Why I hate using windows
Aaron Sloman
Last updated: 15 Jul 2010; 26 Jan 2013
Installed: 14 Jun 2010

My work is all done on linux, but I sometimes use windows (e.g. helping my wife, who has to use it) ... and hate it.
This file lists the reasons why I hate it. (It was previously part of my home page.)

• After I removed OpenOffice and replaced it with LibreOffice on my wife's machine running XP, it proved impossible to set up the right file associations for LibreOffice until I discovered that the Windows program remover leaves a mess behind and a registry cleaner may be required to fix the mess, which I eventually managed to do as explained here.

ADDED 15 Jul 2010: Restrictive selling practices

• I should have added this earlier: too often if you wish to buy a nice piece of computer hardware (like the Dell Latitude E6410 I purchased in June 2010) you are forced to pay for a windows licence that you don't want in order to get the machine you do want.

  The upshot is that people who don't want to use Windows, e.g. lots of linux users, are forced to subsidise those who do.

  This practice should be illegal. Perhaps it is -- and the law is ignored?

• The need for virtual desktops
  How many times have you seen a powerpoint user have to temporarily minimise powerpoint in order to open up something else during a presentation? I cannot understand how all those users (even Professors of Computer Science) have allowed Microsoft to go on providing a system that does not support multiple virtual desktops - which I used for several years on Suns running Solaris, from before 1990, and have been using since 1999 on Linux, both on my desktop PC and on laptops.

  I used to use four virtual desktops at a time at first, then later switched to six. Then for a few years I switched to eight. Now I regularly have ten (using the CTWM window manager for many years, until I recently switched to using OpenBox instead as described here).

• Moreover, I don't have to fiddle about with a mouse in order to switch between virtual desktops, or even between windows. Both Ctwm and Openbox (and other unix/linux window managers) allow me to to cycle through virtual desktops, switching in a fraction of a second using the keyboard -- though I can use the mouse, which is a bit slower).

  So when I need to give a presentation I leave the workspaces dedicated to my long term current tasks (e.g. browser open with frequently accessed web sites, editor instantiations for papers I am working on, PDF reader instances with papers I am currently reading) and set up my presentation (prepared using latex and shown using xdvi or xpdf), and in other desktops set up other things I want to show, e.g. videos, software demos, images. Then during my talk I can flip to the required workspace and instantly run whatever I have set up, and then instantly return.

  If I want to give two or three demos each of which requires several windows to be open, I can prepare a virtual desktop for each of them in advance and flip to them when needed. This is helped by the fact that I don't use powerpoint for presentations, so I don't have to go in and out of 'full screen' mode. Instead I use a PDF viewer or xdvi on the output of latex.

• My wife has to use windows, because she uses the wonderful orienteering map-making package, OCAD. But watching all the hassle she has to go through because of the crummy user interface MSWindows inflicts on her is very sad. She can't even push most of a window temporarily upwards off the screen leaving just the bit she wants to see exposed while she does some work in another window. And she can't type into a small part of a partly covered window without making that window come up and cover everything else. I've had these options on Sun workstations and Linux machines for years. (Input focus follows mouse without raising the window automatically. That should at least be an option on any sensible system.)

• Other things that make me hate having to use MSWindows to help my wife: panels that come up showing a tiny unexpandable window into a large file or large list of options so that you have to waste time scrolling -- as if we still used 800x600 screens. On linux/unix, almost all such panels (that I encounter) are stretchable, except those produced by foolish web site designers. (Maybe some of these problems on windows are the fault of application providers. Does microsoft make it easy for them to use stretchable as well as scrollable text panels?)

• Another stupidity is that when shown directory listings there appears to be no way to filter what is shown by pattern as in

      ls *smith*.txt
  

  in unix/linux, and various other operating systems, which we have been able to do since long before Windows was even invented.

