2. Before you begin
2.4 The special 'scholar' account
2.5 Taking notes
3. Preparing the hard disk
3.1 Scrubbing the disk
3.2 Partitioning the disk
3.3 Formatting the disk
4. Basic installation
4.1 Copying the .CABs
4.2 Running SETUP.EXE
5. Saving your work
5.1 Documenting your work
5.2 Creating milestone images
5.3 Where to take it from here
6. Installing drivers
6.1 Chipset drivers
6.2 Network interface card
6.7 Unknown devices
7. Network connectivity
7.1 Running the Connection Wizard
7.2 Finding the MAC-address for computer C21
8. Windows Update
9. Do the samba
9.2 My Documents
9.3 Roaming profiles
10. Configuring devices
10.1 Printers overview
10.2 Configuring the JetDirect network printer
10.3 Configuring a local printer
10.4 Sharing the local printer with the network
10.5 Configuring a locally shared printer via the server
10.6 Connecting a USB printer
10.7 Checking printer functionality
10.8 Other (local) devices
11. What's next?
11.1 Recommended software
11.2 Tweaking the image for ServerAtSchool
11.3 Cleaning up
12. Concluding remarks
COPY, and the programs
RAWRITE.EXE. When these commands and programs are unfamiliar to you, please ask help from someone who knows these basic operations. In the text, we try to be as elaborate as possible.
UPPERCASEcharacters are used. This does not mean you have to type uppercase. Commands etc. can be issued in lower or upper case, even mixed.
H:\My Documentsrather than the otherwise equivalent
[Enter]. We will omit "... and press the
[Enter]key on your keyboard" with every instruction in order to keep the documentation as readable as possible.
[Ctrl-Esc]or the special Windows button on the keyboard. The choice is yours.
[Enter]for the 'Enter' key and
[Tab]for the tabulator key. Combinations of keys are displayed as follows:
[Alt-U]means: 'press and hold the Alt key and hit the key labeled U' and
[Ctrl-Alt-Del]means: 'press and hold the Ctrl key and the Alt key and hit the Del key'.
If ncecessary, consult section 3. Preparing the server in chapter II. Preparing the hardware for installation of the ServerAtSchool Installation Guide. The suggestions given there also apply to the workstations.
If you use g4u to manage disk images it is especially important that
all the hard disks are identical or at least have the same
geometry. The g4u program simply rewrites a complete disk, including
information about this geometry. Sometimes it is possible to use an
image of a (say) 4 GB disk after restoring it to a 6 GB disk, but
success is not guaranteed. An in-depth explanation is beyond the scope
of this manual. Use
SYS.COM. You are on your own.
In practice there will be slight differences which are acceptable and allow for using the same image on all workstations. One difference might be the optional CD-ROM player. Even if you use workstations without a CD-ROM player, you need at least one workstation with a CD-ROM player. Otherwise you will have a hard time to install Windows from CD-ROM and/or additiona network drivers from CD-ROM. Note that you could also add a CD-ROM player to a workstion temporarily. As soon as you have the workstation up and running (including network connectivity) you might be able to proceed without CD-ROM player. This would make your workstations more identical. You can consider removing the CD-ROM player just before creating the final image (see 4. Creating the golden image in chapter VIII. Managing user profiles).
When hardware differences occur, a restored client will certainly show problems during its startup. Some of those may just be annoying, e.g. a different video monitor connected to the workstation which yields a message like 'Windows has detected new hardware'. These problems are solved by a simple '[Next], [Next], [Finish]'. More severe problems are caused by different chip sets and indeed different disks.
NOTICE: A special note on built-in modems: you may want to remove those devices from the workstation. Generally they are not needed to communicate with the outside world because the ServerAtSchool server takes care of that. Furthermore, sometimes these devices are recognised as a networking device. This can yield problems that are difficult to resolve.
Helper for Software Installation
same as userid (default)
janitors(see notice below)
USER.DATand whatnot by someone guessing the correct password for the 'scholar' account.
It is very convenient to make the 'scholar' account a member of the
'janitors' group. This grants various privileges to this account, in
particular it allows access to shared directories like
H:\My Documents\janitors\shortcuts). That is very handy
for storing shortcuts ('icons') as you install educational and other
software. You build your collection of shortcuts systematically this
way without risking to lose them to an accidental deletion.
The partition that is to be used for installing software on the workstation has to be large enough to hold all necessary files. However, it should not be much larger because eventually you could end up with random data in the unused space on the disk. As a rule of the thumb: a 4 GB partition is usually large enough to hold all necessary files (both of the operating systen, the utilities and the educational software). Given the fact that all modern disks are (a lot) larger than 4 GB, it is necessary to manually partition the hard disk. On a 20 GB disk, you would use only 4 GB. The remaining 16 GB would be initialised to zero's once and remain in that state. That allowes for good compression with g4u and therefore a quicker restore.
If you have not yet created an Autoclave diskette, you can follow the instructions in sections 3.1.1 Creating an Autoclave diskette on the server or 3.1.2 Creating an Autoclave diskette on another computer below to do so. If you have the Autoclave diskette, you can proceed with section 3.1.3 Zeroing the hard disk below.
In short the procedure is as follows.
dd if=/home/share/install/goodies/autoclave/clave03.img of=/dev/fd0
dd(1)is done and your bootable diskette is ready.
You need to have root privileges for this procedure. You can obtain
those by logging in as root at the console. You can also logon to the
server using your regular account and escalate to root privileges
sudo(8), if you are authorised to do so.
You can now proceed with the procedure in section 3.1.3 Zeroing the hard disk below.
Both files are available on the server (in the directory tree under
/home/share/install/goodies). They can also be accessed
via the 'Network Neighbourhood' on a correctly configured workstation.
Both files should be copied to a directory on the workstation.
However, you need a workstation that is
already configured to communicate with the server. This is a classical
chicken-egg situation since we are setting up a workstation in this chapter.
Fortunately there is another way. You can simply use the
ServerAtSchool distribution CD-ROM on an existing DOS/Windows
computer. Here is the step-by-step procedure:
RAWRITE.EXEto a temporary directory on the workstation. Remember in which directory you saved the file.
CLAVE03.IMGto the same temporary directory on the workstation.
RAWRITE.EXEand click on the [Open] button.
Enter source file name:
Enter destination drive:
Please insert a formatted diskette into drive A: and press -ENTER-
Done., indicating the diskette writing process is finished.
