don’t know. What are they for? ................................... 37
Q. Should I back up my entire system? .......................................................................................... 38
Q. How often should I back up my files? ........................................................................................ 38
Q. What backup software should I use? ......................................................................................... 39
Root/User Accounts, Groups and Permissions ............................................................................... 40
Q. Sometimes instructions say that I must “be root.” What does this mean?................................. 40
Q. What is the root account and how does it differ from a user account?...................................... 40
Q. Why would I want to add other user accounts to my system?................................................... 41
Q. How do I add users? .................................................................................................................. 41
Q. How do I delete users? .............................................................................................................. 42
Q. Is there any way to change a user ID? ...................................................................................... 42
Q. Is there a way to disable a user account?.................................................................................. 43
Q. How do I add or change personal information in an account?................................................... 43
Q. How do I add groups? ................................................................................................................ 43
Q. How do I add/delete users in a group? ...................................................................................... 43
Q. How do I change a password?................................................................................................... 43
Q. What is a shadow password? .................................................................................................... 44
Q. How do I display the permissions for a file or directory? ........................................................... 44
Q. How do I change permissions?.................................................................................................. 44
Communications ................................................................................................................................. 45
Q. How do I set up my system for dial-up Internet access? ........................................................... 45
Miscellaneous ...................................................................................................................................... 46
Q. What is the difference between Linux and UNIX? ..................................................................... 46
Q. Why does Linux seem so haphazard?....................................................................................... 47
Q. Where can I download or buy Linux software? .......................................................................... 47
Q. Where can I go for assistance on Linux?................................................................................... 48
Preparing Today for Linux Tomorrow
5
Installation
These questions all deal with some aspect of installing the operating system.
Q. During installation Linux creates a swap space partition. Why do I
need this and how is it different from a Windows swap file?
Like Windows, Linux uses a certain amount of space for holding programs temporarily, when
there is not enough available RAM (random access memory) to hold all the programs that are
running concurrently. Generally, the least recently used program (or part of a program) is
copied from memory to a file on your hard drive until it is needed again, at which time the
current least recently used program is swapped out in its place and the first program is loaded
back into memory. (This is an over-simplified explanation; there is much more to it, but this will
do for this question.) This file is called a swap file in Windows or OS/2 and “swap space” in
Linux, but in either case it is a form of data file that is read from and written to off and on as
long as your system is running.
Windows puts the swap file (a hidden system file with different names for different versions of
Windows) in the bootable data partition by default. OS/2 does the same, but by changing the
CONFIG.SYS file a user can put the swap file in any directory in any partition on any drive
they like. Linux, by default, requires a special swap partition in which to store the swap file.
(Actually, Linux does allow swap files to be put in data partitions, with caveats—see below for
Q. During installation, I have a choice of creating swap space in a
Primary or an Extended partition. Which should I choose?
Either will work. There is no technical advantage to doing it one way or the other, but there are
only a limited number of primary partitions that can be created, so if you plan on installing
more than one operating system you might want to put the swap space in an Extended
partition and save a Primary partition for other uses.
Note: If you have Microsoft Windows installed, it may not boot if you have more than one
primary partition installed on the bootable disk drive (what Windows sees as C:). In this case,
Linux would have to be installed in an Extended partition, or in a Primary partition on a second
hard disk drive. For more information, refer to the lilo (LInux LOader) documentation on a
Linux system, using the command: man lilo (no, man is not a politically incorrect
command—it is merely short for manual). If you do not yet have a Linux system set up where
you can read the lilo documentation there are a couple of other options available: 1) If you are
installing from a CD set, look for a directory or separate CD (as in Red Hat 6.2) containing
documentation, possibly in HTML format. 2) Visit Linux Web sites, such as
http://linux.ctyme.com and http://www.linuxdoc.org, for online documentation.
Q. How large a swap partition should I create?
Although it can be smaller, for best results the partition size should be at least equal to the
amount of memory installed in the system—preferably twice the amount of physical memory.
In other words, if you have 64MB of RAM installed use double that amount for your swap
space partition (64MB * 2 = 128MB).
If you need more than 128MB of swap space, but you are using an older distribution that does
not support a swap file larger than 128MB, there are ways around the 128MB swap file
limitation: 1) Linux allows more than one swap partition on a hard drive, and 2) Linux also
allows swap files to be put in data partitions (i.e., in the same partitions as your programs and
data files).
Preparing Today for Linux Tomorrow
6
If you create a swap partition and then later decide that it is not large enough for your needs,
you can delete the original one and replace it with a larger swap partition if you have the
contiguous space to do so—in other words, if there is no other partition immediately following
the original swap partition. If there is insufficient space to create the larger swap partition, you
can still create a second partition elsewhere on that drive. They do not have to be back-toback.
(Or you can use a disk partitioning program, such as Partition Magic, to reallocate
partition space on the fly between existing partitions, without losing data.) Linux supports up to
eight swap partitions.
Alternatively, if you have multiple physical disk drives, splitting your swap space between the
disk drives may improve system performance, because both drives can be used concurrently
for swapping. (This performance increase is more apparent on a system with SCSI drives,
because SCSI devices run in parallel—at the same time—while standard IDE devices run in
serial—they alternate. Even so, on a system with multiple IDE drives, it can be faster to switch
back and forth between physical drives (an electronic operation) than to shuttle the read/write
heads of one drive all over the disk, dividing their time between partitions on the same drive.)
Also, drives added to the system later may be faster than the original drive that came with the
system. Using these newer drives will give you faster disk read/writes when swapping.
Dividing the swap space across multiple drives may also increase system reliability. Like any
device, hard drives have a finite life expectancy, so spreading the workload between multiple
drives means that the primary drive doesn’t have to work as hard, which in turn may (no
guarantees, though) extend its useful life. (For instructions on how to do this, refer to the
question, How can I create multiple a swap partitions on one or more drives?, below.)
Q. How can I speed up performance by using multiple swap partitions?
If your system (especially a server) has multiple drives set up as a RAID 0 array (data
striping), it would automatically spread the data across all the drives in parallel, greatly
increasing the speed of disk writes versus a non-RAID setup, where the swap partition is on
only one drive. Of course, not everyone can afford a RAID setup; fortunately, there is an
alternative that offers many of the same performance benefits in a multidisk environment, at
least as far as swap space performance is concerned.
As it happens, there is a way to “parallelize” swap file read/writes through the use of the
priority setting in the /etc/fstab file. (Note: This works best with multiple SCSI drives and/or
controllers. Two IDE drives on the same controller will see little or no performance boost from
this technique, but multiple IDE drives on multiple controllers may benefit somewhat.)
Use an appropriate tool (such as kfstab [available from kfstab.purespace.de/kfstab], which
knows the exact layout of the file that fstab requires) to open the file /etc/fstab. This file lists
the partitions used by Linux, including swap space. (Note: It is possible to use a simple text
editor such as vi, but if you do not get the columns lined up just right, or if you use tabs instead
of spaces or vice versa, you can corrupt your system; so if you elect to go this route, rather
than using a tool specifically designed for editing fstab, such as kfstab, be sure to make a
backup copy of fstab first, so that you can recover if your system crashes.)
When you open fstab, if you have multiple SCSI drives with a swap partition on each, you
might see something like this (without the headings):
<partition <mount <partition< mount <backup
name> point> type> options> dump> <pass>
/dev/sda2 none swap sw 0 0
/dev/sdb2 none swap sw 0 0
/dev/sdc2 none swap sw 0 0
/dev/sdd2 none swap sw 0 0
Preparing Today for Linux Tomorrow
7
Under normal circumstances, Linux would use the swap partition /dev/sda2 first, then
/dev/sdb2, and so on, one at a time, until it had enough space to meet its current swapping
needs.
On the other hand, if you change the settings to:
/dev/sda2 none swap sw,pri=3 0 0
/dev/sdb2 none swap sw,pri=3 0 0
/dev/sdc2 none swap sw,pri=3 0 0
/dev/sdd2 none swap sw,pri=1 0 0
you will be assigning the first three partitions the same priority level—a higher one than the
fourth partition. (Swap partitions are used in order from highest priority to lowest—where a
priority level of 32767 is the highest and 0 is the lowest.) This has the effect of forcing Linux to
write to the first three partitions in parallel, greatly increasing the read/write throughput of the
swap space. The fourth partition would be used only if the first three fill up.
In addition to parallelizing the swapping, this technique also allows you to prioritize access to
the fastest drives first. Therefore, you can set up an emergency swap partition on an old, slow
drive, using low priority, so that it will be used only when all high-speed swap space is
exhausted.
Q. How can I create multiple swap partitions on one or more drives?
To create multiple swap partitions:
1. First, use the: su – command to enter superuser mode. (If you don’t do this you will be
unable to use the fdisk command.)
2. Enter the command: fdisk /dev/hda if your bootable hard drive is IDE, or use: fdisk
/dev/sda if your bootable hard drive is SCSI. (If you are not sure which type of disk
drives your system uses and you guess wrong, the worst that can happen is that Linux will
issue an error message saying “Unable to open xda.” You can’t hurt anything.) This is a
text-mode program, so it is not pretty. Where it says, “The number of cylinders for this disk
is set to xxxx” note the number displayed for xxxx (for example, 1024).
3. At the prompt “Command (m for help).”
a. Type: p (and press Enter) to see the current partitions. Check the number shown
under the End column for the last cylinder listed. If it is less than the number of
cylinders from Step 2, above, there is room for another partition. (It is only necessary
to create another partition on the same drive if your Linux distribution doesn’t allow
swap partitions larger than 128MB and you want additional swap partitions.) The
difference between the two numbers determines how much space is available for use.
b. (If there is already an existing swap partition, but it is not large enough, use the: d
(delete) command to remove it; then create a new, larger one.)
c. Enter: n to create a new partition.
d. If you are creating a Primary partition:
1) Enter: p to create the partition. When prompted for the partition number, use the
next higher number than the highest existing partition number. (In other words, if
the last device description on the left shows: /dev/hda1, use 2 as the new
partition number.)
Preparing Today for Linux Tomorrow
8
2) Type in the number suggested for the first cylinder (which should be one higher
than the ending cylinder for the preceding partition, or press Enter to take the
default.
3) Enter the partition size in megabytes, such as: +128M. (Be sure to include both
the + and the M or you will get a “Value out of range” error.)
4) Enter: p again to verify that the new partition was created. It should show an Id
type of 83 and a System type of Linux.
5) Enter: t to change the partition type, then enter the partition number you used in
Step 1), above. Finally, when prompted for the “Hex code,” enter the swap
partition type code: 82. If everything worked correctly, you should see a message
like “Changed system type of partition [x] to 82 (Linux swap).” (where [x] is the
partition number from Step 1).
e. If you are creating an Extended partition:
1) Enter: e to create the partition. When prompted for the partition number, use the
next higher number than the highest existing partition number. (In other words, if
the last Device description on the left shows: /dev/hda1, use 2 as the new
partition number.)
2) Type in the number suggested for the first cylinder (which should be one higher
than the ending cylinder for the preceding partition, or press Enter to take the
default.
3) Press Enter to reserve all the available space.
4) Enter: p again to verify that the new Extended partition was created. It should
show an Id type of 5 and a System type of Extended.
5) Enter: t to change the partition type, then enter the partition number you used in
Step 1), above. Then, when prompted for the “Hex code”, enter the Linux
Extended partition type code: 85. If everything worked correctly, you should see a
message like “Changed system type of partition [x] to 85 (Linux extended).”
(where [x] is the partition number from Step 2). Note: This step is not strictly
necessary. You can create Linux swap partitions in a “regular” extended partition,
but changing the partition type to Linux Extended takes only a few seconds and it
clarifies the partition structure for future fdisk users on this system.
6) Enter: n to create a new partition. When prompted for the first cylinder number,
press Enter. Enter the desired swap partition size in megabytes, such as: +128M.
Enter: p again to verify that a type 83 partition was created with the same starting
block as the Linux extended partition you created in Step 1).
7) Enter: t to change the partition type, then enter the number for the Linux
extended partition you just created. When prompted for the “Hex code,” enter the
swap partition type code: 82. If everything worked correctly, you should see the
message “Changed system type of partition [x] to 82 (Linux swap).
8) If you did not use up all of the available partition space in Step 6), create one or
more additional partitions to hold data, but this time do not change the partition
default type of 83.
f. To write your changes to the hard disk and exit the program, type: w and press Enter.
4. When you are returned to the Hard Disk Selection panel, select No further hard disk
changes and press Enter.
5. Repeat steps 2-4 for each additional hard drive that you wish to add a swap partition to,
changing the fdisk command from: fdisk /dev/hda to: fdisk /dev/hdb; or from:
Preparing Today for Linux Tomorrow
9
fdisk /dev/sda to: fdisk /dev/sdb, depending on the type of hard drives you are
using. When you are done, enter: q to exit fdisk. (For more on disk drive naming
conventions in Linux, see question Wait a minute! Are you saying that there are no drive
letters in Linux?)
Note: To see the complete list of partition types available for your Linux distribution, at the
“Command (m for help)” prompt, enter: l (for “list partition types”).
For performance reasons, if you have more than one hard disk drive your first choice should
be to create multiple swap partitions, rather than multiple partitions on the same drive or one
swap partition and a swap file in a data partition. However, if you have a shortage of available
Primary partitions (or there is no more unallocated disk space from which to create new
partitions), you can create swap files in existing data partitions. To do this, see the question
How do I create a swap file in an existing Linux data partition?, below:
Q. How do I create a swap file in an existing Linux data partition?
If for some reason you can’t, or do not wish to, create Linux swap partitions—or if you have no
more available disk space to use for a new swap partition—it is possible to create swap files
within partitions that currently hold programs and/or data files. However, your first choice
should be for one or more Linux swap partitions.
Important Note: There are two disadvantages to locating swap files outside of dedicated
swap partitions: 1) The performance of swap files that are located in data partitions is slower
than that of swap files in swap partitions (due to file system overhead and noncontiguous data
blocks), and 2) It is also possible that if the swap file is ever damaged it could corrupt the
filesystem for that partition, resulting in lost data. For these reasons it is strongly
recommended that you use swap partitions whenever possible, leaving swap files in data
partitions as a last resort. Swap files might be appropriate if you temporarily need some extra
swap space.
You can create as many as eight swap files, if necessary. Linux will apportion the swapped
data between these files as appropriate.
To create a swap file in a data partition, use a command similar to the following:
dd if=/dev/zero of=<swap path> bs=<size> count=<size>
For example:
dd if=/dev/zero of=/Swapdir/Swapfile bs=1024 count=65536
(Note: If /Swapdir does not yet exist, it must be created before running the dd command. For
example: mkdir /Swapdir or mkdir /home/Swapdir. In order to create the directory, you
must be running as the root operator.)
