Article: "Fixing linux" - opinions?

[David Gerard]

2008/12/12 P.B. <pb at fsfe.org>:
> simo wrote:

>> IT just shows that the audio subsystem is not mature, and no one of
>> the so far proposed and built one was good enoguh to make everybody
>> just switch to it.

> I think with Alsa, Jack & pulseaudio things should be fine. As far as
> I know, both pulseaudio and jack rely on alsa underneith, so there's
> no competition there. Jack is for low-latency and pulse is, let's say,
> for the desktop.
> Jack and Pulse and rather young, and there's a lot happening in those
> corners.


Sound on Linux has been a complete mess as far back as I can remember.
Every new solution claims to solve all the problems of the previous
solutions and provide backward compatibility, and fails on both.

e.g. at present sound in Wine on Ubuntu is a PITA, because
PulseAudio's ALSA support is lacking and no-one's coming forward to
write an entire new PulseAudio backend for Wine itself. The obvious
answer is to fix the ALSA support in Pulse, but this appears to be a
Simple Matter Of Programming, i.e. not that easy. Result: no sound in
Wine unless you disable Pulse and use ALSA instead, i.e. an imposition
on the end-user that should surely be fixed if at all possible.


> - database, apache and shell:
> If you're only dealing with English, you're lucky. There are several
> different environment variables in the right scopes to be set properly
> in order to get special characters working properly across such a
> system. Even if you use UTF8.
> I've done it and I know I'll have to search the web for it if I have
> to do it again, because even if you've been through this, it's tough
> to remember.


Yes, internationalisation is a mess. Too many apps get US-ASCII
working and everything else is an afterthought, even if the app is
UTF-8 clean.


> Files with umlauts can't be copied. One might think that if you're
> using UTF8 it's fine, but it's not. I know how to fix it, but I'd say
> it's a no-go for a computer novice.


This one's a very real problem for me too. It's silly to presume a
homogenous ext3-supporting environment.


> That's what I currently explain to them. Using some examples from the
> "real world" often helps, but I'd like some of the issues I've
> mentioned to be "reminders from the past", and not something which is
> here to stay. Things are already getting better.


Acknowledging broken things are broken is a first step. "Sorry, we
know it sucks, we're working hard on it."


- d.