A:first and retry.
[Enter], or wait a while for system to boot from the diskette automatically after a time out.
I understand.and press
[Enter]. Do type the dot and do not type the double quotes. Please note that using 'Autoclave' really removes all data from the hard disk.
Are you absolutely sure you want to do this? (Y/N). Type:
Donefollowed by a line with a single
The complete hard disk has now been overwritten with null bytes. An image of a disk filled with exactly the same bit pattern is very easy to compress (for g4u). Compression levels up to 100 times are possible. For example a 20 gigabyte disk. Say four gigabytes are used for the software, so 16 gigabytes of null-butes remain and can be compressed to 160 megabyte. That is a time- and diskspace saver.
FDISK.EXE. This tool can be found on the Windows 98 SE CD. Although we assume the usage of these progams is well known, here is a short reminder. If you want to create a partition of 4 GB on the (say) 10 GB hard disk of the workstation you are installing, you need to boot the computer from the CD and subsequently use
NOTICE: Even when you have a small disk, say 4.3 GB we advise to make the partiiton a bit smaller. If the computer manufacturer equipped the computers with 4,3 GB disks of two different brands or two different types, small differences may exist in the total amount of disk space. It is better to be safe than to be sorry. You won't miss the 300 MB on a 4.3 GB disk.
Take the following steps to boot the computer from the CD-ROM.
2. Boot from CD-ROMusing
[Enter]or simply by pressing '2'.
2. Start computer with CD-ROM support, again using
[Enter]or by simply pressing '2' followed by
If you wish to partition the hard disk of the workstation you should take the following steps.
[Enter]in order to start the partitioning program.
Do you wish to enable large disk support (Y/N)......? [Y]by pressing
FDISK Optionsscreen appears. Four options are available. Choose option
1. Create DOS partition or Logical DOS driveby pressing '1' followed by
Create DOS Partitions or Logical DOS Drivescreen appears.Three options are available. Choose
1. Create Primary DOS Partitionby pressing '1' followed by
Verifying drive integrityfollowed by an increasing percentage is displayed. This means that the disk is being checked.
Create Primary DOS Partitionscreen appears. You are presented with the option
Do you wish to use the maximum available disk size for a Primary DOS Partition and make the partition active (Y/N ......?[Y). Do not accept the default choice, but press 'N' followed by
Verifying drive integrityfollowed by an increasing percentage is displayed. This means that the disk is being checked, again.
Enter partition size in Mbytes or percent of disk space (%) to create a Primary DOS Partition ......: . Note that the full disk size (here: 10257 megabytes, 10 GB) is the default value. Do not accept this default but enter a suitable value for your situation, for instance 4 GB or 4000 MB. Enter '4000' and confirm the dialogue by pressing
Create Primary DOS Partitionscreen appears once again. Notice the message at the bottom of the screen:
Primary DOS partition created. Also note that the screen shows a table with a single entry describing the newly created C: partition of type
PRI DOS, a size of some
4000 Mbytesand a still
UNKNOWNfile system. Follow the instruction at the bottom of the screen and press
FDISK Optionsscreen appears again. Note that the following message is displayed on the screen:
WARNING! No partitions are set active - disk 1 is not startable unless a partition is set active. Choose option
2. Set active partitionby pressing '2' followed by
Set Active Partitionscreen appears. It shows the partition table with the single entry as before and the following question:
Enter the number of the partition you want to make active ....[ ]. Answer the question by typing '1' followed by
Ais displayed in the column labeled
Status. Also note the message
Partition 1 made active. Follow the instruction at the bottom of the screen and press
FDISK Optionsscreen appears once again. Follow the instruction at the bottom of the screen and press
[Esc]to exit FDISK.
You MUST restart your system for your changes to take effect. Any drives you have created or changed must be formatted AFTER you restart. Shut down Windows before restarting.Once again, follow the instruction at the bottom of the screen and press
[Esc]to exit FDISK.
The result of this procedure is that a single partion of 4 GB on the hard disk now exists. Also this partition is designated the 'active' partition which means that this partition can be used to boot the computer from after the operating system has been installed.
Note that it is necessary to reboot the computer for the changes to take effect. After rebooting, the partition has to be formatted. This is discussed in the next section.
FDISK.EXEit must be formatted using
FORMAT.COM. However, after FDISK the computer must first be rebooted from the Windows CD-ROM, using the same procedure as before.
2. Boot from CD-ROMusing
[Enter]or simply by pressing '2'.
2. Start computer with CD-ROM support, again using
[Enter]or by simply pressing '2' followed by [Enter].
At this point you can use
FORMAT.COM to format the
C:-partition. However, note that The
is not on the
A: drive but on the
WIN98 directory. Therefore you have to either
start the program including the exact pathname or change to the
D:\WIN98 first, taking advantage of the
built-in search path that always includes the current directory.
In order to format the C:-drive you should take the following steps.
A:\>prompt, issue the following command:
D:\WIN98\FORMAT C: /S
WARNING, ALL DATA ON NON-REMOVALBLE DISK
DRIVE C: WILL BE LOST!
Proceed with Format (Y/N)?
12 percent completed
Writing out file allocation table
Caclulating free space (this may take several minutes)...
Volume label (11 characters, ENTER for none)?
[Enter]here because there is no point in assigning a volume label to a drive that is subsequently cloned, thereby giving all drives exactly the same name.
FORMAT.COMnow shows an overview of total space, space used by the system and available space. This refers to the space on the formatted drive (partition), and not the complete disk. Then the program terminates and once again you are back at the
At this point the C:-partition of the disk has been formatted, i.e. a file system has been created in the partition. This file system is of type 'FAT32'. This is the standard type for large drives, i.e. drives larger than 2 GB.
Note that the
/S switch made sure that the files
necessary for the minimal bootstrap (
COMMAND.COM) were copied to
the disk. The effect is that the computer is now able to boot from the
hard disk, eventually yielding the well known
prompt, without the need for a boot floppy or indeed the bootable
Windows 98 CD-ROM.
NOTICE: It is very efficient to keep the computer running at this time, i.e. for once you do not need to reboot. The C:-partition is ready for use, the three essential system files are transferred and you are ready to proceed with the basic installation. This will be discussed in the next section.
The computer we use to illustrate the whole process is a 'difficult' one. It is a Dell Optiplex GX110 with network, video and audio on the motherboard. These devices require additional drivers not available in the standard Windows distribution. Also this computer has no way to set the BIOS to a well known state via an option 'Restore factory defaults' or something like that in the BIOS Setup.
