Other GUIs on Linux/BSD/UNIX

KDE is not the only window manager that will run on UNIX-like systems. Choosing one depends upon various criteria such as how confident your users are, how much memory you have to play with (remember that Linux and other UNIX-like OSes will run on a fraction of what Windows XP needs to run on), the speed of the machine (again, you can run the OS on quite a low spec machine compared to Windows XP), how much disc space you have and so on.

(Note... click on any of the images on the right to display a 1024x768 version in a new window)

KDE

At the risk of starting a flame war, KDE is the easiest to use and most fully-featured of the common UNIX desktops. I've run it on Linux and various BSDs and it is very flexible. You can make it look and perform like just about any other GUI you want and it supports transparency as you can see with the panel across the bottom and the Midnight Commander window. There are plenty of games for it and OpenOffice.org and the GIMP will run without any problems. Burning CDs is extremely easy using KIIIB and there are plenty of other programs as well. KDE will also run Gnome programs.

Running in and designed from the outset for a multi-user, networked environment, the UNIX-like OSes allow KDE to run many network management tools such as Ethereal and so on.

Gnome

Gnome also runs well, has a fully-featured and easy to use interface and, as a warning, people are passionate about Gnome. Gnome has a sense of style that KDE has not got but by choosing to go down that path, it has lost some of the flexibility that KDE has. Again, network management and office/graphics support is excellent and Gnome has little if any trouble running KDE programs.

IceWM

The Ice Window Manager is beautifully lightweight - taking only around 2 seconds to load on a 1.8GHz machine.

Again, you have access to all of your favourite programs including the office and graphics programs. In addition, there are literally dozens of different themes.

WindowMaker

Window Maker is another lightweight widows manager and to get to the menu on this, you click the mouse anywhere on the blank screen. Again, there are dozens of different themes and you have access to your programs - in this case, you have the SUSE menu to go in addition to the WindowMaker menus.

BlackBox

BlackBox is very lightweight and menu access is via a mouse click.

Again, SUSE menus are accessed through this and you can use the same programs as mentioned above.

MWM

Just when you thought that BlackBox was lightweight, think again.

Menu access is the same as above.

FVWM2

Just when you thought that it was safe, fvwm2 is even more lightweight. This one has the interesting feature that you can drag a window off, say, the right hand side and it will then appear on the left of what was the desktop on the right which now becomes the current desktop. (Note that you can also do that with Gnome.)

This is not the most lightweight though as some essentially just provide a graphical means of accessing more than one terminal at the same time.

For the beginner, I would recommend KDE.

Google Maps and browser compatibility

Google Maps allows you to see satellite images of just about anywhere on the planet in whatever resolution is available - or, at least a reasonable one. The site is at http://maps.google.co.uk/ and if you have the right browser, it will display cornfield profanities, flying cars and much more - possibly ending up as even more of a waste of time than the initial Windows 95 option for the user to change the colours in the GUI.

However, Google Maps needs your browser to be able to play a few tricks and not having scripts will break it. Also not having an as up-to-date as required JavaScript interpreter will also break Google Maps. There are a lot of different browsers out there and as this relies upon more than just standardised HTML to run on, the number of browsers that are supported is limited.

A plug-in is not the answer until Google Maps becomes stable (ie, gets out of 'beta') and you would essentially have to update the plug-in each time Google changed anything.

So, which browsers does it support?

If you look at http://maps.google.co.uk/support/bin/answer.py?answer=16532, you will see that (at the time of writing) it says...

Proxy your web server - domain names on LAN

If you have a web server on your LAN and the server is visible from the Internet, you will have a route from your firewall to your server. If you are running Apache or some other, easy-to-configure web server, you might have multiple virtual hosts as well. But how do you make sure that they are all working properly?

External users connect to the Domain Name Server (DNS) which gives them the IP address of your firewall and the request is then routed to your server.