The dd command (sometimes called “data duplicator,” although the original meaning of the dd
acronym is lost in the mists of time) is used to create the swap file. The if= parameter
identifies the input file source for creating the swap file (device “zero”), of= is where you
specify the path to the output (swap) file you want to create, bs= is the block size to use and
count= is where you define how large to make the swap file, in increments of the block size.
In the preceding example, we are creating a 64MB swap file (65,536 * 1024 [1K] equals
65,536K, or 64MB), called Swapfile (use whatever name you prefer), in a directory called
Swapdir.
Next, to initialize (or make ready for use) the swap file, use the following command:
Preparing Today for Linux Tomorrow
10
mkswap /Swapdir/Swapfile 65536
(Use the same path as in the dd command and the file size you specified in the count=
parameter. If you get a “No such file or directory” error, verify that you typed the upper/lower
case correctly and did not omit the leading /.)
After initializing the file, “synchronize” it to ensure that it is properly written to disk:
sync
And, finally, to tell Linux to start using this new swap file:
swapon /Swapdir/Swapfile
This will activate the swap file temporarily (until the next time you log out or reboot). To make
it permanently active, use an appropriate tool (such as kfstab [see the earlier question, How
can I speed up performance by using multiple swap partitions?, for details], which knows the
exact layout of the file that fstab requires) to open the file /etc/fstab.
When you open fstab, you might see something like (without the headings):
<partition <mount <partition< mount <backup
name> point> type> options> dump> <pass>
/dev/hda1 / ext2 defaults 0 1
/dev/hda2 /usr ext2 defaults 0 1
/dev/hda3 none swap sw 0 0
(There may be other disk partitions listed if you have other swap and/or data partitions.) To
activate your new swap file, add a line similar to the following immediately after the last
partition listing. Then save and close the /etc/fstab file.
/Swapdir/Swapfile none swap sw 0 0
Should you later decide that you no longer need the swap file, it is easy enough to remove.
First, use the swapoff command to disable the use of the file:
swapoff /Swapdir/Swapfile
Then, simply delete the file:
rm /Swapdir/Swapfile
Q. I already have a large swap file in my Windows partition. Is there a
way for Linux to use that swap space instead of creating another file?
Yes, although Linux is rather meticulous about what format the data takes in its swap space.
Thus, there is more to it than simply pointing Linux to the path of the Windows swap file.
I have found several descriptions of how to accomplish this (and all of them reportedly work)
but I have not had the need to try any myself. They all appear to be terribly complicated, so
unless you consider yourself well versed in Linux, it might be best to avoid this topic entirely
until someone develops a utility to simplify the process. If you would like to look over one such
procedure, to make up your own mind whether or not to try it, read the document at
http://www.linuxdoc.org/HOWTO/mini/Swap-Space-1.html. It seems rather thorough. (Use the
Preparing Today for Linux Tomorrow
11
left and right arrows to move back and forth through the document. Use the up arrow to go to
the document table of contents.)
(If you already have a Linux swap partition before you decide to try this technique to use your
Windows swap file instead, you should later delete it using fdisk, to free up the disk space
used by that partition. For directions on how to use fdisk to check your partition tables, and to
delete a partition, see the question How large a swap partition should I create?, above.)
Another possible way to share swap space is to have a swap partition shared by both
operating systems. This is not a “no-brainer” either, but if you are interested do an Internet
search for “Linux swap space” and you should find several such documents.
Q. If I install Linux on an IBM® ThinkPad® with a TrackPoint® II or III
pointer, what kind of mouse should I select?
The installation program should correctly recognize the TrackPoint II/III as a PS/2-type mouse.
Most such ThinkPad notebooks have two “mouse” buttons, but some newer systems have a
third button. When installing on a ThinkPad with two buttons, select the Emulate 3 Buttons
option on the Mouse Configuration screen. If your ThinkPad has three buttons do not select
this option, or it will cause problems. To use the “third button” on a two-button ThinkPad (or
any other system with a two-button mouse), press both buttons together (called “chording”).
(These instructions should apply equally to other brands of notebook computers using IBM’s
“pointing stick” technology, identified as a little “pencil eraser”—usually red, green, or black—
nestled between the G, H and B keys; but to be sure, contact the notebook vendor.)
Migration from Windows
If you are in the process of moving from Windows to Linux, or are setting up a Linux system to
coexist with Windows computers and perhaps share data between the two types of systems,
these questions may save you some time and effort.
Q. How do I get Netscape for Linux to recognize my Netscape for
Windows bookmark file?
For best results, try to use the same version of Netscape on both systems, or as close as
possible (e.g., 4.72 and 4.7, but not 3.0) to minimize the chance of inconsistencies in the
bookmark file format.
Use the Linux Kfind, or similar, utility to determine where the default bookmark file is kept. It
should be in a directory with a name like /home/username/.netscape (substitute your
account name for username). If you simply copy the bookmark.htm file from your Windows
PC to a floppy and then to the appropriate Linux directory, when you restart Netscape and
open the bookmarks you will see the default bookmarks, not your customized list. The
problem is in the file name. The Linux version is called bookmarks.html (note the two extra
letters). Therefore, you must rename the bookmark.htm file to bookmarks.html before
Netscape for Linux will recognize it (reverse the process to copy an updated file back to your
Windows system).
First, copy the bookmark.htm file from the Windows system. If you use a floppy, Zip disk or
other removable medium, you can rename the file on the removable disk before copying to the
Linux system. (The file should be stored in a directory called something like \Program
Files\Netscape\Communicator\Users\yourusername, but you can use the Windows Find or
Search function—depending on the Windows product—from the Start menu to locate it.) Once
the file is copied to the removable disk, use the Windows file manager to rename it
bookmarks.html.
Preparing Today for Linux Tomorrow
12
On the Linux computer, KDE graphical interface users can copy the file to the
/home/username/.netscape directory using the K File Manager (just click on the icon on the
taskbar, or Panel, that looks like a house in front of a manila folder). Gnome users can use the
Gnome File Manager (click on the Gnome icon--the vaguely G-shaped footprint--on the Panel,
select Programs, then File Manager). If you prefer to work from a command line interface, use
the following command:
cp /dev/fd0/bookmarks.html /home/yourusername/.netscape
(Note the use of forward slashes and that /dev/fd0 is used in place of a:. Linux and UNIX® do
not use drive letters, unlike DOS, Windows or OS/2.)
Alternatively, you could combine the renaming and copying steps into one, using the following
command:
mv /dev/fd0/bookmark.htm /home/yourusername/.netscape/bookmarks.html
(Note: The diskette drive may already be “mounted” as /dev/floppy or /mnt/floppy in your
Linux distribution. If so, use that mount location instead of /dev/fd0.)
If you transfer the file directly to the Linux system (via network connection, for example) or
prefer to rename the file once it is on the Linux system go to the /home/username/.netscape
directory (cd /home/username/.netscape) and use a file manager or the move (mv)
command to change the name: mv bookmark.htm bookmarks.html. (If you use the mv
command without a target location, it simply renames the files.)
Note: In addition to the bookmarks file, you might also want to copy the following files from the
same Netscape directory on your Windows system, to your Linux system: liprefs.js, prefs.js
and proxy.cfg. They contain the user preferences information (font selections, proxies,
servers and so on). In addition, if you did not have “cookies” disabled on your Windows
system, copy cookies.txt as well (some Web sites store information about your user account
in this file). This will save you from having to manually reenter your preferences on your Linux
system. (You will probably have to overwrite the default files with the same names on your
Linux system.)
If Netscape is still running, close and restart it so it will read the new bookmarks file (as well as
the other files you copied). If it still doesn’t work, verify that you put it in the correct directory
and have spelled the file name accurately—it must be in all lower-case.
Q. Are there any Linux programs that will read and write Microsoft Office
files?
Yes, there are. For a list of several such programs and how compatible they are, see the white
paper entitled What Good is a Linux Client?, available from the same sources as this paper.
Configuration & Setup
Here are some tips for tailoring Linux or configuring devices.
Q. How do I get my “winmodem” to work with Linux?
Winmodem is actually a trademark of 3Com Corporation, but is often used generically to
describe a class of low-cost software-based modems, designed to be inexpensive by
replacing some hardware functions with software. This software typically incorporates low
Preparing Today for Linux Tomorrow
13
level Windows functions that can’t be duplicated in Linux. Most such modems won’t work with
Linux, however, some can be made to be compatible. These sometimes are referred to as
“Linmodems”. Go to http://www.o2.net/~gromitkc/winmodem.html for more information on this
topic. In most cases, the easiest “fix” is simply to disable the software-based modem and
install a traditional modem adapter, or attach en external modem to a serial or USB port. Many
56K modems are quite inexpensive these days. (For information on how to disable the built-in
modem, consult the user’s guide for the computer, or contact the computer vendor’s technical
support organization.)
Q. How can I tell how much memory Linux is using?
From a command shell, use the “concatenate” command: cat /proc/meminfo for memory
usage information. You should see a line starting something like: Mem: 64655360, etc. This is
the total memory Linux thinks it has available to use. If you have more than 64MB of RAM
installed and see a number close to the one above, Linux is not using the remainder. To see
how much memory is currently unused, use the free command.
Q. If Linux is not using all the memory I have installed, how do I make it
use the rest?
To force Linux to recognize and use the full amount of memory you have installed, open a
command shell and login temporarily as the “root” operator, with the su command. (For
instructions on how to do this, see the Root/User Accounts, Groups and Permissions section,
below.) When prompted, type in the root password. At the command prompt, use an editor to
open the file /etc/lilo.conf, such as: pico /etc/lilo.conf. The first line of the file should
start: boot=, followed by other lines for map=, install= and so on. Position the cursor in
front of the “b” in boot=, then insert the following: append=“mem=128M” and press the Enter
key to move the boot= line down. Finally, close the editor and exit the file (Ctrl-X, in the case
of the Pico editor; Q to quit various other editors). Save the changes, if prompted. From the
command prompt, type: /sbin/lilo to restart LILO, then type: exit to logout as the root
operator and then exit again to close the shell session. The next time you reboot Linux
should be using all the memory installed. (Use the cat /proc/meminfo command to verify.)
Warning: It is possible to render Linux unbootable by specifying a memory size larger than
that physically installed and available. If your computer reserves some memory, rendering it
unavailable to Linux, you must specify the lesser amount. For example, during power up
(POST), if your memory count shows something like 130496 (kilobytes), rather than the full
131072 KB for 128MB of RAM (128 * 1024), this means that some of the memory is reserved
by the hardware. In this case, do not use mem=128M in the example above. Instead, use the
exact amount of 130496K (append=“mem=130496K”), or play it safe and “step down” to
127MB, as in: append=“mem=127M”. If you somehow manage to leave your system in an
unbootable state (i.e., it locks up on boot when it tries to use the reserved memory), the
remedy is to edit the /etc/lilo.conf to correct the memory size.
Q. I have a printer attached to the parallel port. What is this port called in
Linux?
When prompted for the name of a parallel port in Linux, in place of LPT1, LPT2, or LPT3 use
/dev/lp0, /dev/lp1, or /dev/lp2, respectively.
Q. I have a device attached to a serial port. What is this port called in
Linux?
When prompted for the name of a serial port in Linux, in place of COM1 through COM8
substitute /dev/ttyS0 through /dev/ttyS7, respectively.
Preparing Today for Linux Tomorrow
14
Q. What is my floppy drive called in Linux?
If prompted for the path of a diskette drive in Linux, in place of “a:” substitute /dev/fd0,
/dev/floppy, or /mnt/floppy, depending on the Linux distribution you are using. Unlike the
DOS/Windows world, UNIX/Linux doesn’t use drive letters. Devices are assigned “mount
points” (or mounting locations). To determine which mount location is assigned to the floppy
drive on your system, type: mount at a command prompt. You should see something like:
/dev/hda1 on /type ext2 (rw)
none on /proc type proc (rw)
none on /dev/pts type devpts (rw,gid=5,mode=620)
/dev/fd0 on /mnt/floppy type vfat (rw,nosuid,nodev,user=username)
Look for the part that starts with /dev/fd0 on…. This tells you what “soft” mount location to
use, instead of /dev/fd0: “/mnt/floppy”, in this particular case.
Q. Wait a minute! Are you saying that there are no drive letters in Linux?
That is correct. Hard drives, floppy drives, CD-ROM drives, Iomega Zip and Jaz drives, and
other devices that get drive letters in DOS/Windows are instead assigned mount points by
Linux. Depending on the device involved, here are some of the designations used by Linux,
following their typical Windows drive letters:
Floppy drives
A: drive (in Windows) = /dev/fd0 (or /dev/floppy or /mnt/floppy in Linux, as explained above)
B: drive = /dev/fd1 (see note, below)
IDE/EIDE disk drives
C: drive (boot drive; first/primary drive on the first disk controller) = /dev/hda
D: drive (second/slave drive on the first disk controller) = /dev/hdb
E: drive (first/primary physical drive on the second controller) = /dev/hdc
F: drive (second/slave physical drive on the second controller) = /dev/hdd
(CD-ROM drives will often be attached as the primary drive on the second controller, or
/dev/hdc. IDE/ATA Zip drives will generally take the next available drive letter after the hard
drives and/or CD-ROM drives.)
SCSI disk drives
C: drive (boot drive; lowest-numbered drive on the primary controller) = /dev/sda
D: drive (second lowest-numbered drive on the primary controller) = /dev/sdb
E: drive (third lowest-numbered drive on the primary controller) = /dev/sdc
(And so on. There can be many SCSI drives, on one or more controllers. The mount point
designations just keep incrementing up through /dev/sdz. If there are more than 26 physical
drives, they will be incrementally labeled from /dev/sdaa up through /dev/sdzz.)
Note: Unlike the primary floppy drive (/dev/fd0) Linux doesn’t automatically map a second
diskette drive (/dev/fd1) to something like /dev/floppy or /mnt/floppy. However, if desired
you can set up your own alias, such as /mnt/floppy2. The mapping entries that are used for
this are located in the /etc directory, in a file called fstab. To add an alias for fd1, use an
editor to open /etc/fstab and look for the entry for fd0, which should look something like:
/dev/fd0 /mnt/floppy auto noauto,owner 0 0
This line maps /dev/fd0 to the directory /mnt/floppy. To create an alias for fd1, create a
subdirectory in the /mnt directory called floppy2. Then add a new entry (using whatever alias
you wish) to the fstab file in order to map the second floppy drive to a mounting point, such
as:
Preparing Today for Linux Tomorrow
15
/dev/fd1 /mnt/floppy2 auto noauto,owner 0 0
Save the file and close the editor. Now, when you use the Mount command it should include
something like:
/dev/fd1 on /mnt/floppy2 type vfat (rw,nosuid,nodev,user=username)
Q. What about disk drive partitions? Don’t they have drive letters either?
Nope. Linux just tacks numbers onto the end of the drive identifier. For example, if the first IDE
hard drive had three primary partitions, they would be named/numbered, /dev/hda1,
/dev/hda2 and /dev/hda3. If a second hard drive had two primary partitions, they would be
called dev/hdb1 and /dev/hdb2. Likewise, a SCSI hard drive with three primary partitions
would be labeled /dev/sda1, /dev/sda2 and /dev/sda3.