C:drive. This is done by copying the Windows
.CABfiles and all other setup files in the
WIN98directory to the hard disk. It is very convenient and time saving to have your
.CABfiles on the C: drive. Whenever Windows needs a file from a
.CABfile, it will not ask for the CD, but simply look for it on the disk.
In case you wondered:
.CAB files are so-called 'Cabinet'
files. These archive files each contain one or more other files much
in the same way as in a
.tar file or a
At this point we assume that you have partitioned and formatted the C: drive according to the instructions in the two previous sections and that the computer still is running, having been booted from the Windows CD.
Here are the steps you should take to copy the relevant installation files from the CD to the disk.
COPY D:\WIN98\*.* .
[Ctrl-Alt-Del]. The computer will reboot from C:.
NOTICE: Some very self evident steps during the basic installation are not described below. For instance, when you want to create a Windows rescue diskette, you have to have a blank, formatted, HD diskette ready with the erase protection set to 'off'.
NOTICE: We assume that the workstation is properly connected to the server, either directly via a crossover cable or via a hub or switch. The server should have a working connection to the Internet. That allows for downloading additional software (drivers, security fixes, documentation etc.) if necessary.
NOTICE: You may want to take a look at the file SETUP.TXT. This file contains some more information about the Windows 98 setup process. This file was copied to C:\WIN98\SETUP.TXT in the previous section.
Here are the steps you should take for the basic installation of Windows 98 SE.
C:\WINDOWS) and press [Next].
C21(see notice below)
EXEMPLUM(see notice below)
Dell Optiplex GX110
.CABfiles. This may take some time.
Exemplum Primary School
The reasons to choose computer name 'C21' in the 'Identification'
dialogue are explained in section 5.1 Default configuration in
ServerAtSchool, in chapter V. Configuring all ServerAtSchool
components in the ServerAtSchool
Installation Guide. The computer name 'C21' or
c21.exemplum.serveratschool.net is the first name in the
range of static IP-addresses. This requires a mapping between the
hardware address of the computer's network card and the
IP-address. This will be discussed in general in section 2.4 The list of MAC addresses
in chapter V. Managing disk images with ghost for unix. The
mapping for this workstation is discussed in section
7.2 Finding the MAC-address for computer C21
below. Note that by convention the fully qualified domain name of the
c21.exemplum.serveratschool.net) is written
in all lowercase characters, whereas the computer name in the Windows
C21) is written in all capital letters.
NOTICE: The name of the workgroup in the 'Identification' dialogue should match the name entered in the Samba-configuration, see section 12.1 The Samba configuration file in chapter V. Configuring all ServerAtSchool components in the ServerAtSchool Installation Guide.
NOTICE: It is very important that both the client and the server agree on the same time zone. See section 4. Time zone in chapter III. Using the text mode installation program in the ServerAtSchool Installation Guide. If the client and the server have different opinions about the modification times of files, it may become difficult for the server (the hourly backup) to determine which is the latest version of a particular file.
There are different approaches possible. The first one is to keep this
documentation on the server, in a file that is accessible for members
of the group 'janitors'. A logical location for such a 'journal' or
'diary' would be a directory
Another approach is to keep track of the status of a workstation (or a
workstation image) by documenting changes in a file on the
workstation itself. A good place to keep such a track record is a plain text
file in the root of the C: drive, e.g.
You can edit this file with
NOTEPAD.EXE. You should simply write down every step you
took during the installation and save the file on the
C:drive before you make a new image. In this way
you keep track of your actions. It is not uncommon to end up with 15
images when installing many applications. Of course when the final
image is there and proves to be working good over some period of time,
you can delete some of the intermediate images.
Chapter V. Managing disk images with ghost for
unix discusses the processes involved, notably section 4. Restoring an image on a workstation and
section 5. Creating an image of a
workstation. With the g4u program you simply would have restored
the workstation to the state before some vague, ill documented
application ruined your up until then fine installation.
While the workstation is being restored, you can relax and think what action to undertake with the suspect application. Did you overlook something during the installation? Try again! Install it with other options? Go ahead! All these possibilities can be tried without the danger of losing all your work. When you have found the correct solution, you again restore the last 'known-good' image and install you final solution.
The bottom line is that you should periodically create a new image using g4u. Let's call them 'milestone images'. After you have reached the 'ultimate' image, you may want to consider to delete a few of the older milestone images. Until then it is best to keep them all.
It is not a bad idea to keep at least one basic image. For example, the one with the basic installation, with Windows Update (security fixes), with all drivers installed, but without any additional software.
If you need to search for drivers on the Internet, it is important that the network interface is fully functional. However, before that is done, the network interface itself has to have the correct drivers. This is yet another chicken-and-egg problem. In this section we will discuss potential problems with the low-level drivers (including the network drivers). In the next section we will return to the additional configuration of the network interface. It is possible that you have to go back-and-forth between configuring the network interface and searching for drivers and downloading them from the Internet.
You can check if all chipset drivers were found by navigating to: Start | Settings | Control Panel | System | Device Manager. When you see big yellow question marks at items in 'System' like: PCI, bridge, System board, etcetera, the drivers were not found. Consult the motherboad documentation, perhaps on the driver/motherboard CD-ROM, etc. When nothing is available, your quest on Internet starts here. When you have found drivers, install them and repeat the check.
Here is a possible procedure to install the NIC driver software:
PCI Ethernet Controller. Press [Next] to confirm the search.
At this point the driver for the network interface controller is installed. The standard Windows Desktop is displayed on the screen.
NOTICE: There are many other ways to install driver software. Much depends on the effort the manufacturer has made if there is a full-fledged software installer or just a diskette with a few kilobytes of bare driver software. Please consult the appropriate documentation.
The NIC drivers are now installed. This can be inferred from the fact that the yellow question mark is no longer visible when you navigate to: Start | Settings | Control Panel | System | Device Properties. Also you can see that a new icon is added to the Desktop: the 'Network Neigbourhood'. For now this is enough evidence that the installation of the drivers was successful. We will return to configuring network connectivity shortly, see section 7. Network connectivity below.
If the driver is installed. check if you can change screen resolutions by navigating to: Start | Settings | Control Panel | Display | Settings. ServerAtschool recommends a 800 x 600 pixel resolution as standard, but you can use other resolutions as well, depending on the capabilities of your hard- and software and the requirements educational software may impose. The resolution is set in 'Screen Area'. Another recommendation is to set 'Colors' to High Colors 16 bit. This is done in 'Colors'.
