Subject: Re: [linux-audio-dev] plug-in mania and the sound editor du jour
From: Juhana Sadeharju (kouhia_AT_nic.funet.fi)
Date: Fri Apr 07 2000 - 18:51:54 EEST
>From: Dave Phillips <dlphilp_AT_bright.net>
>
>it might be interesting to enumerate the Linux audio apps which already
>have plug-in systems. Here's what I found at a quick glance:
If somebody doesn't have anything to do, would it be nice to have
a document summarizing all those plug-in systems. Then it would be
easier to think what should be done for them --- how easy it would be
to add LADSPA support to them.
I have yet taken a new task of designing a free MIDI-controller for us,
and therefore cannot take yet another task.
> So then I decided to see how the most recent stack of soundfile
>editors was coming along. I decided to look at these apps:
I find these overview postings very helpful. It is good to know
how others are progressing.
> I also took a look at Xwave again. Juhana, if you're reading this
>message, perhaps I could convince you to release your improvements to
>Xwave.
Seriously, XWave code is a terrible mess. When I improved the XWave,
I had to quit because the code just wasn't meant to be extensible.
For example, adding a proper loop-play feature would have needed
a complete rewrite of the core.
I'm continuing my work on "Waver the Vaporware" (but I'm using it for
simple editing). I have progressed from last I told about it but
still it needs more work for the first release, or even for alpha
release.
> Martin Wilz's Kwave is abandoned. Too bad, it has a lot of nice
>features but needs some more attention to push it over the top.
The code looked a bit messy for my taste. It would need a complete
rewrite of the core.
> So I see a lot of duplicated and wasted effort. I've been observing
>the development of Linux audio software for a few years now, and I see a
>certain "project life cycle" at work. Ideas are easy, planning is easy,
>talk is easy, even getting a distributable prototype is easy. But
>building a solid application is not easy,
That is so sad. I know there are people who could turn my Waver to better
in a couple of days by inserting appropriate IPC/thread routines but
I want to think how to do it as programmer-friendly as possible, not
as a kludge with messy code.
Instead of categorizing software projects to Mud Balls etc., those people
should give us answer on how to improve the situation. Categorizing
project types may be good but completely pointless.
It doesn't help that some people seriously suggest that projects should be
"code hacking"-only, and suggest to quit if one doesn't fit for brainless
code writing orgies.
-*-
A couple of weeks ago, it was suggested that we should have LAD competitions.
Inspired by Jazz, next competition is held on most well-documented code!
Documentation should include (i) documented code, (ii) design/algorithm/
implementation docs, and (iii) user documentation. [ Jazz has barely (iii)
only. ]
I think two major reasons for why nobody take orphaned programs are
(i) no documentation and (ii) messy code.
Juhana
This archive was generated by hypermail 2b28 : Fri Apr 07 2000 - 19:56:28 EEST