That is just for primary partitions. If you also have extended partitions, they begin numbering
at 5 and continue from there. If, for instance, you had one IDE drive with two primary and two
extended partitions, the partitions would be called: /dev/hda1, /dev/hda2, /dev/hda5 and
/dev/hda6.
Q. Wow, that’s confusing! Isn’t there a simpler way to access drives?
Yes. Most Linux distributions provide desktop icons for some devices, such as floppy and CD-
ROM drives that can be clicked on for direct access. If you install software for accessing an
Iomega Zip drive, it should also provide an icon to use. And, of course, the graphical file
managers, such as the K File Manager and Gnome File Manager allow you to access files on
various drives without worrying about drive letters. (When you open KFM, look for an icon
called mnt/. When you single-click on it, KFM will display all the devices that are ready for use.
Click on the icon called cdrom/ to display the contents of the currently inserted CD, or click on
the one called floppy/ to see what is on the inserted diskette.)
Q. Is there an easy way to see what devices, such as drives, have been
mounted?
Yes. Red Hat users running KDE can click on the KDE menu and select Red Hat, then
System and finally Disk Management. This will bring up a panel that shows the devices
mounted, and provides Mount and/or Unmount buttons (so that you can take a device offline,
then remount it). The Mount/Unmount buttons will toggle back and forth between the two
options being used. There are also buttons provided to let you format the drives.
Gnome users (of any distribution) can select File manager from the Gnome icon on the Panel
(the Linux “taskbar”) and then click on the /mnt icon to view the various devices.
I have not seen a KDE equivalent for the Red Hat Disk Management tool for other
distributions. As an alternative, the Mount command can be used, as described earlier, in the
question What is my floppy drive called in Linux?. Also, for a less cluttered view of disk drives
mounted, with total and available space, try the disk filespace command (use: df –h to
display measurements in “human-readable” form—gigabytes, megabytes, etc.). It won’t show
non-disk devices, however. (Read the info manual pages at: info df, for more information.)
Q. Is there any way to change the default language used by the operating
system and applications?
The short answer is “Yes.” Of course, as with most things Linux, how you do it depends on the
distribution you are using, the user interface (desktop environment) you are using, your
Preparing Today for Linux Tomorrow
16
applications and perhaps other factors. The default language is set during installation,
however there are ways to change this once the system is up and running.
If, like many users, you are running the KDE environment, the way to change the language is
very simple. From the KDE menu:
1. Select Settings, then Desktop, then Language.
2. Specify your first choice from nearly three dozen languages (including Catalan, Slovak
and Macedonian). You are allowed to choose up to three languages. (When you start a
program, KDE will try to find a translation for the menus, dialog boxes and messages in
the first language. If it fails, it will try the second choice and, if necessary, the third. If all
three fail (or if you do not select three languages), it will use whatever was the original
default language specified during installation.)
3. Press OK to save your changes and exit. (Apply will save your changes but not exit. The
Default button will return all three language settings to the original “Default language”
choice.)
4. Log out and then log in again to see your changes.
Note: This procedure only affects certain system utilities, help files, error messages and the
like. Your applications may or may not be affected, depending on what support for other
languages are programmed into them. Therefore, you may end up with a mix of two or more
languages depending on the various programs you are running at any one time. You may be
able to set up language choices within specific applications as well.
If, after logging out/in you discover to your horror that you somehow selected the wrong
language and you can’t read most icons, menus and other system functions, try to follow the
instructions above. If you can manage to return to the language settings panel, but can’t read
the language names, the little flag icons should allow you to find the language you want, or
you can press the Default button at the bottom—the second one from the left, if you can’t read
it. If you are unable to find your way back to the language settings you may have to reinstall
Linux to specify the default language all over again.
Gnome users can use the same KDE tool, if they wish. The only extra step is to go to the
Gnome menu, select the KDE menus option, then proceed as directed, above.
If you use another desktop environment besides KDE or Gnome, refer to the manual or tutorial
that came with the environment for information on how to change the language.
General Usage
How to perform various tasks using Linux.
Q. How do I format a floppy (diskette)?
If you are using the KDE user interface, the easiest way is to use the KDE Floppy Formatter
utility supplied. Simply click on the KDE icon on the Panel, select Utilities, then KFloppy. The
File System option gives you the choice of the Dos (FAT) file system (for DOS/Windows or
OS/2) or the Linux native ext2fs file system.
If you are using Gnome, go to the Gnome icon (the fancy G-shaped footprint) on the Panel
and select Utilities, then GFFloppy.
From the command shell, to format a diskette for the native Linux file system, use the mke2fs
command. You can use the mtools utility to access DOS/Windows-formatted diskettes. Use
the commands: mformat, mcopy and mdel without parameters to see a list of command
Preparing Today for Linux Tomorrow
17
options. (Using mtools, to change from one directory on the floppy to another, use the
following command: mcd a:subdir, where subdir is the name of the directory you wish to
change to.)
Q. Is there online documentation for most programs?
Yes. The official help tool for Linux is called man (short for manual). To read the manual for
various programs (including many commands), type: man <something> (where
<something> is the name of the program you need help with) at a shell prompt. For example,
to read the online manual for the man program itself, use: man man. You can also use a
viewer called less to view the manual: man man | less (where | is the piping symbol—
Shift-Backslash—located above the Enter key on a 101-key keyboard). Using less, press
the Spacebar to move ahead a page at a time; to back up, use the B key. When you are
done, press Q to quit.
Many newer programs have shifted to the info tool rather than using man. Like man, you can
view info pages using the command: info <something>. Some programs now supply
HTML or XML pages, kept in the /usr/doc/<programname> directory, in place of either man
or info pages. These documents may be viewed from a Web browser or other HTML/XMLcapable
program.
Another option is to redirect the manual to a text file for browsing with an editor. For example:
man mv | col –b > move.txt (to redirect the text for the mv command) will strip out the
screen formatting codes that do not translate into print and then copy the output to a file called
move.txt (col is the name of the formatting utility). Once it is in a text file, it can be read with
Emacs, Pico, Vi, or any other text editor. (Note: Substituting >> for > in the example above
would append the text to the end of an existing file, whereas using > a second time would
overwrite man.txt with new data.)
Likewise, to print a manual you will need to pipe (send) the document to a formatting program
that will strip out the screen formatting codes before printing. To do this for info pages, type:
info mv | col –b | lpr. (lpr is the program that sends the document to the print
spooler.)
Another “quick and dirty” source of helpful information is the whatis command. Whatis
provides a brief synopsis of many commands. For example, try: whatis man. You will be
shown several “definitions” of what the man command does. The same goes for: whatis
whereis, to find out what the whereis command does. If there is no information available for
a particular command you will see a “nothing appropriate” message instead. Note: If the
whatis command doesn’t work, it may mean that the whatis database has not yet been built.
To create it, first use the: su – command to log in as root, then enter:
/usr/sbin/makewhatis. This may take several minutes. When the process finishes, try the
whatis command again.
Q. How do I kill a program that has locked up?
If you are not the owner of the program (the user account who installed it), you must log in as
root. (For instructions on how to do this, see the Root/User Accounts, Groups and
Permissions section, below.) Then, from a command line, you can type: ps (process status) to
see what programs are running. To the right of the process name will be a Process ID (PID)
number, such as 506. This is what you will use to kill the program. From the command prompt,
type: kill pid (for example, kill 506). This tells the program to either shape up (respond
normally) or ship out (close itself). This will usually solve the problem, but occasionally
stronger measures are required. Preceding the PID with –9 (such as: kill –9 506) will kill
Preparing Today for Linux Tomorrow
18
the program. (This is sometimes called “signal 9,” not to be confused with “Plan 9 from Outer
Space”, a cheesy B-movie.)
An alternative for killing a background process, or daemon is to use the –HUP (hangup)
parameter, as in: kill –HUP 506. It essentially signals the program to close itself, then
restart and reread its configuration file. Again, if the kill command doesn’t work the first time,
you may have to resort to the –9 parameter rather than –HUP.
Q. Is there a way to increase the priority of a program to make it run
faster?
Yes. If you have a program that normally takes quite a while to run and you really, really need
it to complete faster, you can boost its job priority level so that it uses more processor cycles.
Of course, this means that all other currently executing programs will run slower. But if this is
what you want, the nice command will do the job for you.
Linux supports priority levels from 19 (lowest) to –20 (highest), with the default being 0. To
change the priority to a negative number (increasing the priority), you must first log in as root.
Then, to start a program called Myprog with a high priority-level of –10, use the command:
nice –n –10 myprog. If you wish to increase the priority of an already running program,
use the renice command, identifying the program by its PID: renice –10 506 (the –n
argument is not used).
Conversely, if you have a long-running program that is a low-priority job and it is slowing the
entire system down while the program is running, you can lower its priority using the same
techniques by providing positive numbers—such as 8—rather than negative numbers. By
default, nice will reduce the priority by 10. In other words, the command: nice myprog is the
same as: nice –n 10 myprog.
It may take some experimenting for you to find the right priority levels to assign to specific
programs. Too high a priority may bring the rest of the system to its knees; too low and the job
may take forever to complete. As a rule of thumb, unless you have a reason to be changing
priorities, leave them alone. The default setting is best most of the time.
Q. Is there a spelling checker that I can use outside of a word processor?
Yes. Linux includes a program called ispell, which will read any text file, highlight
questionable words, offer alternatives and let you add the words to the ispell dictionary if
desired. It runs from the command line, with a text-mode (i.e., non-graphical) interface, but is
fairly easy to use. There are a number of parameters that you can use (type: ispell --? for
a list of them or info ispell for more detailed explanations). But the simplest usage to
check a file called memo.txt would be: ispell memo.txt. It will even check a series of files
in order, such as: ispell memo.txt mydata.txt or even ispell *.txt (to check all
files in the current directory with a .txt extension).
When the program runs it will present a “questionable” word within the context of the line of
text containing the word and offer suggested replacements in a numbered list. To select one
of the suggestions from the list to replace the original word, simply type in the number of the
word you prefer. There are also several menu options displayed along the bottom of the
program. To Replace the word with something you type in, type R (or r, ispell options are not
case-sensitive) and then whatever word you want to replace the original with. (You also can
type in the number of a replacement word from the list. This is the same as just typing in the
number directly, but is there in case you forget you can do it that way. Use A to Accept the
existing word as is (i.e., ignore it) for the rest of the spell-checking session. (Using the
spacebar tells ispell to accept/ignore the word this time only.) I lets you Insert (or add) the
Preparing Today for Linux Tomorrow
19
existing word into the ispell dictionary. Entering an L lets you Look up other words in the
dictionary. (Perhaps you would like to find another word not on the suggestion list.) If the word
is correct but capitalized inappropriately, choosing U will Uncapitalize it. Either Q or X will end
the program. If you have specified a list of files to spell-check, X will stop checking the current
file, save it and begin on the next one. This is essentially “save and exit.” On the other hand, Q
will ask whether you really want to Quit immediately without saving your changes. (Type Y or
N, depending on your preference.)
Be sure to check out the associated online help for that program by entering ?. In addition to
explaining the above options, there are others listed that you might not know about or
remember, including Redraw screen (^L).
Warning: ispell is not discriminating as to what files it checks. It is entirely possible to spellcheck
a word processing document or a binary program file. Saving changes to a document
created by a word processor may (or may not) corrupt it, but the word processor probably has
its own built-in spelling-checker anyway, so why bother with ispell? On the other hand, you
will corrupt a program if you change anything and then save the changes with the X option. If
you accidentally open a binary file (easily identifiable by the “gibberish” displayed that ispell is
trying to correct), be sure to Quit immediately with the Q option!
GUI/Desktop
The following questions involve window manager/desktop and other graphical user interface
questions.
Q. My copy of Linux came with KDE and Gnome. Why do I need both?
Technically speaking, you probably do not need both. However, both desktop environments
are popular, so this gives you the option to decide which you like better. Also, some programs
will work only with one or the other, allowing you to use both types of programs, switching
between the two desktops as needed.
Q. I have both KDE and Gnome installed. How do I switch between them?
From the graphical interface, simply log out. Then at the Log in screen, type your login ID and
password and choose which desktop you wish to load (Session Type). This choice will remain
your default until you change it to something else.
Q. Can I add more desktop themes to use with Gnome and KDE?
Yes. One source for more Gnome themes is http://gtk.themes.org. To install a new theme in
Gnome, after downloading it, start the Gnome Control Center by clicking on the “toolbox” icon
on the Panel. Then click on Theme Selector (under Desktop) in the panel on the left. Press the
Install new theme button (in the center of the right-hand panel), and when the dialog box
opens, scroll down to the name of the file you downloaded (called something like
new_theme.tar.gz and most likely in your login directory). Click on the name of the theme you
downloaded and press the OK button. When the theme appears in the Available Themes box
you can press the Preview button (if Auto Preview isn’t already selected) to see how it looks
before applying it. If you think you might like it, press the Try button to conditionally apply it to
your desktop. If you approve of the results, press OK to save the changes. If you decide you
do not like it after all, use the Revert button to back out the theme and return to your previous
theme.
KDE themes can be found at http://kde.themes.org. To install a new KDE theme, start the
KDE Control Center from the “home” icon (it looks like a little house in front of a manila folder)
on the Panel. Then, click on Theme Manager (under Desktop) in the panel on the left. Press
Preparing Today for Linux Tomorrow
20
the Add button (in the upper-right part of the right-hand panel). When the dialog box opens,
find the name of the theme file you downloaded. Double-click on the name to install it. Then
click on the theme name in the Installer box. The window in the middle shows how it will look.
If you like the appearance, press the Apply button at the bottom of the panel to change your
desktop theme. If you do not like the appearance, press Cancel to exit the KDE Control
Center.
Other sources for themes include many Web sites that offer software downloads. For a list of
some of these sites, refer to the What Good is a Linux Client? white paper, available from the
same sources as this Q&A paper.
Q. How can I add programs to the Panel?
KDE users can click on the KDE icon, select Panel, then Add Application and then choose the
desired program from the various menus shown. Once the application has been added to the
panel, you can rearrange the order of the icons by right-clicking on an icon and selecting
Move, then dragging the icon to wherever you want it.
Gnome users must first log in as root, then click on the Gnome menu icon, then Programs and
Settings, then select Menu editor. When the popup menu appears you can create a submenu
or add to an existing menu.