NOTICE: When your school uses different brands and types of monitors it's a good idea to connect each one of them to the workstation in the course of the software installation process. They will be recognised and a 'New hardware found' message will be displayed. As a rule a simple [Next], [Next], [Finish] is enough to add this monitor to the list of 'known' monitors, preventing the 'New hardware found' message in the future.
Check if the sound driver is installed properly:
At this moment there is little more to do than ckecking if the mouse functions. In the section 11.2 Tweaking the image for ServerAtSchool we will give some additional tips.
You can check if the driver was found in: Start | Settings | Control Panel | System | Device Manager. Now yellow question mark serves as an indication that the drivers are installed correctly. Configuration of actual USB-devices is discussed in section 10. Configuring devices below.
[Esc]] on your keyboard. Do not specfy a username and password.
These simple tests prove that the network interface is working and you have a connection with the server and with the Internet.
NOTICE: There may have been an icon on your desktop by the name of 'Connect to the Internet'. Doubleclicking that icon should have presented you with the same 'Internet Connection Wizard'. However, it also might have taken you to a setup procedure for one of the 'Online Services'. You have prevented this unexpected behaviour by going through Control Panel and Internet Options. Either way, if you have completed the wizard, the icon will be removed from the desktop automatically.
NOTICE: It is possible that the system tries to seduce you to setup an e-mail account and/or install Outlook Express. Resist the temptation and just say no. ServerAtSchool comes with its own mail server including webmail with access control based on client IP-address. For a school this is a much better proposition.
NOTICE: It is possible that your local systems administrator has enabled the proxy server. In that case you should enable the proxy settings in the Internet Connection Wizard too; both the proxy server and the client software should agree on this or else it will not work.
In order to retrieve the MAC-address of the workstation, you should take the following steps.
NOTICE: The IP address you observed is an address from the so-called dynamic pool. You really want to have an address from the fixed pool. See also section 5. DHCPD in chapter V. Configuring all ServerAtSchool components in the ServerAtSchool Installation Guide.
If you want to link the fixed IP-address 172.17.2.21 to the MAC-address 00:B0:80:9B:DC:F7 you just retrieved, you should take the following steps.
/etc/dhcpd.conf, e.g. using
vi(1)and add the following lines inside the pool with static addresses.
At this point you can reboot the workstation (do not enter
username or password but press [Cancel] or hit
and use WINIPCFG once again to see if the workstation now has the
NOTICE: The mapping between MAC-address and IP-addres has to be done for all workstations in the school. This will be discussed in section 2.4 The list of MAC addresses in chapter V. Managing disk images with ghost for unix.
You might already have noted the MAC-address in an earlier stage if
you would have used the
ifconfig ex0 command with g4u as
discussed in section 3.3 The client's MAC
address in chapter V. Managing disk images with
ghost for unix.
Here is the procedure (somewhat sketchy):
At this point your workstation is as patched as is possible. Now is
certainly a good time to make a fresh 'milestone image' as per the
procedure described in section 5. Creating an
image of a workstation in chapter V. Managing
disk images with ghost for unix. Do not forget to update the file
C:\CONTENTS.TXT before creating a fresh milestone
image. After you have created an image you can return here and
continue with a bit of fun.
[Esc]'ing the 'Enter Windows Password' dialogue, continue with the following steps.
scholar(see section 2.4 The special 'scholar' account above)
Adp+w4K3(see section 2.4 The special 'scholar' account above)
EXEMPLUM(this is what you entered previously in step 5 under 'Windows NT Domain')
NOTICE: With leaving the 'Windows Password' blank (by removing the
asterisks in the pertinent dialogue) you save yourself some
typing. Windows still creates a file named after the user (here:
SCHOLAR.PWL in the directory
will take care of disabling this feature lateron (by setting the
DisablePwdCaching), in section 11.2 Tweaking the image for ServerAtSchool below.
NOTICE: The 'scholar' account used here to logon will be the account under which all software will be installed. That is its only raison d'etre. Therefore this account should not be used as a real production account; it should be kept as clean as possible.
Scholar on 'Server'(H:). You see two or three directories: a hidden directory '.profile', 'My Backups' and 'My Documents'. Select the 'My Documents' directory.
Check if you really are looking at the 'My Documents' on the server by double clicking on the 'My Documents' icon on the desktop. You should at least see the 'janitors' directory. For further explanation on this 'My Documents' directory see chapter XI. My Documents and My Backups.
Of course it would have been possible to simply change
H:\My Documents by simply typing in
the new location. However, this only works if you do not make typign
As an extra precaution you can warn users who try to save documents on
the local drive, i.e. in
C:\My Documents directory, by
creating a file there and naming it something like 'DO NOT SAVE YOUR
DOCUMENTS HERE.txt'. This document will show up whenever a user
C:\My Documents in order to save a file.
The contents of this file can fully explain why saving on the local
drive is not a good idea.
USER.DAT, stored in
H:\.profile\USER.DAT. This is the part of the Windows Registry that holds all personal settings.
H:\.profile\Desktop. This collection of
.PIFfiles basically defines the user's Desktop.
H:\.profile\Start Menu. This collection of
.PIFfiles basically defines the user's Start Menu.
Some configuration has to be done before these personal settings are in effect. The procedure will be discussed shortly. However, it is best to postpone flipping this switch until most of the educational software has been installed. This gives you the least headaches with 'unknown settings' and endless prompts of applications insisting that they need to be installed first even though this has already been done. We will enable personal profiles in the aforementioned chapter IX. Shortcuts, desktops and start menus. Please be patient.
By the way, now is a good moment to make a fresh 'milestone image'.
NOTICE: We mention roaming profiles here but do not do this step at this moment! This final step must be postphoned at least until after chapter VII. Installing educational software. Consider yourself warned.
