22 Feb
2004
22 Feb
'04
12:43 a.m.
On Sun, 22 Feb 2004 12:42:31 +0900
Patrick Shirkey <pshirkey(a)boosthardware.com> wrote:
Sorry Patrick, but I tend to agree with Peter. The projects that
actually make progress are the ones where one or more people sit down
and start cutting code with very little planning. That is the same
for small projects like Peter's Specimen and larger projects like
JACK, Ardour, ecasound or anything else.
In the case of JACK, there was a whole bunch of talk on the LAD list
about how it should be done. Paul contibuted to that, but then went
away and wrote an implementation. Once there was something that at
least worked, others joined the effort and Paul is now one of about
5 or 6 people contributing to it. I strongly believe that without
Paul's implementation, we would still be discussing it on this list
and there would be no working code.
Another case in point is libsample:
http://libsample.sourceforge.net/
I told Dominic Mazzonni (Audacity) that I was working on a sample
rate converter quite early on and he suggested that I join the
libsample effort which was also working on doing sample rate
conversion. I looked at the web page and saw that they had
the web page, a header file, a mailing list (with zero traffic)
and some ideas along the lines of "lets use the JOS resample code".
When I looked the page, the project had already been dormant for
a number of months. I decided to press ahead with my own efforts
which resulted in Secret Rabbit Code being released in Nov 2002.
There has been zero progress on libsample since May of 2002.
In the Free Software world code talks and bullshit walks. Working
code, however simplistic, is worth 1000000 web pages with project
plans. My guess is that if Pete wanted to take Specimen into more
advanced areas he will get there long before the LinuxSampler
project and along the way he will have a useful tool that he and
others enjoy.
Erik
--
+-----------------------------------------------------------+
Erik de Castro Lopo nospam(a)mega-nerd.com (Yes it's valid)
+-----------------------------------------------------------+
Windows NT : An evolutionary dead end.
Pete Bessman wrote:
Fsck that.
They can come help Specimen, which is fully usable now, if
not feature rich. Or they can work on SimSam (simsam.sf.net). Not to
be a polemicist, but I predict this will go the way of Octal
(www.gnu.org/software/octal). Too much planning, too little product.
Failing that, when it does reach maturity in a few years, I'll just
whore the read-from-disk and network control code from 'em.
Nice attitude. While this ethic is understandable it doesn't help
progressions very much. Maybe now you've had a couple of days to think
rationally about it you have changed your mind anyway.