Q. What is a virtual desktop?
When running a number of programs concurrently, the screen sometimes becomes cluttered
with several windows open at once, covering each other up. One solution, of course, is to
minimize some of the windows until they are needed. However, this entails minimizing and
restoring those windows repeatedly, as different windows are accessed. This can be tedious
over the course of a day.
An alternative is to use a program that creates “virtual desktops.” Each desktop is a clean
slate where you can open one or more programs. Rather than minimizing/restoring all those
programs as needed, you can simply shuffle between virtual desktops with programs intact in
each one. (If this doesn’t seem like much of an improvement, consider that with eight open
programs, a user might have to minimize and restore each of those eight programs multiple
times in turn. With eight virtual desktops, each program could stay open at all times, needing
only one click to change to the next, instead of two.) Virtual desktops can help reduce desktop
clutter.
There are several programs of this type available for Microsoft Windows users to buy or
download, but Linux users do not even have to look for such a program. Both KDE and
Gnome include virtual desktop support by default and users can specify how many desktops
they want to use—or disable them entirely. To switch between desktops, click on the buttons
on the Panel (for One, Two, Three, etc.), start one or more programs, then select another
virtual desktop, start more programs and so on. To cycle through all the desktops in turn, use
Ctrl-Tab.
Q. You said I could have eight virtual desktops, but I see only four.
Where are the rest?
The default setting is four, but you can easily change the number in increments of two (i.e., 2,
4, 6 or 8). To do this in KDE, click on the KDE icon on the Panel, then select Panel and
Configure. This will display the KPanel Configuration window. Use the Visible slider to change
the number of virtual desktops. (The Width slider lets you control how big the icons are on the
Panel.)
Preparing Today for Linux Tomorrow
21
If you are using the Gnome environment, from the Gnome icon on the Panel go to Program
and select Setting, then Gnome Control Center. When the Gnome Control Center panel
appears, click on Desktop, then Windows Manager and finally Workspaces. Select the desired
number of virtual columns or desktops and press OK to close the panel.
Q. Is it possible to share a program across multiple virtual desktops?
Yes. In the upper left-hand corner of a program window there is an icon that looks like a
pushpin. Pressing this button will “pin” that application in place, making it appear in all virtual
desktops, in the same position onscreen. To remove it from other virtual desktops, press the
button again and it will remain on only the virtual desktop from which it was “unpinned.”
Q. How can I change my desktop background color?
KDE users can click on the KDE icon on the Panel and select Settings, then Desktop and
finally Background. From the Display Settings dialog box you can pick which desktop (from
among the virtual desktops) to modify, then choose one or two colors and/or wallpaper to use.
To test out your choice, press the Apply button. When you are finished, press OK to save your
changes, or Default to go back to the original system default colors. (Cancel will not abort the
changes once you press Apply; it will only close the panel.)
Q. Is there a faster way to copy/cut & paste text than to use the pulldown
menu?
Yes. There is a quick way to copy text with the mouse. If you are using a three-button mouse,
first position the cursor where you wish to place the text, then select (highlight) the text as you
normally would with the mouse. Finally, click the middle mouse button. (If you have a twobutton
mouse set up to emulate a three-button mouse, press both the left and right mouse
buttons simultaneously to achieve the same effect.) The highlighted text will be copied to the
cursor position. This should work in most text editors and command line sessions, as well as
in word processors and other graphical programs. Alternatively, you can paste the text with
Shift-Insert (hold the Shift key while pressing the Insert key), rather than using the third
mouse button.
For a keyboard method, you can highlight the text (using the mouse or holding down the Shift
key while using the left and right arrow keys to “paint” the text) then use Ctrl-C to copy (or
Ctrl-X to cut) the text, then Shift-Insert (or the third mouse button) to paste; similarly to the
Ctrl-C/Ctrl-X/Ctrl-V trio used in Windows.)
Note: Because of all the different user interfaces, editors and applications used in Linux, not
all of these techniques will work in every program, so you may need to experiment to see what
works where.
Command Shell
The next series of questions are specific to using commands and command shells.
Q. If I want to issue a command, how do I open a command prompt?
In Linux parlance, it is generally referred to as a command shell, or simply a shell. Each Linux
distribution includes at least one shell. The most commonly used one today is bash (short for
Bourne Again SHell), an enhanced version of the older Bourne shell. To open the default shell
for your distribution, KDE users can click the Kvt icon on the Panel (the icon that looks like a
computer terminal), or press Ctrl-Alt-F1. This will provide a command line interface (CLI) from
which you can run commands as needed.
Preparing Today for Linux Tomorrow
22
Gnome users can go to the Gnome icon on the Panel, select System and open either a
regular or color Xterm (terminal) session, or command prompt window, by clicking on the
icons that look like a computer terminal with a large red or multicolored X in front of it,
respectively. Alternatively, you can right-click anywhere and open a new terminal by selecting
New, then Terminal.
If you need only a “quickie” command line, just long enough to execute one command and
then close itself, press the Alt-F2 keys. When the pop-up panel appears, type the command
and press Enter. Instead of a command, you can enter a data file name (with the path
necessary to find it) and Linux will launch the appropriate program, based on the file
association. (For example, typing: /home/username/.netscape/bookmarks.html will
cause the default browser to launch.) If you use a path to a directory instead of a file, Linux will
open the directory using the default file manager, such as KFM.
Q. What are the Linux equivalents to DOS/Windows commands?
There are a number of commands in Linux that are similar or identical to either the name or
function of DOS commands. Here are some common DOS commands with their Linux
counterparts:
Command Purpose DOS Command Linux Command
Change to the parent directory
cd.. or cd .. (short for
change directory)
cd ..
Change to the root directory cd\ or cd \ cd / (note the forward slash)
Clear the screen cls clear
Close a command prompt window exit exit
Compare the contents of two files fc (file compare) diff (difference)
Copy a file copy cp (copy)
Create a directory
mkdir or md (make
directory)
mkdir
Delete files del or erase rm (remove)
Delete directories rmdir or rd (remove dir) rmdir
Display file contents, a screen at
a time
more more or less
Display the amount of available
RAM
mem free
Display/change the date date date
Display/change the time time date
Echo output to the screen echo echo
Find a file by filename dir or attrib locate
Find a file by date or other file
attributes
dir (not by date) find
Find a text string in a file find grep
Format a floppy format
mke2fs (or mformat, for a
DOS/Windows floppy)
List files in directory dir (directory) ls (list) or dir
Move files move mv (move)
Rename a file ren (short for rename) mv
Show the current filepath cd pwd (print working directory)
Even though some of the commands are the same, the parameters for these commands may
differ between DOS/Windows and Linux. For example, let us look at the differences between
changing the date and time in Linux versus Windows.
Preparing Today for Linux Tomorrow
23
In Windows, at a command prompt you would type: date 12-31-2000 to change the date to
December 31, 2000.
Then, to change the time to 2:00 p.m., you would use: time 14:00:00 (or simply time
14:00).
In Linux, on the other hand, first you would have to be logged in as the Superuser (or use the
command: su – to temporarily grant yourself that authority), then you would use the
command: date 123114002000. Notice that the 14:00 is embedded in the date string,
between the day and the year. This is the way Linux wants it. You can’t specify seconds using
this command—just hours and minutes.
However, if you want to change only the time—but not the date—use the following command
instead:
date –s 1400
or, if you want to specify seconds:
date –s 14:00:35
(The colons are required when specifying seconds.) In both examples, the –s parameter tells
Linux that you wish to set the time.
For more on the su command, see the Root/User Accounts, Groups and Permissions section,
below.
Q. How do I get help for command parameters?
There are several options available. First, many common commands will have short “manual”
pages built into Linux. You can access them by typing: man, followed by the command name,
such as: man mv to see the parameters for the mv (move) command. Some utilities provide
more detailed (and sometimes newer) command help using the info tool; for example: info
mv. If these tools do not produce results, try following the command with –help or –h, as the
only parameter (as in mount -h or mv --help). Some commands, such as: mformat, will
display the list of available arguments whenever the command is used without parameters, or
with invalid parameters. When in doubt, try them all until you find the one that works.
To learn how to use the info command, type: info info (the tool uses itself to display the
tutorial). To learn more about man, use: man man and/or info man. (The tutorials are
different, so it is not a bad idea to consult both.) When you are finished reading the help
displayed by man or info, press the Q key to quit and return to the command shell.
Q. How do I repeat a command?
If you tend to use the same commands over and over, or recently typed in a rather lengthy set
of parameters you wish to use again, there are a number of ways you can repeat a command
without having to retype it. Linux stores the last 500 commands in a file (called .bash_history)
for quick retrieval. The simplest way to find a recent command is to use the up and down
arrow keys to scroll back and forth through the list until you find the one you want. If you wish
to change a command you used recently, use the up/down arrow keys until you find the
command you want, then use the left/right arrow keys to go to the part you wish to change.
Type in anything you need to add and use the Del key to delete unneeded text.
Alternatively, typing: !! (pronounced “bang, bang”) and pressing Enter will execute the last
command entered.
Preparing Today for Linux Tomorrow
24
If the command is too far back up the list for scrolling to be convenient and you know the
name of the command, you may prefer to jump to it directly, using a shortcut: ! string
(where string represents the command). For example: ! mcopy might find the previously
used command: mcopy –tnm something.txt c:/mydir and execute it directly.
If you are not sure of the command name or have used several variations of the command
and are not sure which one you want, you can view the entire history file using the command:
history (without parameters). The commands will be numbered in the order they were
executed. When you find the command you want, you can execute it simply by typing: ! 452
(or whatever the correct number is). If you do not want to view the entire 500 line history, you
can elect to see just the last x number of commands, such as: history 50 (to see the last
50).
If you need to search the entire 500 item list but do not want to read every line looking for a
command, you can use the grep command to search the contents. For example, to find that
mcopy command from earlier, type: history | grep mcopy (where |, the piping symbol, is
the Shift-Backslash character above the Enter key, on a standard 101-key keyboard). You
can even use partial-word searches with grep, such as: history | grep m (to find the
mformat, mdel and other commands starting with “m” that were executed previously), or even
wildcard searches (history | grep m*y). If none of these methods finds the command
you are looking for, then most likely it was executed more than 500 commands ago and has
“fallen out” of the .bash_history file.
Q. What can I do if I don’t remember the name of a command I need?
You can page through a manual looking for the command you want, but a quicker way is with
the apropos command. It allows you to search a database of available commands, using
keywords. For example, if you want to operate a floppy drive but do not remember what
command you used the last time (or perhaps this is your first time and you do not know where
to start), type: apropos floppy. You should see a result like:
fd (4) - floppy disk device
fdformat (8) - Low-level formats a floppy disk
mbadblocks (1) - tests a floppy disk and marks the bad blocks
in the FAT
mformat (1) - add an MSDOS filesystem to a low-level
formatted floppy disk
mkbootdisk (8) - creates a stand-alone boot floppy for the
running system
setfdprm (8) - sets user-provided floppy disk parameters
(The numbers in parentheses indicate the section in the man pages where this information
can be found, such as: man mformat to look at section 1 (the default) in the man pages for
the mformat command, or: man 8 mkbootdisk to go directly to section 8 to see the rmt
command.)
If you were to use the command: apropos diskette instead, Linux might return a “nothing
appropriate” message. This indicates that apropos doesn’t recognize the term “diskette.” If
this happens on a search, simply try another term to look for, such as “floppy.” It is also
possible that apropos won’t find every possible related command. For instance, if you search
for information on tape drives using: apropos tape, it would list a number of commands
related to operating the tape drive, including mt and rmt, but it would not include tar, which
can be used for tape backup, so you may need to be creative to find what you are looking for.
Preparing Today for Linux Tomorrow
25
Note: If the apropos command doesn’t work, it may mean that the whatis database (which is
accessed by the apropos command) has not yet been built. To create the database, first use
the: su – command (and type in your password when prompted) to log in as root, then enter:
/usr/sbin/makewhatis. This process may take several minutes. When it finishes, try the
apropos command again. For more on whatis, refer to the question, Is there online
documentation for most programs?
Q. I just used a command and got completely different results from the
last time I used it. Why?
Remember that Linux is case-sensitive. Verify the case of the parameters you used for the
command. For example, if you type: tar –xvf filename.tar one time, and: TAR –XVF
FILENAME.TAR on another occasion you will get different results, because x and X, v and V, f
and F, are all valid, yet different, parameters for the tar command. In fact, because
commands are case sensitive as well, you could have two completely different commands,
called tar and TAR—with different syntax and parameters—on the same system.
Q. Long pathnames are a pain to type in. Is there any sort of shortcut I
can use?
Yes, in some cases. If you are using the Bash shell, or other shell with similar capabilities, you
can take advantage of a feature called filename expansion, by using the Tab key. You can
type the first part of a directory name, then press Tab and it will complete the name for you (as
long as there are no other names starting with the same letters.
For example, if you want to list the /home/johnson/projects/2000 directory, you could type:
ls /ho[Tab]/jo[Tab]/pr[Tab]/2 and the shell will expand the pathname for you. But,
what happens if you have two directories under /home/johnson, called /projects and
/previews? If you know this ahead of time you could just change /ho/jo/pr/2 to /ho/jo/pro/2,
to eliminate the conflict. Alternatively, if pressing the Tab key after /pr doesn’t provide the
pathname expansion, pressing Tab a second time will bring up a list of all names that start
with “pr”. (In other words: ls /ho[Tab]/jo[Tab]/pr[Tab][Tab].) Then simply use the
keyboard up-arrow key to repeat the previous command and change “pr” to “pro”, or whatever
else is needed to create a unique match; then finish the command with the /2 at the end.
Q. Is there a way to use one command to start more than one program at
a time?
Yes. The command shell allows you to “stack” commands on one command line, to be
executed in order. For example, if you type: ls –l; cd ..; ls –a Linux will list the
current directory contents in “long form” (ls –l), then change directory to the parent of the
current directory (cd ..), and finally list all files in that directory (ls –a). The “semicolonspace”
separating the individual commands is how Linux knows where one command ends
and the next begins. If you wanted to copy a file from one directory to another and then to a
floppy (while changing its name), and then display the contents of the floppy, you could type:
cp /dir1/myfile.dat /dir2; cp /dir1/myfile.dat /dev/fd0/myfile2.dat; ls –a /dev/fd0
Stacking commands can be an efficient way of starting several programs that must run
sequentially, without having to stand around waiting for each to finish in turn. Simply type in
the stacked commands and walk away until the last one is done. Of course, if you need to
perform several tasks end to end on a recurring basis, it makes more sense to create a “script”
file (like a batch file, for you DOS/Windows users) containing those same statements. Then
you just have to type in the name of the script file, instead of a long multi-part command.