The procedure to enable roaming profiles:
At this point the personal settings will take effect. It means that as
soon as a user logs in, the current settings as they are
stored on the server are copied from the server to the
workstation. This includes a copy of
USER.DAT and all the
items found on the Desktop and in the Start Menu. As soon as the user
logs out, the (possibly changed)
items on the Desktop and in the Start Menu are copied back to the
The first time ever a new user logs in, nothing is copied from
the server because at that time there is nothing to copy from
the server (the directory
H:\.profile\ is basically
empty). As soon as this user logs out the first time ever, a copy of the
USER.DAT and the Desktop and Start Menu of the
Default User are copied from the workstation to the server. This explains why it is
important to first get all the settings right (by installing
educational software etc.) without having personal profiles, only to
enable the personal profiles after the workstation is 'complete'. In
other words: the process of installing (educational) software
eventually builds a single
USER.DAT on the workstation
which will be used as a template for new personal profiles as soon as
they are created, i.e. when a new user logs out for the very first
As stipulated above, enable this functionality only when you have installed all software. Please follow the instructions in this section when you have progressed to section 3.1 Enabling roaming profiles in chapter VIII. Managing user profiles.
A useful guide during this part of the installation is chapter 3 of Using Samba, 2nd ed., Jay Ts, Robert Eckstein & David Collier-Brown, February 2003, O'Reilly & Associates, Sebastopol, CA, USA, ISBN 0-596-00256-4. See [Ts, Eckstein, Collier-Brown] in Appendix B. Bibliography.
Most peripheral devices, printers in particular, either have a parallel port or a Universal Serial Bus (USB). The best way to configure a device with a parallel port is to start with the workstation switched off. Then the device should be connected to the parallel port and switched on. Only then the workstation should be switched on. You will get a familiar warning 'New hardware found' or the drivers are automatically installed. In the first case, when you navigate to: Start | Settings | Printers, you find the installed printer. In the latter case, you can install the printer software in much the same way as the NIC driver. You can check the installation via Start | Settings | Printers.
The best way to configure a USB device is not connecting it to the USB port initially. First install the driver (from a CD-ROM or otherwise), if necessary reboot, and only then insert the USB device or its cable into the workstation. New hardware will be found and automagically recognised.
NOTICE: Serial devices with RS-232 interface, Firewire and Bluetooth devices are beyond the scope of this manual. Please consult the manufacterers documentation.
This section focusses mainly on the installation of printers on the workstation, be it that various flavours are discussed.
In section 10.2 Configuring the JetDirect network printer below we will configure the workstation for an HP Laserjet with built-in network interface. The server part of this printer is already configured during server installation (see 20.1 Setting up a JetDirect network printer with CUPS in chapter V. Configuring all ServerAtSchool components in the ServerAtSchool Installation Guide).
In section 10.3 Configuring a local printer below we will configure the workstation for an HP Deskjet 930C with parallel interface.
In section 10.4 Sharing the local printer with the network below we will discuss the way a locally connected printer can be shared with the network. This step is necessary before other workstations (or the CUPS server for that matter) can print to the printer connected locally to a workstation.
In section 10.5 Configuring a locally shared printer via the server below we will discuss the necessary steps that make the example in section 20.2 Setting up a shared printer via SMB of chapter V. Configuring all ServerAtSchool components of the ServerAtSchool Installation Guide actually work.
During the installation of the server, you have made the decisions depending on your hardware. The Exemplum Primary School has one network printer (the HP Laserjet) which is installed in the central hall. There is also a parallel printer (the HP Deskjet 930C) which will be connected to workstation 'C34' using a parallel cable. This printer will be located in the Faculty office.
NOTICE: This subject can be quite confusing. This is mainly caused by the fact that the 'master' workstation you are preparing (computer 'C21') will eventually be cloned to computer 'C34', too. Among other things this means that if computer C34 needs to do something that requires particular software, the software needs to be installed on the image, i.e. it has to be installed on computer 'C21'. And, if computer 'C34' is to share the parallel printer with the network, then computer 'C21' necessarily has to do so too, for 'C21' is 'C34' and 'C34' is 'C21' once the single workstation image is restored on all workstations.
NOTICE: The third option mentioned in the Installation Guide (a printer connected to the parallel port on the server) is comparable to installing the networked printer discussed in section 10.2 Configuring the JetDirect network printer below, at least as seen from the workstation. Therefore it will not be discussed. Furthermore, this is rather uncommon in a school where the server usually is located in an unaccessible place such as a broom closet.
There are even more ways to connect a printer but an in-depth discussion is beyond the scope of this manual.
HP LaserJet 4/4M PostScript. This can be inconvenient and also confusing for end users. Therefore change the name to something more meaningful, e.g.
Printer Central. Confirm the dialogue by pressing [Next].
The printer is installed and when the printer is switched on, it should print the test page. Please check if everything looks OK.
NOTICE: We chose another name for the printer; we changed 'HP LaserJet 4/4M PostScript' into 'Printer Central'. The reason is that we think end users can recognise and select a printer with such a name more easily than some gobbledygook combination of brand name and model name. Furthermore, if one day the central printer is replaced by another brand/model, users could still use the familiar 'Printer Central'. In general end users don't really care about brands and models as long as their print job is processed quickly.
The real 930C will be connected to workstation 'C34'. However, we are now busy creating an image that eventually will be cloned to computer 'C34' too. Therefore we have to install the local printer on the master image machine (computer 'C21'), even if this particular computer will never have a local printer connected to it when the school network goes into production. It is a good idea, though, to go to the Faculty office now and fetch the printer. It would be handy to be able to test the parallel connection right here, right now, on machine 'C21'. You do not want to wait until your 'golden image' is distributed, only to find out that the printer does not work. So go ahead, and hook up the 930C to your workstation (for now). Make sure it is switched on.
In order to configure the local printer you should take the following steps.
HP DeskJet 930C. In this case accept the suggested name (it helps you distinguish between local printers, see notice below). Leave the option for default printer as-is. The laserprinter ('Printer Central') defined in the previous section is a good choice as default printer. Confirm the dialogue by pressing [Next].
Note that if a print job is submitted to the printer called 'HP DeskJet 930C', it will always be directed to the parallel port of the workstation. Obviously this will only work if a real printer is in fact connected to the parallel printer port. In the Exemplum Primary School this will be the case only on computer 'C34'. That is: the printer will be connected to the workstation located in the Faculty office. For all other computers the 'HP DeskJet 930C' is more or less meaningless. However, it is necessary that the drivers are installed, even if they would only be used on computer 'C34' and not on the machine you are working on right now (machine 'C21').
