[linux-audio-user] some thoughts about Linux audio software documentation

[John Check]

On Friday 13 August 2004 06:33 pm, Erik Steffl wrote:
> John Check wrote:
> > On Thursday 12 August 2004 03:34 pm, Erik Steffl wrote:
> >>John Check wrote:
> >>...
> >>
> >>>One act gigging with this stuff is worth a dozen coders when it comes to
> >>>legitimizing the platform. There's so much potential with what's here
> >>>today that it blows my mind, but if it's "by geeks, for geeks" it really
> >>>limits were
> >>
> >>   not that long ago it wasn't even that. The sound/audio/music software
> >>in linux is improving rapidly. obviously, you get the by geeks for geeks
> >>stuff first because it cannot be any other way - it takes time to make
> >>the program stable enough to be usable by general public.
> >
> > Yup. There's a definite progression. I'm not unfamiliar with development
> > cycles, as far as does it _have_ to be that way, it's a debatable point.
>
>    not really, you will always have nothing, then something incomplete
> sort of usable and only after that there's something usable (if you're
> lucky:-)
>

The operative word is _always_. How long is such a thing tolerable?


>    developer could opt to not release unfinished stuff but then how
> would he get requirements?
>
> ...
>
> > There's is a tendency though, for free software developers to put some
> > things off because "We'll make it right for 1.0". Which in itself is
> > good, but we have a quality obsession (which is also good) whereby a 1.0
> > release really means something, unlike commercial software.
> > The upshot of this is as the code matures, decisions that were made early
> > on become hairier and hairier to address. Ironically, this can lead to
> > massive redesign and rewrites which pretty much wipes out most of the
> > accumulated documentation in terms of accuracy.
>
>    I don't remember a project (hobby or commercial) that would not go
> through major design changes. This is true for major project (with all
> the docs that osi 9000 (or whatever it is for software) requires) all
> the way to small (but non-trivial) scripts.
>
>    It is a problem but I don't think there is a way out, it's kinda like
> there's no perpetuum mobile, you always loose some energy (=docs always
> lag, requirements change, bugs are found etc.)
>
>    OTOH I guess now is a good time to focus on usability since the tools
> are on the verge of actually working (for general public, i.e. non
> programmers, non sysadmins)
>
>    I just find it somewhat not entirely appropriate to make it sound as
> if developers were ignoring users/documentation/user-interface etc.
> That's the whole mindset that I see expressed fairly often, don't mean
> to accuse you of accusing developers, actually your email was quite
> reasonable (so I am, to some extent, using your email as a vehicle to
> rant a bit:-)
>
> 	erik