Preparing Today for Linux Tomorrow
26
Warning: Like many powerful tools, command stacking can be a double-edged sword. It can
hurt you as easily as help, if you are not careful. As an illustration of what can happen if you
are careless, years ago I was using the OS/2 operating system to write a book about OS/2,
which also supports command stacking (although without requiring the semi-colon between
commands). I decided to delete the existing files from an old diskette so that I could back up
my book chapters (one file per chapter) to it. I intended to type in the following command: del
a:\*.* which would have deleted all files from the root directory of the a: drive.
Unfortunately, what I actually typed was del a: *.* (a space where the backslash should
have been). OS/2 interpreted that as “delete all files from a: (del a:) and then delete all files
from the current directory (del *.*)”—all my original files! I realized what was happening
about halfway through the files on the hard drive and was able to cancel the job before the
rest were deleted. Fortunately, I had a backup from the night before, so I lost only a few hours
work, but it could have been much more serious. The moral of this story is be very careful with
command stacking (or script files), because a lot can happen when you run programs
unattended. If you were to issue a command to first copy a file to a floppy and then delete the
original, and the floppy copy was unsuccessful (the disk was full, or not formatted, or
defective), it is possible for the file to be deleted without having been backed up first.
Q. Is there a way to stack commands and have them execute
concurrently in other command sessions?
Yes. Running the stacked commands in the examples from the previous question will cause
them all to execute one after the other in the same session window, rendering it unavailable
until all the commands have finished running. If you would rather the commands launch
another command session and run the commands from there, simply append an ampersand
(&) symbol to the end of the first command, such as:
cp /dir1/myfile.dat /dir2&; cp /dir1/myfile.dat /dev/fd0/myfile2.dat; ls –a /dev/fd0
In this case, Linux would launch a new session for the first command, then continue to
execute the remaining commands in that session when the first one is finished. Meanwhile,
the original session is available for your use.
If for some reason you want each command to run in its own session, append the ampersand
to all of them, as follows:
cp /dir1/myfile.dat /dir2&; cp /dir1/myfile.dat /dev/fd0/myfile2.dat&; ls –a /dev/fd0&
Of course, if you are doing this just to free up the original session so you can keep working
while the other sessions are busy running those commands, a simpler solution might be to just
manually launch a second command session to work from while the first one is busy.
Launching several programs in different sessions offers possible advantages. If, for example
one program is doing heavy database updating while a second is performing spreadsheet
recalculations and a third is displaying complex graphics onscreen, each is primarily stressing
a different part of your computer’s hardware. (The database update is using the hard drive
heavily, the spreadsheet is grabbing a lot of processor cycles to calculate numbers and the
graphics program is mainly using the video controller.) Running them all at once could greatly
improve the performance of your computer by using system resources that would otherwise sit
idle until the current program finishes.
Bear in mind, however, that how useful this is will depend to a great degree on the kinds of
programs you are running and the type of hardware your computer has installed. For example,
if you run two programs at once that both make heavy use of your hard drives, people with
multiple IDE drives (or one SCSI drive) will likely see little or no performance improvement by
doing this. This is because the time spent jumping back and forth between hard drives, or
Preparing Today for Linux Tomorrow
27
even between areas on the same hard drive, offsets much of the time savings of running the
programs together. IDE devices are “dumb,” relying on the processor on the IDE controller to
handle all the disk drive instructions, but the controller can access only one drive at a time, so
it has to stop working on one drive, switch to the other, work for a while, then stop and return
to the first. This is very disruptive. On the other hand, for those computers with SCSI drives,
the situation is very different. SCSI devices are intelligent (which is why they typically cost so
much more than IDE drives) and the SCSI controllers can multiprocess, allowing all SCSI
devices to execute concurrently and independently of one another. So if you have two
programs accessing data on two different SCSI drives, they can both run full speed all the
time. Of course, regardless of disk type, if you do not have enough memory to run both of
these programs concurrently, Linux will have to alternate between them in memory, swapping
all or part of one program out to the hard disk temporarily. This extra disk activity will reduce
the amount of time that can be spent reading/writing the data files, again hurting performance.
Similarly, if one program is making heavy use of the processor and the other is primarily
graphics-oriented, there still may not be as much an advantage as you might think to running
them concurrently. If you have a high-end video card with lots of memory, it may be able to
process all of the graphics commands using the video controller’s onboard processor and
memory. However, many computers—especially those at the low end of the price range—use
inexpensive video controllers with little or no video memory (some share the computer’s
memory instead, using something called Uniform Memory Architecture, or UMA). This means
that the computer’s processor may have to process all of the spreadsheet calculations and the
graphics calculations at once, reducing the performance of both programs. Moreover, using
the main computer memory for both may again result in swapping, unless you have lots of
system RAM.
Still, it can’t hurt to try executing your stacked commands concurrently. If they all complete
faster than if run sequentially, terrific. If not, you can always go back to running them the “old
fashioned” way.
Logging In/Logging Out/Shutting Down
This section contains questions or problems involving shutting down Linux, or logging out of
an account.
Q. I can’t seem to log in, even though I’m using the correct password.
What’s wrong?
Linux is case-sensitive, meaning that a password of a1b2c3 is not the same as A1B2C3. Be
sure that the Caps Lock key is turned off before typing your password (unless, of course, your
password is all caps).
Q. Why do I need to use Shutdown? Can’t I just turn the power off?
Just as with Windows, using Shutdown allows the system to save open files, flush the system
cache and do other necessary system maintenance. Turning the power off abruptly
circumvents this housekeeping and can result in lost data and other problems. Also, because
the file system is left in a “dirty” state, it must be cleaned up the next time the system is
booted, slowing the reboot considerably.
Q. I can find a Logout option, but where is Shutdown?
Using the KDE user interface, click on the KDE icon (the large stylized K) on the Panel and
select Logout. When the Login panel pops up, click on Shutdown and then one of the three
options: Shutdown, Shutdown and restart or Restart X Server. Shutdown shuts down the
Preparing Today for Linux Tomorrow
28
operating system and turns the system power off; Shutdown and restart does a restart, and
Restart X Server restarts only the X Server process without restarting the system.
From the Gnome desktop, click on the Gnome icon (the large stylized G that looks like a funny
footprint) on the Panel and select Log out. When the Log Out Confirmation panel pops up,
select one of the three choices: Logout, Halt, or Reboot and press the Yes pushbutton. Logout
logs you out of your account and brings up the Login panel; Halt logs you out and turns off the
system; Reboot logs you out and restarts the system. (Halt and Reboot will require your
account password to proceed.) Pressing the No pushbutton will return you to your Gnome
session.
From a command line, use one of the following commands: shutdown –r now (to reboot
immediately), or shutdown –h now (to halt, or stop, the system). To set an interval before
shutting down, replace “now” with “+number”, such as: shutdown –r +5 (for a five minute
delay). You will be prompted for your account password before the system shuts down.
Q. Is there a faster way to “warm-boot” the system than using
Shutdown?
Yes. As with Windows, you can use the Ctrl-Alt-Del key combination to restart the system,
but first you must open a full-screen virtual console session. A simple way to do this is by
pressing Ctrl-Alt-F1. Once the text mode screen appears you can press Ctrl-Alt-Del. You will
not be asked whether you are sure; reboot is immediate. Be sure that you do not have any
files open before doing this, as you can lose data otherwise. If, after opening a virtual console
you change your mind, press Ctrl-Alt-F7 to close the console and return to the graphical
desktop.
Q. Do I have to use Shutdown if I am just turning over the system to
another user?
No. In this situation, the preference would be to simply use Logout, allowing the next user to
login without the time and effort of rebooting the system.
Q. I clicked on Logout and now the session is locked up. How do I log
out?
Just press Ctrl-Alt-Backspace to kill the session and return to the Login screen.
Q. How can I enable Linux to automatically restart applications that are
running when I use Shutdown or Logout?
On both the KDE Logout panel and the Gnome Log Out Confirmation panel there is an option
to Save current setup. Make sure these options are selected before logging out of these
environments.
Files and Directories
The following questions relate to working with files and directories.
Q. What is the root directory?
The roots of a tree form the foundation for the entire tree. The trunk of the tree has a few large
branches; these branches lead to smaller branches, which in turn have smaller and smaller
branches that eventually lead to leaves.
Preparing Today for Linux Tomorrow
29
Using this analogy, the root directory is the base (roots and trunk) from which all directories
(main branches) and subdirectories (smaller ones) branch-off. Inside these directories are
data files and programs (the leaves). This sort of arrangement (used by UNIX/Linux,
DOS/Windows, OS/2 and other operating systems) is often referred to as “tree-structured.”
Q. Can you explain more how the directory structure of Linux works?
The main difference between how this works in Linux versus Windows, is that Windows uses
backslashes (\) instead of slashes (/) in the directory pathnames. To illustrate, here is the
hierarchy of a series of directories branching off the root:
/
/home
/home/mark
/home/mark/.netscape
/home/mark/.netscape/cache
/home/mark/StarOffice51
/home/mark/StarOffice51/database
/home/mark/StarOffice51/gallery
/home/mark/StarOffice51/gallery/clipart
/home/susan
/home/susan/.netscape
/home/susan/.netscape/cache
/usr
/ust/etc
To show it another way, this is how it looks lining up those same directories by equivalent
levels of “depth”:
/
/home
/mark
/.netscape
/cache
/StarOffice51
/database
/gallery
/clipart
/susan
/.netscape
/cache
/usr
/etc
The / indicates the root, or base, directory. /home is a directory off the root and /home/mark
and /home/susan are directories branching off from /home. (These subdirectories hold
account-related files for user IDs “mark” and “susan”.) Within /home/mark there are
directories called .netscape and StarOffice51, which in turn have directories branching off
from them. Susan has a similar directory structure for her Netscape data files. Outside of the
/home directory structure, there is also a directory called /usr branching directly off the root.
To change analogies for a moment, if /mark (and its subdirectories) and /susan (and its
subdirectories) are each children (and grandchildren) of the /home directory, then /usr is a
sibling (or peer) directory of /home. Both are direct offshoots of the root. (This whole naming
convention is referred to as the “file namespace”.)
From the root directory, to delve deeper into directories and subdirectories, use the cd
(change directory) command. For example, cd /mark/StarOffice51 would immediately
Preparing Today for Linux Tomorrow
30
move you down the directory path from the root to the contents of the subdirectory
StarOffice51. (Using the command: ls –a at that point would show you the files and
subdirectories stored in StarOffice51.) If you then wanted to change to the cache directory
inside of StarOffice 51, you have two ways of doing so. You could type the “absolute” path of:
cd /home/mark/.netscape/cache (which says, “starting from the root, travel down these
directories”) to get there, or you could simply use the shorter “relative” path of: cd cache to
get to the same place. A relative path means, “starting from where we are now, go to the next
level directory, called cache.” (Had there also been a directory below cache, called etc, we
could get there from the StarOffice51 directory with the command: cd cache/etc).
Remember, preceding the pathname with a forward slash (/) tells Linux to start from the root
directory. Therefore a command of: cd /cache/etc would fail in the preceding example
(unless you also happen to have a cache directory branching off from the root, with its own
etc subdirectory). Why? Because it is looking for the cache directory starting from the root
rather than from the StarOffice51 directory.
Reversing our direction, if you use the cd .. command from within a directory it will step you
back up a directory level (such as from /home/mark/StarOffice51 to /home/mark). You can
repeat this command as often as necessary to return to the root directory (/). Alternatively, to
jump directly to the root, use the cd / command. (This is equivalent to cd\ or cd \ in
DOS/Windows.)
Q. There are a number of directories I don’t recognize in the root
directory. What are they?
These directories store essential system files and programs. For example, the /etc (et cetera)
directory contains configuration files, where system and application settings are stored.
Beneath that, the /etc/skel directory contains shell, or “skeleton” files used to create user
accounts. The /usr (user) directory stores other important files, including program libraries
(shareable system program modules) in /usr/lib, manual pages for the man utility in /usr/man
and other documentation in the /usr/doc and /usr/info directories. Other such system
directories include /lib (libraries: security files, system modules and libraries containing
software libraries and informational databases), /opt (optional: pre-loaded software packages
are often stored here) and /var (variables: spooling, logs and all messaging files). Unless you
have a specific reason, and know what you are doing, it is a good idea to leave those
directories alone.
Q. Is there a command line shortcut to the login directory?
Yes. From any shell prompt simply type: cd and press the Enter key.
Q. Some filenames are preceded by a dot. What does this mean?
A filename that starts with a dot—such as .bashrc or .netscape—is a hidden file. They are
often used to hold user preference settings, and other configuration information. These files
are hidden to make it less likely that they will be accidentally deleted. Hiding files also makes a
file listing shorter and less cluttered with files the user will rarely be interested in. There are
different ways to see these files. For example, the ls (list) command with no parameters won’t
display these files, but ls –a (list all) will.
On the other hand, if you are using the K File Manager (KFM) under KDE, simply select View,
then Show Hidden Files from the pull-down menu to see these files. (To start KFM, just click
on the Panel icon that looks like a house in front of a manila folder.)
Preparing Today for Linux Tomorrow
31
Gnome users can click on the Gnome icon (the funny G-shaped footprint) on the Panel, select
Programs, then File Manager. From the Gnome File Manager, choose Settings, then
Preferences; finally, click on Show hidden files.
Q. When I look at a list of files and directories, the names are followed by
something like -rw-rw-r-- or lrwxrwxrwx. What does this mean?
These cryptic notations actually describe the “permissions,” or access levels, assigned to
three groups: the file owner, groups and other users. The first position identifies what type of
object the item is. Possible types include a physical (actual) Directory/folder (d), a physical file
(-), a mere Link/alias (pointer) to a file or folder (l)—similar to a Windows shortcut, a Socket
(s), a pipe (=), or a device (b is for “Block” devices, such as disk drives, that handle data a
block at a time; and c is for “Character” devices, like serial and parallel ports, that deal with
data one byte, or character, at a time). The remaining nine positions are composed of three
groups of three characters each.
The first three apply to the file owner (usually its creator). The next trio describes the
permissions granted to any authorized groups of users and the final three characters denote
any access available to unspecified other users. Each triad of positions reveals whether the
user has Read (r), Write (w)—i.e., create/modify/delete, or Execute/run (x) permission, or no
permission to perform one of those actions (-).
Using the first example from the question, -rw-rw-r-- tells us that the object is a file (because
the first character is -). The first rw- indicates that the owner has read/write access to the file,
but not execute (so it can’t be a program, otherwise the owner would be able to run it).
Likewise, any groups that have been granted access to the file can Read and Write to it (the
second rw-). However, any “other” users (those not expressly added to the authorized group)
have only Read access to the file, as evidenced by the r-- at the end. If the permissions were
drw-rw-r-- instead, the only difference would be that the object is a folder/directory rather than
a file.
In the other example from the question, lrwxrwxrwx means that the object is a link to a file or
folder and all three groups have read, write and execute permission to access that object.
Execute permission means that if it is a program, script or other executable file, an authorized
user (in this case, anyone) can run it, and if it is a folder then the user can read the contents.