Also note that in this case we kept the suggested name for the printer ('HP DeskJet 930C') rather than change that name to something like 'Printer Local'. Of course there is nothing wrong with that (the same reasoning as that for 'Printer Central' applies). However, if the school would have another local printer but of a different model connected to another workstation, say an 'Epson Stylus 800' connected to computer 'C49' it would be confusing to distinguish the local HP printer and the local Epson printer. By keeping their default names you can easily distinguish the printers. Eventually you end up with an image that has two 'local' printers, one called 'HP DeskJet 930C' and one called 'Epson Stylus 800'. Because you are making an image both printers have to be defined.
The rule of the thumb is that printers accessed via the network could have a name indicating their physical location (e.g. 'Printer Central', 'Printer Faculty', 'Printer Principals') whereas locally connected printer should have their default names (e.g. 'HP DeskJet 930C', 'Epson Stylus 800'). If you stick to this rule you can tell by the name of the printer how it will be accessed.
If yo have not yet already added this service previously, you should take the following steps.
SERVER.VXDalready exists. Confirm the dialogue and keep the newer file by pressing [Yes]. (The newer version was installed during Windows Update).
Enabling file sharing is not necessary, however it can be
convenient. First of all, it's not that unsafe. You are not sharing
anything. You only enabled the service. A real-world application of
file sharing would be the following example.
The school has workstations without diskette drives; a security precaution. Only one workstation in the school has a diskette drive. It's the only client in school with the ClamWin virus scanner. On that client you can share the
A: drive for
the other workstations. So, when someone has data on a diskette, they
would need to take it to that particular computer, where it can be
scanned for viruses before the data are retrieved from the
diskette. The data retrieval itself can then be done from another
workstation via the shared drive
HP930C(see notice below)
HP DeskJet 930C (parallel)
The 'Share Name' you entered (here:
HP930C) should match
the share name configured in section 20.2 Setting up a shared
printer via SMB in chapter V.Configuring all ServerAtSchool
components in the ServerAtSchool
Installation Guide. This share name will be used to link the print
deskjet on the server to the local printer on
computer 'C34'. If these share names do not match, printing to the
printer via the network will not work.
Note that the 'Comment' you added (here:
HP DeskJet 930C
(parallel)) is displayed when you browse the network via
'Network Neighbourhood' with all details visible. Unfortunately,
because you are working with an image that is cloned to all
workstations, this comment will always be the same for all
workstations. Therefore you can not use this field to say something
like 'Maria's DeskJet' or 'Lev's Epson'; it has to be a generic
HP930Cin our example.
deskjetin our example. This queue is visible in the 'Network Neighbourhood' as
What remains to be done is to add yet another printer to the workstation.
NOTICE: In a way you are now installing the HP DeskJet 930C for the second time, in the same workstation image. The first time was in section 10.3 Configuring a local printer. Here we will be installing the same printer but under a different name: 'Printer Faculty'. This name and hence the printer can be used from any workstation in the network. The local printer added in section 10.3 is only meaningful to one workstation in particular, namely workstation 'C34'. On that workstation the local printer must exist because something must be there before it can be shared; you can not share 'nothing'.
If you want to add the printer to the workstation you should take the following steps.
HP DeskJet 930C (Copy 2). This can be inconvenient and also confusing for end users. Furthermore, a printer with a similar name already exists in the image (see notice below). You chould change the name to something more meaningful, e.g.
Printer Faculty. Leave the option for default printer as-is. The laserprinter ('Printer Central') defined earlier is a good choice as default printer. Confirm the dialogue by pressing [Next].
Keep in mind that on the server this network printer queue is configured to connect to workstation 'C34'. Since you are currently working on the image, using workstation 'C21', there is not yet a workstation 'C34'. This means that at this point you can not properly test 'Printer Faculty'. If you really want to make sure that everything works like it should, you should take the following steps.
At this point the default printer is 'Printer Central'. This may be the correct choice for most workstations. In section 4.4 Setting the default printer in chapter V. Managing disk images with ghost for unix the procedure for changing the default printer is discussed.
It is possible to add a printer that prints directly to the
printer share on workstation 'C34', e.g a network printer that prints
\\C34\HP930C rather than
\\SERVER\deskjet. However, this can cause severe
headaches if you restore the image containing this printer on computer
'C34'. Windows appears to be completely confused when it discovers it
has to print to a networked printer share that happens to be served
from the same machine. You really want to avoid this
situation. Printing to the printer share
\\SERVER\deskjet, which in turn prints to
\\C34\HP930C is OK, because the Windows on 'C34'
considers these two printer shares to be different and hence printing
is not a problem, even from workstation 'C34'. Of course we know
better, having 'fooled' Windows to print to itself using a detour via
A USB printer can be shared with the network in much the same way as the parallel printer we used as an example in the previous section.
NOTICE: Many USB printers come with complex printer control software. This software installs itself as a system service on the workstation. The software will then automatically start when the workstation is booted, occupying resources (memory, processing capacity) that could be useful for applications. If at all possible you should disable this kind of 'fancy' control software. It serves no real purpose and it only works on the particular machine which is connected directly to the USB printer. On the other workstations this software is simply useless.
NOTICE: Depending on the bundled driver software, a special 'USB printer port' that is printer-specific may be created by the software. It means that when you have a total of say 5 HP DeskJet 845C printers, 5 different 'USB printer ports' will be necessary. Each port is linked to an individual printer by using the printer's serial number as part of the 'USB printer port' name. This means that it is not possible to use one of those 5 printers as a pars pro toto but that you really have to install all printers.
The installation depends on the way of connection; via the USB or the parallel port. Please refer to the device's documentation. As always, try to minimise the amount of installed software. Most of the fancy features are less than useful but it still takes up valuable space and processing capacity.
NOTICE: Depending on the available exotic peripherals your school may have (camera's, scanners, CD/DVD burners, etc.) it may make sense to concentrate all of those on a single workstation. Perhaps the workstation of the school's caretaker is the right machine? Obviously this machine will then no longer be a 'standard' machine that the 'standard' image as created on 'C21' can address. However, having two different images, one generic and one very specific, may be good way of dealing with the odd one out.
If you have a multi-purpose 6-in-1 card reader or similar (allowing
you to copy images from different memory cards from different types of
electronic cameras) you may encounter the strange phenomenon that you
sometimes can not access your own
H: drive on the
network. This may be due to the fact that drive letters
J: are 'taken' by the card
reader. A work-around is to insert a memory card after you
have logged on to the network successfully. You should also remove the
card again when you are done with it. Sometimes it may be possible to
configure the software of the device to skip specific drive
letters. This depends on the driver software that comes with the
By the way, now would be a good time to create yet another milestone image.
