I've found that software that relies on sound is problematic. Problems include:

* Different software uses different levels, making one have to adjust their volume up, down, or off such that its current state is unpredictable.

* Some sounds get annoying after hearing them over and over such that they are not wanted anymore.

* Sometimes it outright does not work, and becomes yet another aspect of desktops to troubleshoot.

* There are too many "levels" of volume control: the physical speaker knobs, the OSS settings, and individual applications and/or players. Many users don't know or don't bother to check or understand these. There should be one and only one master control, or at least one "control panel" even if app or vendor level control is possible.

----
''(DeleteWhenRead: This was originally C''''''omputerSoundUseless.  I added a comment questioning whether the above amounted to uselessness, but someone else felt more strongly about it and moved the page to C''''''omputerSoundMildlyInconvenient, which seems like a minimization of the original poster's (valid) complaints. So I've moved it to the more-neutral ComputerAudioUsability, partly because the original page name looks too much like 'Computers Sound Useless'.  Some days I'd agree with even that sentiment, though... --TimLesher''
----
Talking of relying on sound... did you know that WindowsVista has SpeechRecognition?
* http://www.youtube.com/watch?v=tLa3Wac4O2A - DemoMeltdown in full effect
* http://www.youtube.com/watch?v=Uy9dRkCzRRQ - behaving a little better
* http://www.youtube.com/watch?v=KyLqUf4cdwc - on the other hand, not so good for PerlLanguage hacking.