  However, I have noticed that the Linux developers who try to make Linux look and behave like Windows, in the interests of winning converts often fail to use the power of the Unix mechanisms in Linux. E.g. as far as I know, the standard graphical file browsers in Linux (e.g. nautilus, pcmanfm) go slightly beyond windows explorer in allowing pattern matching on the first character of a file or directory name, but do not allow wild cards, e.g. 'a*.p' to get all files starting with 'a' and ending '.p'. I find that exclusion daft, given the powerful regular expression matchers available in Linux.

  I used that in my 'toy' pop11 based file browser announced as long ago as 1999:
  http://www.cs.bham.ac.uk/research/projects/poplog/figs/rclib/rclib-browser.html

• I have been using symbolic links for files very fruitfully for many years on unix (since the 1980s). This wonderful invention allows a unix directories to share subdirectories, and has many powerful uses.

  As far as I can tell, symbolic links are still not available in recent versions of windows.

  For Windows users who have no idea what I am talking about look here: http://en.wikipedia.org/wiki/Symbolic_link. Shortcuts are a poor substitute, for the reasons given there.

• On unix, as on earlier systems, the Control key was close to the little finger, i.e. next to the "A" key, and the CapsLock key, when that became available, was down at the bottom, rightly because for many people it was used a lot less often, and more importantly, did not need to be held down at the same time as other keys. I suspect the position of the windows Control Key at the bottom of the column causing the wrist to have to be contorted for many combinations, e.g. Control+W, Control+Q, control+A is responsible for much repetitive strain injury.

• Not distinguishing devices from directories in file name syntax was a brilliant idea which I first encountered in Unix in the mid 1970s. So the fact that this directory tree starts off on a magnetic drive

      /places/towns/cities
  

  and then is later moved to a different device (e.g. a USB drive, or to a subdirectory of another directory)

      /things/places/towns/cities
  

  does not affect the syntax required for the name 'places' on a unix (and now also linux or Mac) file structure allows many things to be done elegantly. Even the terminal can be treated as a file /dev/tty and processes reading from and writing to any device can be conceptualised as reading from and writing to a file, so that the symbols '>' and '<' for writing to and reading from have a suitably general interpretation. Where a non-linear mode of access is required, that can be handled by software used, not the syntax of the path name.
• And then there was the totally unnecessary decision to use '\' as separator in path names in DOS and Windows, where '/' (a much older character in pre-computer usage) had been used previously. Why do people who are supposed to be clever do these things? The backslash was originally introduced to allow computers to represent things like '/\' and '\/' as complex symbols. Later backslash was used in the C programming language and Unix as an character used to transform the interpretation of the following character, e.g. putting \ before an apostrophe to prevent it being interpreted as the end of a string, or before itself, as in \\ to stop it being interpreted as an escape character. Then along came Microsoft, and IBM's failure to understand what they were doing.
• I can't push a window upwards beyond the screen boundary, so that a small portion is left for me to type into or read, as it scrolls, etc.
• In windows 7 the operating system keeps suddenly grabbing a window I am moving it and making it fill the whole screen. That could be useful sometimes, but the first time it happens there should be a warning panel asking whether the user wants it and if not explaining how to turn it off (it took me ages to find that option buried somewhere).
• Fror someone who uses the CTRL and ESC keys a lot their default location on PC keyboards (unlike old Sun keyboards) is a source strain and potential RSI. On linux I have been swapping Ctrl and CapsLock, and swapping ESC and Grave keys for many years. As far as I know windows does not directly support this. But I did eventually find a useful tool that allowed me to do it. SharpKeys

  I am amazed that the developers at Microsoft have not learnt the importance of giving users freedom to tailor keyboard maps. (It should be easier than it is in linux too.)

• I have never found a way to us 'focus follows mouse' on windows, so that you don't have to click on a window in order to type into it.
• I have found a way to allow a window that is partly covered to remain partly covered while I type or click something in it.

Is it ignorance, arrogance, bad management, sheer stupidity, lack of imagination, ..?

Or is the design of windows so non-modular that they cannot accommodate these features.
needs

For some hunches about what Microsoft may be doing see the last section of this file.

Maintained by Aaron Sloman
School of Computer Science
The University of Birmingham