This tool provides automation for the Microsoft QuickRes screen mode
change in Windows 95 and 98. It allows you to change color depth and
screen resolution without restarting the computer. QRes is very useful
for older educational applications that require low resolution and
color depth. The latest version of QRes can be found on the
ServerAtSchool CD. By default the QRes installation package
qres-setup1090.exe) is also copied to
/home/share/install/goodies/qres at install time.
If you wish to install QRes you should take the following steps.
qres-setup1090.exeand press [Open] in the 'Browse' dialogue.
If the installation routine complains about 'QuickRes is not enabled', you should enable it. You can do this by taking the following steps (after the workstation has rebooted an you are logged in with the 'scholar' account once again).
When an application needs adjustment of size and color, use the configuration options in this QRes tab. The options speak for themselves. If you are unsure, you can consult the help-function by navigating to: Start | Programs | QRes | QRes Help.
For some versions of Windows, notably Windows ME, it is necessary to
apply a patch. Please refer to the ZIP-files located in
\\SERVER\install\goodies\qres\patches. For a generic
Windows 98 environment these patches are not needed.
PuTTY can be used to connect to the ServerAtSchool server from any workstation in the school or from home. A systems administrator will be particularly happy with PuTTY because it enables working in a comfy chair or at home on the couch with a laptop rather than hanging upside down in a broom closet where the server happens to live.
You can download the latest version of
At the time of writing the latest release version was 0.58.
If you want to use PuTTY you should at least download the program file
PUTTY.EXE. A good location to store it would be the
C:\WINDOWS\COMMAND on the workstation. The
next-best choice would be on the server, in the directory
P:\BIN as seen
from the workstation.
By keeping the program in the workstation image you always have it at hand, even if for some reason the Samba server has stopped working (you could use PuTTY to login and restart Samba). An advantage of keeping PuTTY on the server is that it is easier to upgrade it to the latest version, when necessary. The choice is yours.
Another useful program that you might want to install in much the same
PSCP.EXE. This is the Windows-implementation of
scp(1). It can prove to be very useful to securely
transfer files from one computer to another where one of the machines
If you want to have all the benefits of the complete suite of
programs you may want to download the WIndows-style installer with
'everything'. Look for
If necessary, create a shortcut to
PUTTY.EXE the usual
way. Note that you need to be a member of the group 'janitors' if you
actually want to do something useful at the server command line after
logging in with PuTTY. After logging in you can use
sudo(1) to gain root privileges (when you know the root
password). The proper commands and more information can be found in
the section 21.3 Managing
privileges via sudo in chapter V. Configuring all ServerAtschool
components' in the ServerAtSchool
DOSKEY.COMis a program in DOS which allows you to recall and edit DOS commands from the command line. It is part of the Windows 98 software and is located in
C:\WINDOWS\COMMAND. If you add the call to this program in
C:\AUTOEXEC.BATit will be automatically loaded each time the computer is started. It may save you retyping DOS-commands because you can recall commands from the command history using
[CursorUp]. Here is an example of a complete
AUTOEXEC.BAT, with the line of interest emphasised. Note that your version of
AUTOEXEC.BATmay differ from the example below, maybe because of sound drivers that are loaded or other programs that are loaded via this file. The important thing is to add the line with
You can read
DOSKEY.COM's help function by opening the
DOS environment and issuing the command
In 'Item' you can set 'Icon Spacing (Horizontal)' from 43 to say 55 and line up icons.
This way the workstation is always in a ready state. If you (or a pupil) wants to use the computer it should be ready now and not at some point in the future after disks have been spinning up, screens have been restored and monitors have once again warmed up.
Why is this? This is an attempt to make access to the complete system
of workstations, servers and networking more 'userfriendly' and 'easy'
by storing various passwords in a password cache. This password cache
is kept in a file named after the user, with the extension
.PWL. For instance: for the (new) user Mary Astell with
userid 'mastel' the file
created, on the local machine. The password that is asked here (the
so-called 'Windows Password') is used to protect the contents of this
.PWL file. By default Windows suggests you to re-use the
same password as you used for access to the server (the so-called
'Windows Networking Password'). Even worse, this password is already
entered once in the name of 'userfriendlyness'; you only have to enter
it one more time.
There are several things wrong in this picture. The most important being:
The best is to get rid of the
.PWL files and
make sure they do not return automagically. The solution is to disable
the password caching mechanism alltogether. This can be done by
manipulating the Windows Registry.
The registry is, in brief, a central database for storing system-wide and application-specific settings and also configuration information. In the case of roving (or roaming) profiles, part of the settings are user-specific.
Windows provides a tool to edit the registry called
REGEDIT.EXE. Editing the registry can be tricky: if
you make an error the
workstation might no longer work at all. Fortunately with the help of
the g4u images it is easy to go back to a previous known good state.
However, you still need to check and doublecheck whenever you change a
A good way to make modifications to the registry is to prepare the
changes in a simple text file (using
another plain text editor) and subsequently process that file with
REGEDIT.EXE. This gives you the opportunity to check your
changes before you actually apply them. Furthermore by keeping these
.REG files you always know what you did afterwards. In a
way it helps you documenting your changes.
If you want to suppress the password caching by editing the Windows Registry, you should take the following steps.
H:\My Documents\janitorsand create a new directory ('folder') called
regedit. You will use this directory to store the file you are going to create.
NOTEPAD.EXE, create a new file.
PWDCACHE.REGin the directory
NOTEPAD.EXE. You are back at the desktop.
H:\My Documents\janitors\regedit, perhaps via the
My Computericon on the Desktop or otherwise.
PWDCACHE.REGin that directory. The Registry Editor prompts you with: "Are you sure you want to add the information in H:\My Documents\janitors\regedit\PWDCACHE.REG to the registry?"
.PWLfiles (there should only be one:
SCHOLAR.PWL) using the following commands:
After completing this procedure and logging in as user 'scholar' you should not be prompted for the Windows password. New users will also not be irritated by the annoyance of password caching.
NOTICE: Much more registry settings can be used to tweak the configuration. See chapter VIII. Managing user profiles for more information.
Note that if you ever need to re-enable the Task Scheduler, you can navigate to Start | Programs | Accessories | System Tools | Scheduled Tasks and then go to Advanced | Start Using Task Scheduler.