Q. The ls command doesn’t show everything in the directory. What’s
wrong?
The ls (lowercase L, not capital i) utility has many parameters to let you view the directory
contents in different ways. For example, ls –a shows all files and ls –al shows all files in
“long” form (with additional details), while ls –a --color adds a splash of color to
differentiate the various types of files. (For the list of parameters, use: ls --help or type:
man ls for a full description.)
Q. Can I use wildcard searches with the ls command, as I can with Dir in
DOS/Windows?
Sure. You can search for all files ending with the .txt extension using: ls *.txt, or all files
starting with “mc” and with any extension, using: ls mc*.*. If you are looking for a file
among a long list of files with similar names, you might limit the wildcard to just one character,
using the ? symbol. For example: ls hello?5.c would find files called hello05.c, hello95.c,
helloj5.c and so on, but not hello105.c or hellor7q5.c. You can even combine wildcard
characters for very specific searches, such as: ls hello?5.*. (Wildcards can also be used
with other commands, including cp (copy), mv (move), rm (remove), etc.)
Preparing Today for Linux Tomorrow
32
Q. I want to do a wildcard search on files with ? or * in the names. How
do I do that?
Unlike DOS/Windows, UNIX/Linux does allow those characters, as well as ) and ( and | and
other special characters in file and directory names. To distinguish between a wildcard search
using ? or * and a search for those particular characters, you merely need to insert a
backslash (\) before the wildcard character. For example: ls hello\?.txt or ls
myfile\*.c or ls mc\)127.h would find the files hello?.txt, myfile*.c and mc)127.h,
respectively. You can also combine these “regular expressions” with wildcards in searches,
such as ls hello\?.t?t (which would list hello?.txt and hello?.tst), or ls mc\)127.*
(which would list mc)127.h and mc)127.c, for example).
Q. How do I create or rename a file with special characters in the name?
As with the previous question, you merely need to insert a backslash before the special
character. For example: touch hello\?.txt would produce an empty file called hello?.txt;
touch myfile\*.c would result in myfile*.c; and touch mc\)127.h would create
mc)127.h. Similarly, mv hello.txt hello\?.txt would add a ? to the filename,
immediately before the dot. To use multiple special characters, simply insert a \ before each
one; for instance: touch myfile\(\*\*\).txt would produce myfile(**).txt. The
procedure is the same for any other special character. If you are not sure which characters
require the \ treatment, simply try the command without it first. If you receive a “sh: syntax
error near unexpected token ‘yourfilename.’” message (where ‘yourfilename’ is the name you
tried to create), then you need the backslash.
Important Note: Although you can include many special characters in this manner, using
some special characters, including the asterisk (*), question mark (?), hyphen (-), vertical
bar/piping symbol (|), circumflex or caret (^), apostrophe (‘) and ampersand (&, in a file name
is generally not a good idea, because they may be confused with wildcard, parameter
(“argument”) and other symbols by various commands. In addition, using a space in a
filename (such as my file.txt) may cause some commands to treat the name as two
filenames, or a parameter and a filename. (To avoid this problem, enclose the name in quotes,
as in “my file.txt”.) Any of these conditions will cause unintended results. At the very least,
these special symbols can make filenames more difficult to read. Unless you have a good
reason for doing so, it is probably best not to use them.
Q. How long can a filename be?
A filename can be up to 255 characters long, including numbers, letters and special symbols,
as well as a combination of upper and lower case (which are treated as different characters by
Linux/Unix). However, for the sake of readability and to make it easier to type from the
command line, much shorter names are preferable. This 255-character limit does not include
the full path. In other words, if a file called ThisIsMyVeryLongFilename.txt is in the
/user/mark/garbage/files directory, the filename alone can be up to 255 characters,
excluding the characters in the pathname. To illustrate, out of the following path:
/user/mark/garbage/files/ThisIsMyVeryLongFilename.txt
only the 28 characters (bolded) following the final / comprise the actual filename (including the
dot before “txt”).
Q. The directory listing is too long and scrolls off the page. Is there a
way to view a page at a time?
Preparing Today for Linux Tomorrow
33
Yes. There are programs, called “pagers” to do this. Just as in DOS/Windows, where you
could use: dir c:\ | more to list the contents of the root directory, in Linux you can type:
ls –a / | more
Or, using an alternative utility:
ls –a / | less
Using less, press the Spacebar to move ahead a page at a time; to back up, use the B key.
When you are done, press Q to quit.
Q. Is there a way to find files from a command shell?
Just as Windows has a Find feature, Linux has the locate command. It will find data files,
programs, directories and other objects whose names match your search argument. For
example, locate myfile will find myfile.txt, /myfile and other matches. (For more on
locate, including search parameters, use: man locate.) Locate searches a database (called
slocate.db) on your hard drive used for this purpose, which is updated overnight. If you are
searching for a new file that locate is unable to find (or you regularly shut the system down
overnight, so the database doesn’t get updated often), you can force a manual database
update. To do this, use the: su command to temporarily login as root, then from the shell
prompt type: updatedb. Wait a few minutes for the database update to complete and try the
locate command again.
Note: Linux does also have a find command, but for performance reasons locate is better for
searching on filenames. If you have a need to search by file date or other attributes or would
like to automatically execute a program using the found file as a parameter, or any of a large
number of other interesting uses, find is the command to use. For more on find, use: info
find.
As an alternative, if you are merely looking for the binary, source and manual pages for a
command, you might try the whereis command. To illustrate the difference between whereis
and locate try the following two commands: whereis passwd and locate passwd. As you
can see, the whereis command found only a few files, while the locate command listed
dozens. That is because whereis was looking for a few specific files (all called passwd.*),
while locate was busy seeking any file or directory name containing “passwd” in the name
(including gpasswd.1.gz, autopasswd, passwd-1.html, etc.) For another example, try:
whereis *.doc and locate *.doc. As before the locate command came up with many
files sporting a .doc extension. The whereis command, however, found nothing. Why?
Because there is no Linux command called “*.doc” or even “doc”. Remember, whereis works
only on files associated with commands, not just any old data file or program.
Q. Is there a way to browse through a text file from the command line?
Yes. Perhaps the best way is to open the file with a text editor, such as Pico, Emacs, Vi or
others, and scroll or page through the file. However, if you are trying to do the browsing from
the command line without opening the file first, you can use the head and tail utilities, or the
more or less commands. For example, head myfile.txt would display the first 10 lines of
the file onscreen, by default. The command: head –40 myfile.txt would show the first 40
lines. (You can substitute larger or smaller numbers for “40”.) Similarly, tail displays the last
10 lines of a file, as in: tail myfile.txt. (If the file is short, but you are not sure how many
lines long, you could use the cat command (short for concatenate) to display the entire file on
the screen, such as: cat myfile.txt.)
Preparing Today for Linux Tomorrow
34
Obviously, this is fine if you know approximately how close to the beginning or end of a file the
text string is, but not if it is somewhere in the middle of a large file. In that situation more or
less would be more productive. (See the question, The directory listing is too long and scrolls
off the page. Is there a way to view a page at a time? in this section, for information on those
commands.)
Q. Is there a command to search text files for a specific character string?
Yes. Naturally, you can open a file with a text editor and use its search function. However, if
you prefer to search from the command line, you can use the grep utility. It is a very powerful
text file search tool. You can perform wildcard searches on a specific file, but more
importantly, you can search many files at once, also with wildcards. For example, say you
have a directory full of text files and you know that one of them contains the phrase, “Mary had
a little lamb”, but you do not know which file it is. Using the command: grep “Mary had a
little lamb” *.txt might result in the following output:
myfile.txt: Mary had a little lamb. Its fleece was white as
snow. And everywhere that Mary
(Note the use of quotation marks around the phrase to identify the search string when multiple
words are involved.) The grep command shows the entire line of text containing the search
string to make it easier for you to determine whether it is the correct text (because there may
be multiple matches).
Note: The grep utility is by default case-sensitive, so searching for “Mary” would not find mary
or MARY. If you wish to perform a case-insensitive search, add the –i parameter in the
preceding example. Then it will find Mary, mary, MARY, mARy, or any other mixture of case.
For example, a case-insensitive wildcard search such as: grep –i mar* *.txt might yield
this result instead:
myfile.txt: Mary had a little lamb. Its fleece was white as
snow. And everywhere that Mary
personnel.txt: Candidates for a pay increase this year include
Johnson, Martin and Littlefield.
personnel.txt: Due to Smith’s impending marriage, she will be
taking a two-week honeymoon in
Not only were there “hits” in two different files, but there were two separate occurrences of
“mar” in the personnel.txt file as well.
The results of a search can even be copied to another text file or printed, using the piping
symbol (|). (See the question, The directory listing is too long and scrolls off the page. Is there
a way to view a page at a time? in this section, for information on piping.) Search results can
also be redirected to a file using the redirection and append symbols (> and >>). (See the
question, Is there online documentation for most programs? in the General Usage section, for
information on redirection.)
There is much more that can be done with the grep command. For example, the –c parameter
tells grep to display only the count of the number of times a string appears in a file, rather than
the text of the matches (such as personnel.txt:2). For more information, read the manual
page (command: man grep).
Preparing Today for Linux Tomorrow
35
Q. What are “symbolic” and hard links, and how can I identify them?
Symbolic (soft) links work much like “shortcuts” in Windows, as pointers to programs, data
files or directories located elsewhere. Links allow you to start a program from more than one
folder (or the Desktop) without having to duplicate the files in multiple places. Links are
independent files—only a few bytes in size—that can be renamed or moved without affecting
the original programs or files they point to. Conversely, if the original file is moved, deleted or
renamed it will break the link, because it no longer matches the filename and location pointer
stored in the link. (If a link is broken this way, it should be deleted—as it is no longer useful—
and a new link created, if needed, for the new filename and/or location. On the other hand, if
you rename the original file back to what the link is expecting, or return it to its original
location, the link will “magically” work again.)
Symbolic links can be identified, using the K File Manager (kfm) or other graphical file
manager, by a small arrow in the lower right corner of an icon, pointing at the icon. From the
command line, one command you can use is: ls –F (capital F, not lowercase). When the
filenames are listed, the symbolic links will be followed by an @ symbol. For example, if you
were to list the files in the /bin directory, you might see:
arch*
ash*
ash.static*
aumix-minimal*
awk@
basename*
bash*
bash2*
bsh@
and so on, where awk@ and bsh@ are the symbolic links to files in other directories. (The *
following the other names indicates that they are “original” files.) Alternatively, you could use
the: ls -il command (see below).
Link files specific to the KDE user interface will have .kdelnk filename extensions, such as
Autorun.kdelnk.
Another type of link is the “hard” link. This is similar in function to a symbolic, or soft, link
except that renaming, deleting or moving the original file won’t break the link (because it
actually links to the physical file on disk, not just the file pathname in the disk directory
structure). Hard links apply only to data files or programs, though, not to directories.
All files have at least one hard link (the actual filename in the directory listing). Adding another
hard link increments the hard link count, which you can display using the: ls –il command.
For example:
619880 -rw-rw-rw- 2 ownerid groupid 428 Apr 14 19:42 myfile.txt
145389 -rw-rw-r-- 1 ownerid groupid 7285 Sep 30 10:17 coffee.txt
268394 -rw-rw-rw- 1 ownerid groupid 8912 Nov 4 19:42 tea.txt
619880 -rw-rw-rw- 2 ownerid groupid 428 Apr 14 19:42 person.txt
(The –i parameter says to display the “inode” number—explained below, and the l argument
tells Linux to show the directory list in long form, including the permissions, owner and group
Ids, file sizes, and creation date and times.) There are four different filenames listed, but
notice some key similarities between two of them: the files myfile.txt and person.txt have
exactly the same date, time, file size and properties. In addition, they both show “2” for the
number of hard links, but all of these similarities could just be a coincidence (after all, two
unrelated files of the same size and type could have been created within a minute of one
Preparing Today for Linux Tomorrow
36
another). The real clincher, however is in the set of numbers preceding the file permissions
listings. (For an explanation of file/directory permissions, see the question, How do I display
the permissions for a file or directory?) The “619880” number is the inode, or physical disk
address of the file. Only one file can occupy a given block of disk storage (just as only one car
can occupy a particular parking space at any one time). Therefore, these “two” files must in
fact be only one physical file with two hard links (filenames) to the file.
Again, unlike soft (symbolic) links, deleting either of these two files doesn’t affect the other.
Another difference between the two types of links is that hard links can’t span disk drives. In
other words, a hard link can’t be created on /dev/hdb to link to a file on /dev/hda.
Q. How do I create links (symbolic or hard)?
The ln (link) command can be used to create both types of links. For example, to create a
hard link called myfile2.txt to a file called myfile.txt, type: ln myfile.txt myfile2.txt.
This will create the link in the same directory as the original file. To create a symbolic link
called myfile3.txt to the same original file, use: ln –s myfile.txt myfile3.txt (the –s
parameter is what tells Linux to make the link symbolic).
If you use the: ls -il command to list the directory contents, you should see something like
the following:
619880 -rw-rw-rw- 2 ownerid groupid 428 Apr 14 19:42 myfile.txt
619880 -rw-rw-rw- 2 ownerid groupid 428 Oct 2 10:18 myfile2.txt
224516 lrwxrwxrwx 1 ownerid groupid 6 Oct 2 10:18 myfile3.txt -> myfile
Note that all details of the hard link (myfile2.txt), including the inode address (619880), are
identical to that of the original file (myfile.txt) except for the creation date and time. It is
treated as a physical file (as verified by the file size and the permission type of -), but at the
same location as the original file (same inode address) with the same read/write permissions.
The 2 following the permissions reflects the number of hard links to the physical file (the
directory listing for the original file being the first hard link). Conversely, virtually every aspect
of the symbolic link (myfile2.txt) is different. The inode address differs because there is a
physical file created elsewhere on the disk that contains a pointer to the original file. Likewise
the permissions differ, with the symbolic link showing a file type of link (the l leading off the
permissions) and including execute permissions for all user types. The 1 following the
permissions demonstrates that this is not a hard link. Finally, the filename is followed by “->
myfile”, which reveals the name of the original file to which it points.
Note that creating a hard link for this file in the same directory probably would not be the
normal usage of a hard link. More likely, you would create it in another directory so that you
could access the file from either place (or several places, if you make multiple links). For
example: ln myfile.txt /user/mark/myfile2.txt would put the new link in the
/user/mark directory.
Linux itself makes use of hard links. In many cases Linux continues to use programs and files
that were first used in earlier versions of Linux, which can sometimes cause problems. For
example, if the X Window System was first provided in a directory called /usr/X11R4, Linux
software which used X Windows would look for those files in that directory. But when a newer
version of Linux comes out that uses a later version of X Windows, the files might be stored in
/usr/X11R5 or /usr/X11R7 or /usr/X11R10. So how to keep from breaking all those programs
that are expecting the files to be stored in /usr/X11R4? Simple. Put hard links from one
directory into the other. This way, whichever directory the application software looks in for
those files, it will find what it needs.
