richard bowers wrote:
> 
> I personally feel that a database, for whatever purpose, would be the wrong
> place to document 'errors'.

Actually, my idea wasn't just to document known errors, but to document
known non-errors, IOW to document that someone somewhere is sucessfully
using a certain ugen, so that someone can feel confident in using a
certain ugen, knowing that there's a chance that there's no major
problems with it. I understand that just because a ugen worked correctly
for some person in some orc, doesn't mean that there aren't bugs that
wills show up in a different usage. But it's a start.

> I think Larry Troxler's original concern over the status of the ugens for a
> new user of csound may be better addressed by a thoroughgoing review of the
> manual: 
>
> I feel that some of the examples may be added to and the consistent
> format expanded to allow for other common features (eg. default settings may
> be explicitly stated - I can't think of anything else offhand - but you get
> the idea?). It may be good to organise the sections differently etc.

This of course isn't a bad idea, but it's orthogonal to the problem I
was addressing.
Yes, maybe a "KNOWN BUGS" section for each ugen would be helpfull. Of
course, the bigger problem is UNKNOWN bugs, which if we had some idea of
how widely various opcodes are used, could have a better feeling if
we're in for another night of debugging code in attempting to use them.
 
I think one first basic step forward, would be a simple mechanical one:
Provide an archive of the mailing list somewhere. The only archive I've
been able to find is an on-line interactive one, which given the volume
on this list I find impractical to deal with. For me, it would be more
usefull to be able to download a years worth of messages to grep in the
comfort of my own home. I don't care if this means downloading a 100MB
file. The added convenience would be well worth it.

Larry Troxler