The problem is that if you use one of your domain names, your browser will ask the external DNS and you will get your firewall's external IP address and the process will most likely fail.

If you have your own, internal DNS, the solution is to have an entry in it with the address of the server linked to the domain names it uses.

However, if you have not got your own, internal DNS, you cannot do this. Also, putting the domain names in the 'hosts' file doesn't always work.

• DeMilitarised Zone - okay, I know that is 'DZ' but the Americans like TLAs (Three Letter Acronyms of which TLA is itself one).
  
  A DMZ is a third zone on a firewall (a tri-homed firewall - one with three network connections on the back) and in effect, it is where a server will end up being attacked if it is going to be. However, any attack on it will not intrinsically compromise the LAN as that is separate (you can make it compromise the LAN but you have to work at it, not merely work with it).

So, this is for people who either don't have a DNS of their own that they can play around with, don't have the time, or, are too lazy.

That way, you can type in the URL and the server will give back the page from the virtual server that you want.

In the second case (where the server is running on the DMZ), you would again look on 192.168.0.20 on port 80 although here, there would be no proxy for normal use (you can always have one on your LAN and run your web server on a DMZ).

However, the problem is that changing the browser configuration and then changing it back afterwards is quite a lot of hassle. If you just want to check that changes to links and file references work all right, you could do with a quicker, temporary solution - one that you don't have to undo afterwards.

You can see the full command line in the example on the right...

This is not the default virtual host on the server but using links like this checks that all of the file links work all right.

Expanding image density range

In theory, you can take an image of the moon that has been taken during daylight and cancel out the light from the sky - this is effectively just an extrapolation of what happens at night except that the amount of light from the sky is a lot less at night so you don't notice it.

In the image on the right, you will just about be able to make out the moon.

We can go to each layer (red, green and blue) one at a time and expand the left and right density markers so that they match the densities of each layer in turn.

The first thing you will notice (assuming that we are ignoring the contrail that has become apparent - East Midlands Airport was downwind of the place that this was taken from so aircraft were flying towards me) is that there is a lot of noise in the image. This is due to the fact that we have expanded such a short range of densities to cover the full range in this photograph. Another thing to notice is that we haven't cancelled out the blue sky either.

Let's look at where this noise is coming from and see if we can do anything about reducing the level of background 'blueness'.

However, you will notice that this is a gradient. The sky produces a gradient but taking a value away from the whole image is like taking a flat colour from the image.

So, let's try taking values and then subtracting gradients from the image.

The reason I have used full range gradients is so that I can see where they are.

You can see how narrow they are by looking at the output limits (166 to 169, ie it has a range of four values).

Now, the image has all of the right amounts taken off it (well, as best we can do considering the image in the first place). You can see that it looks black.

Note that here, there is still a gradient of background noise although this appears to be a step - this happens when you have a gradient over a small density range. One way of getting rid of this is to start off by expanding the density range of the image so that it is a lot wider than the original. If you do this, when you subtract a gradient from it, you will have more steps and if they are fine enough, you will not see any steps (or at least, not very easily) when you put the gradient in. This is an artefact of only having 8 bits per layer per pixel - if it was 16 and RAW, there would be no JPEG artefacts and the lower limit to the dynamic range would be defined by the sensor instead of the image file format used (remember that these cameras are designed so that people with little understanding of computers can process the images on a Windows PC).

Using 'time' on Linux

It is possible to find out what proportion of the system's resources is used by a particular program by using top.

Here, we are looking at ksirtet which is a tetris game for KDE.

We can see it (PID 22146) and that it has used a total of 2.59 seconds of time. However, top give you the data from the process as it runs. We can get more information about the process once it has finished.

Once ksirtet has finished, time prints out the total elapsed time that the program has run from when it started to when it finished (in this case, 393.62 seconds or 6m33.62s), the amount of user time (2.44s) and the amount of system time (0.34s). If you add up these last two times, you get 2.78s which is reflected in the time that top came up with when it was running.