Preparing Today for Linux Tomorrow
37
As an alternative to the command line interface, some graphical file managers (such as GNU
Midnight Commander), offer options to create either or both types of links by right-clicking on a
file icon.
Q. Can you summarize the differences between symbolic and hard links
and differentiate between using links and simply duplicating a file?
Sure. If you merely copy a file from one directory to another, you have doubled the disk space
used by the file, and editing one file doesn’t affect the other (because they are in no way
connected to one another). This is fine if it is what you intended (so that you can keep a
pristine original while modifying another copy, or if you made a backup copy for safekeeping.)
If you copy a directory instead of a file, you also copy the entire directory’s contents, which
can result in a lot of disk space used. A file or directory copy can be put anywhere on any disk,
either directly attached to the system or on a network or Internet/Intranet drive somewhere
else entirely. Because the original and the copy are unconnected, deleting one has no effect
on the other.
On the other hand, if you create a symbolic link it takes only a few bytes of space and can be
located on any mounted device that you can type in a path for. If you delete, rename or move
the original, you break the link (although recreating the original in the same place will restore
the link). Using the link to access the file or directory, any changes you make to the file or to
the directory contents will affect the original files/directories, because those are what you are
actually touching.
Alternatively, if you create a hard link instead, no additional disk space is used, however the
link can only be created on the same physical hard disk (but multiple partitions on that disk),
not on other local or remote devices). Moving or renaming the original file won’t hurt the link; it
will reflect the change. When you “delete” the original file, you are actually deleting only the
original hard link (the directory name). So, as long as at least one other hard link exists, the
file is still accessible. Hard links can be created for files but not for directories. Because the
permissions indicate that a hard link is a physical file, any software that tries to access these
files will think it is accessing the original files (which, in fact, it is—through the link).
Q. I recognize a number of file types, such as .JPG, .GIF, .WAV, .TXT,
.HTM and .ZIP from Windows, but I see many other file types I don’t
know. What are they for?
There are a number of file extensions that are common in Linux/UNIX but not typically seen in
the Windows world. Here are some of them:
.a – Archive file .png – Image file (like .jpg/.gif)
.au – Audio file (like .wav) .ps – PostScript printer-formatted file
.bzip2, .bz2 – Archive file compressed with
bzip2
.rpm – RPM Package Manager file
.c – C programming language source file .so – Program library file
.conf – Configuration file .tar – Archive file
.cpp – C++ language source file .tar.bzip2, .tar.bz2 – Archive file
compressed with bzip2
.gz, .gzip – File compressed with gzip .tar.gz, .tar.Z – Archive file compressed with
gzip
.h – C or C++ language header file .tga – Image file
.html – HTML file (same as .htm) .tgz – Archive file compressed with gzip
.o – Program object file .xpm – Image file
.pl – Perl script file .z, .Z – Older gzip file
Preparing Today for Linux Tomorrow
38
If you encounter a file type you do not recognize (such as one with no file extension), you can
use the file command to find out more about it; for example: file smith.xyz or file
smith (if there is no file extension).
Q. Should I back up my entire system?
That is not as easy a question to answer as it may sound. A fully configured system with
operating system, many large applications, data files and perhaps a library of multimedia files
(music, video, photographs, etc.) can amount to many gigabytes of information. Unless you
have a high-capacity, high-speed backup option (such as a 4mm, 8mm or DLT tape drive, or
an Iomega Jaz removable disk cartridge, for example), it just is not practical to back up
everything. On the other hand, because you probably installed all your software from a few
CDs, it is not absolutely necessary to back up the operating system and applications. If you
back up only your data files, setup files, account files and others that have been created or
modified since installation, you can always reinstall the software from CD, then restore the
backed up user files.
Of course, the disadvantage to this, in the event of catastrophic disk failure, is the time and
effort needed to install all that software individually, rather than just starting a restore program,
inserting a tape or Jaz cartridge and walking away. But if you do not have a high-capacity
backup solution at your fingertips, a partial backup is the most effective method for all but a
complete system rebuild. Each application will have its own set of configuration files that
should be backed up, as well as data files, but you will have to find out what they are from the
software documentation or the vendors.
From a Linux standpoint, the directories you should be sure to back up include:
• /etc
• /home
• /opt
• /usr/lib
• /usr/local
• /usr/src/linux (if you have modified any of the kernel programming)
• /usr/X11R6/lib/X11 (if you are using X Windows)—Case is important in this folder name!
• /var/spool/mail
Of course, you should also back up any directories containing important user data files. If you
lack large-capacity storage, you can always back up each application’s files to its own set of
floppies using an archiving/compression utility, such as tar, gzip, bzip2 or zip. These and
other tools can span more than one disk or tape to store all the data in one related archive.
This method may not be as quick or efficient as a tape drive, but at least you will have a copy
of your essential files. It is not necessary to back up the program files themselves, as you can
always just reinstall the applications from scratch, but you might want to back up the files that
contain the user setup information (preferences), such as default margins and tabs for a word
processor. (One clue as to which files these are is to look at the file dates. Generally, the user
setup files will have much newer dates than the files created when the program was first
installed. If in doubt about what to back up, contact the software vendor for the filenames.)
Q. How often should I back up my files?
There is no one right answer to this question either. Every user is different. One way to
answer the question for yourself is to use the “ouch” test. Consider how much work you could
Preparing Today for Linux Tomorrow
39
afford to lose before it hurts. If you are a home user, perhaps a once-a-week backup would be
adequate (right after you pay bills, for example). In a high-performance office environment,
once or twice a day—even hourly—may be required.
It is also a good idea to use a “grandparent-parent-child” strategy of rotating backup media. (In
other words, use three or more sets of tapes or cartridges and alternate among them.) This
way, if for some reason your most recent backup is corrupted, at least you have the next older
one to fall back on. Plus, if you discover that you accidentally deleted a file a few days ago—or
it was damaged by a virus—you can restore the file from a backup made before the deletion.
Another option is to back up only what has changed since the last backup (some archival
programs track this information from backup to backup). The advantage is that often very little
has changed since the last time, making for very quick backups. The disadvantage is that you
must keep copies of all the intervening backups since the last complete archive. Losing even
one partial backup along the way can result in the loss of important information. Thus, even
with a policy of incremental backups, it is a good idea to do full backups periodically.
Q. What backup software should I use?
There are a large number of commercial and free programs available for use, some primarily
for tape drives and others for more general usage (magnetic and optical disks, network drives,
etc). Many of these have graphical interfaces to make them easier to use. It is a good idea to
try out a few different programs to see what you like best.
There are too many available tools to try to describe them all here. But there are a few
common commands and utilities—likely to ship with your Linux distribution—that you can start
with:
tar — tar was the earliest and most common “packaging” tool for Unix/Linux (it does the
unpacking of files as well—there is no “untar” command). The tar command creates a file that
contains all the other files and directories, optionally in a compressed state (using gzip or
bzip2), to save space. If you have used zip/unzip or PKzip with DOS/Windows, then you are
already familiar with the concept. Files packaged with tar typically end with a .tar extension. If
also compressed with gzip, the extension may be .tar.gz, .tar.Z or .tgz. Tar files compressed
with bzip2 generally have a .tar.bz2, .tar.bz, .tbz2 or .tbz extension. The gzip and bzip2
compression utilities can also be used independently of the tar command to compress and
decompress files (see below). There are many different parameters available for the tar
command (which does the unpacking of files as well), so it is best to read the manual page for
tar using the command: info tar. Note: It is not advisable to compress tar files on a tape
backup, because tar would not be able to recover from tape errors.
gzip/gunzip — gzip (GNU zip) is the tool traditionally used for compressing tarred files and its
companion gunzip is used to decompress those files. More recent tools, such as PKunzip,
can generally uncompress files created by gzip as well as those compressed by other tools,
such as PKzip and zip. Due to this flexibility and because PKzip and zip are also available on
other operating system platforms, they have become more popular than gzip. Files
compressed with gzip usually end with .gz, or in the case of some older files, .z or .Z. You can
read more about the gzip/gunzip commands using: info gzip.
bzip2/bunzip2 — Think of bzip2 as a newer, faster version of gzip. It has a very similar
command structure (including many of the same parameters), but it tends to be faster at
compressing and decompressing than gzip/gunzip. Bzipped files end with a .bz2 or .bz
extension. Learn more about bzip2/bunzip2 from: info bzip2.
PKzip/PKunzip — PKzip is a popular shareware/commercial packaging and compression
tool, primarily in the DOS/Windows world. PKzip is very fast and offers varying degrees of
Preparing Today for Linux Tomorrow
40
compression vs. speed. Newer versions include a graphical front user interface. PKunzip is
compatible with files created by PKzip, zip, gzip and tar, as well as the UUencode,
XXencode, MIME and BinHex file formats. PKzip files end with a .zip extension. For more
information, read the documentation that comes with PKzip.
zip/unzip — zip and unzip are a pair of command-line tools that will package and compress
(and decompress) files and directories. In addition to being very fast, they are free and
available for other platforms, including DOS/Windows, OS/2 and other UNIX OSes, including
SCO. The unzip command can also unpack files created by PKzip. Like PKzip, zipcompressed
files end with a .zip extension. The zip and unzip utilities may be included with
your Linux distribution; if so, read their manuals using the following commands: info zip
and info unzip.
Other utilities you can obtain separately include Arkeia, Backup Edge, BRU2000, Catchup, CTAR,
Drive Image, Kbackup, KDat, PerfectBACKUP+, Sitback, Taper and many more.
For a lot more on this topic, see the related white paper entitled, Installing Linux Applications
for the First Time, available from the same sources as this paper.
Root/User Accounts, Groups and Permissions
These are questions revolving around administering accounts, passwords and permissions.
Q. Sometimes instructions say that I must “be root.” What does this
mean?
It means that you must be logged in as the root user, or superuser. (Sorry, but you do not get
a cape with a big S on it!) There are two ways to do this. If you will be working as root
extensively, you should logout as yourself and login as root. When you are done, reverse the
process and log back in as yourself.
On the other hand, if you need to be root only long enough to execute a command, you can
open a command shell and issue the command: su (substitute user) to temporarily login as
another user, while still inside “your” login shell. You will be prompted for the root password.
Once that is provided you will “be root” and can issue commands as needed until you end the
root session by typing: exit and pressing Enter or by closing the shell.
Some commands won’t work unless you have logged in as root (i.e., su won’t work, because
Linux “knows” it is still you in charge). An alternative is to use the su – command (su followed
by a hyphen, with a space separating them). This makes you root with the root’s login shell, as
if you had logged off and logged back in as root.
Q. What is the root account and how does it differ from a user account?
When you installed Linux, you were asked to provide a root password. Later you were
prompted for a user account name and password. The root account is equivalent to the
system-level access you are given under Windows, with the ability to delete, create and format
partitions, create/delete directories and files, or do anything else your heart desires. Linux
operates like a network client/server system. The root account is like a system administrator,
able to control what access authority the various clients (users) have. There can be many
different user accounts set up on the same Linux system, each with its own set of
permissions, configuration options, choice of user interface and so on.
This is very useful in a business environment where multiple people share the same system,
as each user account can be tailored for each individual user. On the other hand, it can be a
Preparing Today for Linux Tomorrow
41
nuisance when there is only one user, because there are some system functions that can be
performed only by the root account, necessitating that the user log out of the user account and
log in to the root account long enough to perform the needed operation, then log out of the
root account and back into the user account. (Fortunately there is a shortcut you can use in
some situations as an alternative. See the description of the su - command in the earlier
question Sometimes instructions say that I must ‘be root’. What does this mean?.)
As compensation, however, doing things this way makes it much less likely that the user will
accidentally delete or corrupt important system files that they have no business playing with in
the first place.
Important Note: Most, if not all, Linux distributions require that you create a user account in
addition to the root account when you install Linux. While it is tempting to log on as root and
use the root account for everything, this is a Very Bad Idea. There are too many ways in which
you can accidentally (or someone else can maliciously) corrupt up a system from the root
account. By running mostly from the user account, your system is protected to a great degree
from most disasters. (It won’t prevent a disk crash, of course, but at least you can’t
accidentally delete your /usr/lib directory, for example.) If the Linux distribution you installed
required you to create a user account, use it. If you did not create one during installation,
create one now—and use it!
Q. Why would I want to add other user accounts to my system?
If you are the only one who uses your system, there may be no need to set up other user
accounts (besides your own). On the other hand, if you share your computer with other people
(second and third shift workers, perhaps; or you share a common work area; or other
members of your family use your computer), or even if you wish to make your printer or other
devices available to others on the same network, you will need to set up individual user
accounts and possibly group accounts.
Q. How do I add users?
There are a number of tools available for doing this. For instance, Red Hat Linux 6.2 includes
the graphical Red Hat Disk Management tool (linuxconf). For occasions where you are not
running a graphical interface, or prefer a command line tool, a couple of text-mode programs
that are often included with Linux are adduser and useradd. Despite the similarity of their
names and purpose, adduser is the more functional of the two.
To use linuxconf:
1. Run: linuxconf from a command line under either KDE or Gnome. (Linuxconf is not
located on a menu anywhere.)
In order to use Linuxconf and similar system-level tools, you will need to first be logged in
as root. To do this, use the su – command and enter the root password when prompted.
2. Once the linuxconf panel opens, scroll down to the heading called Users accounts.
Immediately below, there is a subheading called Normal, and beneath it User accounts.
3. Double-click on User accounts to open a panel that lists all existing accounts.
4. To create a new account, click on the Add button at the bottom of the panel. You will see
a subpanel with four tabs: Base info, Params, Mail settings and Privileges.
5. When you are done filling in the necessary information, press the Accept button to save
your changes (or Cancel to reject the changes and not create a new account).S
Preparing Today for Linux Tomorrow
42
Users of other distributions besides Red Hat have other graphical user administration tools at
their disposal. For instance, Caldera OpenLinux 2.3 users have Webmin and COAS.
Although neither individually can do everything linuxconf can do, together they offer similar
functionality.
To add users with the COAS tool:
1. From the KDE menu, select Settings, then COAS, then System and finally Accounts. A
panel will appear.
2. On the tool bar, select User and then Create user.
3. When the menu appears, begin customizing as many users as you need.
If your distribution doesn’t offer a graphical tool, or if you prefer to use the command line
interface, you can use adduser. As with linuxconf, before using adduser you must login as
root. Then type: adduser newusername (where newusername is the user name you wish to
create).
The program will assign the next available ID for both the user and default group, then remind
you to set a password for this new account. (To add the password, see below.) This will
supply the minimum information necessary to create a user. To add personal information to
the account record, such as the user’s name, address and phone number, use the chfn
(Change Finger Name) command (see below).