POLEDIT.EXEto further configure the workstation. Treating this subject is beyond the intention of this manual. A good reference is the Microsoft Windows 98 Resource Kit, Microsoft Press, Redmond, Washington, US, 1998.
POLEDIT.EXEand related topics are discussed in chapter 8: System Policies. However, before you start using
POLEDIT.EXEyou should read chapter VIII. Managing user profiles and specifically section 2.2 Preparing policies. At this point you should not use
POLEDIT.EXEon your workstation, read the aforementioned section first.
H:\My Documents\janitors\regedit\USERNAME.REG. The contents of that file may look like the illustration below.
You can add/modify this register setting using the procedure as
outlined in 11.2.5 Password caching before or by
processing the file from a DOS box with a command line like
REGEDIT /S USERNAME.REG.
After processing this registry setting the 'Enter Network Password'
dialogue wil not show a username the next time. (More precisely: the
name of the last user shown will be an empty string.)
You could also show a message in that field by using a slightly
modified version of
USERNAME.REG as illustrated below.
Note that this 'trick' (an empty name or a message) only works once, at the next logon. Therefore you have to redo this setting every time, just before you create another (milestone) image with g4u. Also note that you can not use more than 20 characters including spaces for this 'username' variable.
It could be very handy if you kept a copy of
C: drive. This way you can reset the username
without the need to be logged in. I.e. you can 'login' by pressing the
[Cancel] button or hitting
[Esc] and still be able to erase the last
username with the
If you never want to show the name of the
previous user on the workstation you could use the register
setting as shown in the file
Note that this is a more 'permanent' setting; the effect will be that the name of the last user is never shown in the 'Enter Network Password' dialogue. It also means that pupils have to enter their nest account ('grade8', 'grade7', ...) every time, even if it is the only account using that particulare computer. That can be annoying too.
You can reverse the effect by changing the dword value from '00000001'
into '00000000' and processing the
.REG file once again.
NOTICE: Much more registry settings can be used to tweak the configuration, see chapter VIII. Managing user profiles.
The feature could be disabled by changing the registry with the following recipe.
An important argument to disable the feature are the so-called copy protected audio CD's. If you try to play one of those on a computer, the CD might automagically install software onto the computer which limits your rights accessing the music on the CD, under the pretense of 'Digital Rights Management' or DRM. In some cases such an innocent-looking music CD might install software that hides itself from prying administrators eyes while at the same time crippling the PC. This is a very good reason to stay away from this Autorun feature. It can also make installation of educational software a little easier.
In the tab 'Buttons' you could adjust the doubleclick speed. Many young children do not yet have enough eye-hand coordination to properly move and doubleclick the mouse. This can result in 'strange' errors and 'weird behaviour' of the computer, even though the machine does exactly what it is told to do.
(Real world) example: if in a list of directories (such as you encounter in standard a 'Save As...' dialogue) you
In the tab 'Motion' you could lower the pointer speed, much for the same reasons of hand-eye coordination. Letting the mouse leave a pointer trail can help identifying the location of the mouse on the screen, especially on 'busy' screens with many 'loud' colours some educational software makers seem to like so much.
Finally in the tab 'Pointers' you can select a complete set of mouse pointers (called a 'scheme') or you can change individual pointers such as the 'Normal Select' pointer or the 'Busy' pointer.
A few 'schemes' are available in the standard Windows 98 distribution. Navigate to: Start | Settings | Control Panel | Add/Remove Programs | Windows Setup | Desktop Themes if you consider installing additional schemes. Note that these take up quite some space (23 MB) for limited added value. (The schemes also contain wall papers and sounds).
Alternatively you can add a few 'bigger' cursors to the system by navigating to: Start | Settings | Control Panel | Add/Remove Programs | Windows Setup | Accessibility. Look in 'Accessibility Tools' for high-visibility mouse cursors.
The ServerAtSchool distribution CD contains a file with a collection
of completely free mouse pointers. Look for the file
CURSORS.ZIP. By default this file is copied to the server
during install, into
\\SERVER\install\goodies. Installation instructions are
CURSORS.ZIPinto some temporary directory, e.g.
V:\CURSORS. The file unpacks into seven subdirectories.
My Computer. You can read the
README.TXTfile and browse the directory tree that was created.
C:\WINDOWS\CURSORS. Put both directories side by side on the screen.
At this point you have added your selection of cursors
.CUR files) and animated cursors (
C: drive. You now have to configure the mouse to use
Note that not everyone likes or needs huge cursors. Depending on the school's policy you might allow teachers to choose their own 'scheme' lateron. See also section 2.2 Preparing policies in chapter VIII. Managing user profiles. For now the choices you make are the 'standard' or 'default' choices.
By copying the cursors individually ('hand-picked') from the cursors
C:\WINDOWS\CURSORS you keep the image as
clean as possible. However, it is necessary to have the mouse cursors
stored on the
C: drive; you can not select cursors stored
on a network drive because they are not available when you are not
NOTICE: You may need to update or extend your browser with various add-ons or plug-ins in order to be able to play 'special' content from the web, e.g. movies that require QuickTime or Flash player. Please refer to Appendix A. Additional software for some more suggestions.
C:\CONTENTS.TXTand/or the journal file you keep on the server (see section 5.1 Documenting your work above).
After that is done it is time to clean up a bit. You may want to take a look at the following ToDo-list before you make your next clean(er) image.
C:drive. In the tab 'Disk Cleanup' mark all four items in the list 'Files to delete':
[Esc]key). Then process the registry file
USERNAME.REGyou created in section 11.2.8 Dealing with the previous user logon before.
Note that this cleanup should be done before every new image. Also
note that cleaning the username (with
be the last thing you do before you create a new image.
The workstation is ready, but not for general use by pupils (or teachers for that matter). It lacks standard tools and educational software. In Appendix A. Additional software we do a few suggestions for additional software that may come in handy on a school. In chapter VII. Installing educational software, we give some hints and examples on various educational software packages and the ways you sometimes have to work around design flaws.
If you have followed all the instructions in this chapter, you have already worked your way through chapter V. Managing disk images with ghost for unix and chapter VI. DAEMON Tools virtual CD player. The next chapter to read would therefore be chapter VII. Installing educational software.
Now, how about a hot or cold beverage while the next milestone image is uploading...?
Authors: Peter Fokker <peter (at) berestijn.nl> and Dirk Schouten <schoutdi (at) knoware.nl> $Id: workstation.html,v 1.26 2006/03/31 08:16:19 peter Exp $