'top' has the advantage that you can see what a program is doing in real time but it has the disadvantage that you cannot interrogate the system once a process has finished. On the other hand, 'time' only tells you what happened once the process has finished.

On the right, you can see the output of a game of tetris without the -p switch and another game with it.

Removing motion blur at source

There are many different effects that you can use to 'enhance' a photograph but the best photographs are those that were good to start with - even before the image was consigned to the camera's memory.

Any way, you might take a small tripod with you like the one on the right so that if the light levels are low, you can at least have some chance of taking a photograph without any great amount of blur.

The image on the right is some bird seed taken close up in low light levels. As a result, the camera is still moving as I take my finger off the shutter release (maybe 'push button' would be more appropriate as there is no shutter to release).

The seeds are around 1.5mm across and you can see the that they are in fact quite shiny.

System monitoring on Windows 98

If you are going to use your Windows 98 box for network- or file-storage-related activities, you might want to keep an eye on what it is up to. You can do this using its System Monitor program.

Look under 'Start', 'Programs', 'Accessories', 'System Tools', 'System Monitor' or similar - you might well have moved it since you got the machine.

If you cannot find it there and have no idea if it the link is even on the machine, look under 'C:Windows\' for 'Sysmon.exe'. When you have located it, right-click on it and select 'Create Shortcut'. Next, drag the shortcut into your 'Start' menu and drop it where you want it to go.

Click on 'Edit', 'Add Item...'

Click on the categories on the left and the items will appear in the list on the right.

You can highlight more than one item at a time by holding down [Ctrl] whilst clicking on the items. However, this only works one category at a time.

Once you have finished, click on 'OK'.

There are plenty of controls to play around with and you will soon get a feel for which ones are of use to you.

You can alter the update interval quite easily by clicking on 'Options', 'Chart...'. The slider takes you through 1, 3, 5, 10, 30 seconds, 1, 5, 10 minutes and 1 hour.

When you close the program, the items that you selected will be saved and next time you load the program, they will be there again.

Adding a program to the panel in KDE

Sometimes you cannot find just the executable you want on the KDE panel. Maybe it's because it is hidden away and you don't know exactly where it is, or maybe it is a program that you have installed or have written yourself. Any way, this is what you need to do to add a program icon to the panel in KDE...

Click on 'Panel Menu', 'Add', 'Special Button', 'Non-KDE Application'.

Files are usually in /bin, /sbin, /usr/bin or /usr/sbin. You can also find them in your home directory - /home/paul/bin for example.

Pick your file and click on 'OK'

If you have one of your own, you can point it at that. You can make one quite easily with the GIMP and remember that you can make them transparent. Just look at those already on the system.

Here it is with icon zooming enabled. If this is not quite the right place for you, you can easily move it.

You can now slide it along the panel until it is where you want it. Note that you don't have to keep the mouse cursor on the panel to do this as the icon tracks the mouse movements.

Once you have it in the correct place, just click the mouse - again, this does not have to be on the panel (if you have panel hiding enabled, the panel might disappear but clicking the mouse still moves it in the same way that it would have if the panel was visible - just move the mouse where the panel is so that it becomes visible again as you would if you were using it normally).

'vi' survival guide

Usually, the only time you will need to know how to edit a text file using a text-only interface will be:

• if you can only access your machine via a terminal (such as if it is a headless server);
• if you have installed a new monitor or replacement graphics card and have no GUI (for some reason you cannot restart X); or,
• if you have rebooted after taking out a drive that is checked automatically and have been dumped at the command prompt (because the system can't find a drive that it thinks it needs) and need to edit /etc/fstab or a similar file.

Whilst vi is quite complex and there are other programs (see below) vi is one that is very likely to be on a system and whilst it can be daunting the first time you use it, for simple text editing, there are only a few commands in vi that you need to know.