Q. How do I delete users?
This is a multistep process:
1. Log in as root.
2. Use the userdel command with the name of the user you wish to delete, as in: userdel
mary. Note, however, that while this deletes the user account name from the /etc/passwd
file, it doesn’t delete all associated files.
3. Follow up with the groupdel command to remove the user’s UPG entry (the default
group), such as: groupdel groupname.
4. Remove any other group affiliations for the user (besides the UPG), using: usermod –G
username username. (Note that the –G parameter is capitalized.)
5. Use a text editor to remove the login ID from the /etc/group file.
6. Delete the user’s home directory and mailbox using: rm –R /home/username
/usr/spool/mail/username. (Note that –R is capitalized.)
Q. Is there any way to change a user ID?
Yes. To change a user ID (login name), use the usermod command. For example, to change
an account name from “mary” to “Mary_Jones”, use: usermod –l Mary_Jones mary. If
you change a login name, you should also change the user home directory, where all the user
account information is stored. To do this, use the command: usermod –d
/home/Mary_Jones –m mary to move the user account information to the directory
/home/Mary_Jones from /home/mary. (The –l parameter tells usermod that you are
changing the login name; –d identifies the directory you are changing; and –m means to move
the contents to the specified directory.)
Preparing Today for Linux Tomorrow
43
Q. Is there a way to disable a user account?
Yes. If you are using “shadow” passwords (see the question What is a shadow password?, in
this section, for an explanation), the simplest way to disable an account is via the usermod
command. Merely assign an expiration date that has already passed, such as: usermod –e
01/01/92 mary. (See below for an explanation of shadow passwords.)
As an alternative, experienced users can edit the /etc/passwd file. To disable the account,
find the line for the appropriate user and insert an asterisk (*) in front of the encrypted
password. The password is located between the first and second colon separators, following
the user name. For example:
mary:*Ms.2jbeTWmUYi:500:Mary Jones:/home/mary:/bin/bash
Warning: This alternative approach should be used only as a last resort, by a proficient Linux
user, because manually editing this file can cause major problems if not done correctly!
Q. How do I add or change personal information in an account?
If you wish to add name, address and telephone number information to your account or
change information already there, there is the chfn (Change Finger Name) command. You will
be prompted for your password, then shown the data fields for the information above.
To change other user accounts, login as root. Then specify which user you wish to change the
information for, such as: chfn fred, and make the additions or changes.
If you prefer a graphical tool, Red Hat Linux 6.2 users running KDE can click on the KDE
menu (the big K on the Panel) and select the Red Hat option, then System and finally About
Myself. This will bring up a panel with entry fields for all the information accepted by the chfn
command. I am not aware of any such graphical tools for desktop environments other than
KDE, so if you are not using KDE you will have to use the chfn command line option.
Q. How do I add groups?
Default groups are automatically created when you add users. Should you need to add groups
manually you can do so with the groupadd command. Normally, group IDs (GIDs) are
automatically supplied (by adding 1 to the last number used) when a group is created.
However, because you are creating a group manually you will need to specify a group ID
yourself. The maximum GID number is 65536, so you might want to specify something very
large (say, 50000 or higher) to keep from possibly running into one already assigned by the
system. To create a group called “mygroup” with a GID of 50000, type: groupadd –g 50000
mygroup.
Q. How do I add/delete users in a group?
To add users, use the passwd program with the –a (add) parameter. For example, to add
user “mary” to group “mygroup” type: gpasswd –a mary mygroup.
To delete users, simply replace the –a parameter with the –d (delete) parameter, such as:
gpasswd –d mary mygroup.
Q. How do I change a password?
To change your own account password, Red Hat 6.2 users running KDE can click on the KDE
menu and select Red Hat, then System and finally Change Password. This will bring up a
panel prompting you for the current password (for the logged in user). If the correct password
Preparing Today for Linux Tomorrow
44
is supplied, you will be prompted for a new password to use. (If it is too short or invalid for
some other reason, the prompt panel will reappear.) I am not aware of any graphical password
management tools for desktop environments other than KDE, so if you are not using KDE you
will have to do it from the command line.
From the command line, simply use the passwd command. It will prompt you for your current
password and then the new password twice, for verification.
To change another password, you must be logged in as root. Then tell the program which user
you wish to change the password for, such as: passwd fred. You won’t be required to know
the current password; simply enter the new password twice.
Q. What is a shadow password?
Shadow passwords are used to protect system passwords (for user accounts) by making the
file containing those passwords (/etc/shadow) readable by the root operator only. When
shadow passwords are used they replace the encrypted password in the /etc/passwd file with
asterisks. (Moving the passwords to /etc/shadow makes it less likely that the encrypted
password can be decrypted, because only the root operator has access to the file.) To see if
you have shadow passwords enabled on your system, use the command: ls /etc/shadow.
If you receive a message like: ls /etc/shadow: No such file or directory, the
files have not yet been installed. For more information on shadow passwords, read the manual
page: man 5 shadow.
Q. How do I display the permissions for a file or directory?
From a command line, use the ls command with the –l parameter. For example, ls –a –l
*.txt would display all .txt files in the current directory, with their permissions, such as:
-rw-rw-rw- 1 ownerid groupid 1428 Apr 2 19:42 myfile.txt
-rw-rw-r-- 1 ownerid groupid 914 Sep 30 10:17 personnel.txt
The first 10 characters (or “switches”) on each line contain the permissions for that particular
file, directory, folder or other object. For an explanation of what the characters mean, see the
question, When I look at a list of files and directories, the names are followed by something
like -rw-rw-r-- or lrwxrwxrwx. What does this mean?
If you are using the K File Manager (KFM) under KDE (just click on the Panel icon that looks
like a house in front of a manila folder), simply select View, then Long View, from the pulldown
menu to display all objects with their permissions. Gnome users will have to use the
command line method.
Q. How do I change permissions?
If you are the system administrator (root), or the owner of a file, directory or other object, you
have the authority to grant or remove permissions for other users and groups, using the
chmod command. The operative symbols to remember are u for User (owner), g for Group, o
for Other, a for All, r for Read, w for Write and x for eXecute. Precede the r/w/x characters
with a plus sign (+) to add permission or a minus sign(-) to deny permission.
For example, to grant Read/Write access for the file myfile.txt to Groups and Others, use the
command: chmod go+rw myfile.txt. To remove Write and eXecute access from all
accounts, use: chmod ugo-wx myfile.txt or chmod a-wx myfile.txt. (Note that a,
for All, is a convenient shortcut for ugo.)
Preparing Today for Linux Tomorrow
45
Communications
These questions relate to configuring or using communications features of Linux.
Q. How do I set up my system for dial-up Internet access?
This is a multipart process. Before you begin you should contact your Internet Service
Provider (ISP). There is some information you will need to provide to Linux that only your ISP
can provide. (A technician who is conversant with Linux can probably step you through the
entire process, but because you may not find a Linux expert at your ISP, here are the steps
necessary to do it yourself.)
• If you are using KDE, use the Kppp utility to start the configuration process.
! Log in as the root operator, then use your favorite text editor to open the
/etc/ppp/options file. If you see a statement with the option “lock” you will need to
“comment it out” to disable it. (Kppp handles locking the connection and the hardware,
so Lock is redundant.) To disable the lock option, insert a # symbol in front of it, such
as: #lock. Save the file and exit the editor. Log out as root and log back in as
yourself.
! To start Kppp, go to the KDE Application Starter menu (the big K icon on the Panel)
and select Internet, then Kppp.
! When the Kppp Configuration panel appears, press the Setup button to create a new
ISP connection record.
! From the Accounts tab, press the New pushbutton.
# When the Edit Account panel appears, you will be in the Dial tab. Enter whatever
you want to use as the ISP name where it says Connection Name. Enter the dialup
connection phone number in the next field (including area code, or anything
needed to dial out of the switchboard, such as 9,). The default authentication type
is PAP. Use that unless your ISP tells you otherwise. (The remaining fields are for
experienced users. Unless you have a reason to enter information in those fields,
leave them blank.)
# Click on the IP tab. When that panel appears either choose Dynamic IP Address,
or enter a static address. Your ISP will have to tell you which option to use, but
generally the ISP will assign a temporary IP address each time you connect, so
most likely the ISP will tell you to select Dynamic IP Address. If you are setting up
a connection inside a firewall, you will probably be using a static address. The
administrator in charge of the firewall will assign you a static address and tell you
the subnet mask to enter in the appropriate fields. Use the “Auto-configure
hostname from this IP” checkbox if you are connecting to a DHCP server or a
Bootp server. (Your network administrator will tell you if you need this.)
# If you are connecting from home, you can skip the DNA and Gateway tabs. If you
are connecting in an office environment, your network administrator will tell you if
you need to enter anything in those tabs. Unless told otherwise, you should not
need to enter anything in the Login Script tab either. If you live outside the United
States or Canada and your ISP charges by bytes transferred, use the Accounting
tab to select the country, the company and whether you are charged by incoming
data, outgoing, or both; otherwise ignore it. Click the OK button to close the Edit
Account panel and return to the Kppp Configuration panel.
! Click on the Device tab to set up your modem connection. If the modem was set up
correctly during installation, you should have to select only the port to use. COM1
through COM4 are mapped respectively to /dev/ttyS0 through /dev/ttyS3. Use the
one that corresponds to the port to which your modem is connected. An internal
Preparing Today for Linux Tomorrow
46
modem may be mounted as /dev/modem. For Flow Control, use CRTSCTS to enable
hardware flow control. For Line Termination, choose CR/LF. For a 14.4Kbps modem,
choose a connection speed of 57,600bps. For faster modems, select 115,200bps. If
you would like to prevent a second program from disconnecting one already using the
modem, check the Use Lock File box.
! In the Modem tab, you will configure the modem. Use the Busy Wait box to specify
how long to wait for a busy modem or dirty phone line before disconnecting. The
Modem volume slider lets you decide how loud to make the connection tones. Use the
Modem Commands button to set the default control codes for certain functions. The
Query Modem button can be used to test the modem, and the Terminal pushbutton
pops up a mini-terminal for entering control codes manually.
! Click the OK button at the bottom of the panel to close the KPPP Configuration panel
and return to the KPPP panel.
! In the Connect to field, select the ISP for which you just created a record. Then enter
the Logon ID and Password you were assigned. Make sure the Show Log Window
box is checked. This will let you see the progress of the logon process, which is
helpful in case there are problems logging on. Then press the Connect button to dial
the ISP. If you are unable to connect, verify all of the information you entered in
previous steps. If you do not find anything wrong, contact the ISP or network
administrator for further assistance.
• If you are using Gnome rather than KDE, there is nothing packaged with the standard
Linux distributions, but there is a downloadable tool you might try called Gnome Dialup.
As of this writing, it is still under development but worth a try if you are looking for a
graphical dial up access setup tool for Gnome. Gnome Dialup can be found at
http://tor-pw1.netcom.ca/~phantom/gnome_dialup/index.html. Other such tools may exist
as well.
• From the command line, you can use the pppd command to start the PPP daemon. This
approach is not for the faint of heart, however. There are many parameters that need to
be set via this command, so unless you are quite knowledgeable about Point to Point
Protocol setup or know someone who is, this probably is not the best approach for you.
For more on the pppd command, read the instructions at: info pppd.
If you need help getting connected to the Internet via your office network, contact your help
desk. If you are trying to set up Linux for DSL or cable modem (non-dialup) access at home,
contact your DSL or cable modem provider for assistance (and hope someone there has ever
heard of Linux).
Miscellaneous
This section contains questions or problems that did not fit into any of the other categories.
Q. What is the difference between Linux and UNIX?
UNIX began as a proprietary operating system developed by Bell Laboratories in the 1960s. It
eventually spawned a number of mutually incompatible commercial versions from such
companies as Apple (Mac OS X), Digital (Digital UNIX), Hewlett-Packard (HP-UX), IBM
(AIX® ), NeXT (NeXTSTEP) and others.
Linux is an attempt to create a nonproprietary operating system for the masses. It is a free,
open source, UNIX-like operating system, originally begun by Linus Torvalds in 1991. “Linux”
really refers to only the operating system kernel, or core. More than 200 people have
contributed to the development of the Linux kernel. The rest of a Linux “distribution” package
Preparing Today for Linux Tomorrow
47
consists of various utilities, device drivers, applications, a user interface and other tools that
generally can be compiled and run on other UNIX-based operating systems as well.
Q. Why does Linux seem so haphazard?
You may have heard the old saw that a camel is a horse designed by committee, illustrating
the difficulty of getting a number of people to agree on the design of anything, and the results
of compromise. If anything, Linux illustrates a horse designed by complete anarchy, with a
healthy helping of Darwinian natural selection thrown in for good measure. Thousands of
programmers over the years have written applications, utilities, games, device drivers, user
interfaces and other software for Linux. In the absence of a central authority to dictate a
mandatory “look and feel” each programmer had his or her idea of what was the best design
for software and developed the programs accordingly. Over time certain software became
more popular than others, which in turn spawned “work-alikes” and “near-clones,” taking
features of the previous products and adding to them, or offering slightly different variations.
Eventually, companies such as Red Hat, Caldera, TurboLinux and SuSE took the Linux kernel
and added what they considered the best available utilities, user interfaces, etc., to form a
“distribution,” or packaged deliverable. Because each distribution chose from among this
programming chaos, it was inevitable that the various programs selected for a deliverable
looked and worked differently from one another, and that each deliverable looked and, to
some extent, worked differently from other deliverables. This is the reason for the
inconsistencies between programs in a deliverable. The flip side is that because the
distribution companies were free to choose from among all available programs, not just those
that slavishly conformed to an arbitrary style, the deliverables were able to include the very
best programs—even the ugly or nonconformist ones.
Q. Where can I download or buy Linux software?
There are quite a few sources for free or commercial Linux software if your local software
store doesn’t carry what you are looking for. Here are just a few of them:
For free software, try:
• DLR Fresh Archive (http://www.go.dlr.de/fresh/linux/src)
• Freshmeat (http://freshmeat.net; love the name!)
• Linux Apps (http://www.linuxapps.com)
• Linux Archives (http://home.linuxarchives.com/software.html)
• Linux Game Tome (http://www.happypenguin.org)
• Linux Games (http://www.linuxgames.com)
• Linux Software Encyclopedia (http://stommel.tamu.edu/~baum/linuxlist/linuxlist/linuxlist.html)
• Linuxberg/Tucows Linux (http://linux.tucows.com)
• Slashdot (http://slashdot.org)
If you see some commercial software that you would like to buy, there are several Web sites
in the United States with extensive Linux inventories:
• Indelible Blue (http://www.indelibleblue.com)
• Linux Mall (http://www.linuxmall.com)
• The Linux Store (http://www.thelinuxstore.com)