Ideally, you should have a play around with vi and get used to the way it works. If the first time you use it is in a real situation, you will be at a disadvantage.

Starting vi

So, suppose you need to remove some lines from the file system table. Log in and then type

vi /etc/fstab

It doesn't matter whether you do this on the machine itself or on an ssh terminal over the network although you will need to be root to edit /etc/fstab. For your convenience, I've included a fstab file here for you to play around with so that you can get some idea of what you need to do to navigate and edit using a text editor that is designed so that it will work over even the simplest of screen-based connections. Just click here to open up the directory in a new window

You should now see your fstab file similar to the one on the right.

Different Modes

For the reason that it needs to be compatible with situations where arrow keys and some other keys are not accessible, some functions need to use normal character keys. As the program needs to know whether you mean to insert a letter 'h' or move one character to the left, it needs a way of telling the difference and it does this by using two different working modes: command mode and insert mode.

Command Mode

Navigate around by pressing the left, down, up and right arrow keys but, should you find yourself on a terminal that doesn't support them, use 'h', 'j', 'k' and 'l' respectively. The 'H' and 'L' keys ('h' character and 'l' character) speak for themselves as they are at the left and right extremes of the four keys. As for which is up and which is down, think of the 'J' ('j' character) as pointing down and you will be all right.

Note that like UNIX, vi is case-sensitive so whilst 'l' takes you one character to the right, 'L' takes you to the beginning of the last line of the file.

To delete a character, just press the [Delete] key or 'x': to delete a complete line, press 'dd'.

Changing Mode

To change from command mode, to insert mode, press 'i' or 's'. Now, each character you type will be inserted where the cursor is, just like in KWrite or even Notepad. To get out of insert mode and back into command mode, press the [Escape] key.

Insert Mode

To insert a new line from command mode, press 'o' and a new line is inserted after the cursor -- you will also be in insert mode ready to start typing. If you were already in insert mode, just press [Enter] for a new line. To replace a line, press 'C' from command mode and then type away until you press [Escape].

So, now, we can insert and delete lines and characters and all we need to do is save and quit.

Saving and/or Quitting

To save, get into command mode and then press ':'. The cursor will move to the bottom. Now type 'w' and vi will save the file, overwriting the original.

If you want to save it as, say, 'newfile.txt', just type ':w newfile.txt'.

To quit without saving, type ':q!'. The exclamation mark forces the quit without saving first. You'll need to do this to rescue yourself from a situation where you make a monumental mistake from which you cannot recover but you haven't yet saved. In this case, just quit out of vi and then go back into it again.

Running PICO and PINE

Pico and Pine are two handy little programs that originate from the University of Washington (in the USA, not the original Washington in the UK).

Pine allows you to create and read email and Pico allows you to edit text files. Both are lightweight and will work via a remote, text-only terminal.

Pine

Pine is a small application that allows you to access your email remotely. You can work your way through your email directories, reading them as you would with any mail reader.

To send an email with pine, at the command line, type something like...

pine paul@bloggs.com

If you take the cursor down to the message text, you can edit it but unlike sendmail, you can go back through it and change other lines as well as the header information.

Once you have finished, just use the control functions to send it.

This presents the menu on the right and will allow you to configure it for sending and receiving emails.

The program is fairly easy to navigate although there is a lot that can be configured if you are going to need to use this as your only mail program.

You can see why HTML email is deprecated.

Pico

Pico is a simple, lightweight text editor and if you have this installed, you can use that to edit /etc/fstab file if you want.

Remember that Pine and Pico both use the control key for commands whereas vi does not.

I have tried Pine and Pico over an SSL link from Windows using PuTTY and there are no problems with the way that the control key is transmitted. However, if there is a possibility of ending up with a link that will only use unmodified keyboard characters (ie, not [Ctrl], [Alt], [Meta], [Win/Tux] and so on) then you need to learn how to use the very basics of 'vi' (see above).