Mail Archives: geda-user/2015/12/30/21:30:23
John Griessen:
> On 12/30/2015 03:35 AM, Richard Hughes (hughsient AT gmail DOT com) [via geda-user AT delorie DOT com] wrote:
> > So, you're looking for the 0.01% of people crazy enough to do PCB
> > design, of which 0.1% of which will know the difference between free
> > software and shareware, of which 0.1% also knows a pretty esoteric
> > language
>
> Even though scheme is naturally good for big lists of data, other high level languages like python are also,
> and the popularity guesses above are probably not more than 10X off.
>
> Even just 1% * 1% * 1% is 1 in a million people,
> so it really does limit contributions.
>
> If there are several more lurkers like Richard Hughes that I cannot recall seeing post here before,
> maybe the fastest way forward is a base up rewrite in python, starting with Roland's xorn,
> with as little C as possible for performance reasons only,
> 8and with existing code maintained for a good while so comparison testing can be done.
> PCB is a repository of tons of layout knowledge -- you would keep it around and use in parallel
> for a couple of years to be sure you had all features implemented, none forgotten. gschem is not as many little rules,
> so it might transition sooner than 2 years. With python and a plugin system of code, from a bottom-up and
> top-down design and rewrite, I might even contribute. It seems like quicksand as is, so I cannot afford the time to.
If I'd choose, I'd have libgeda in c only, because that what I
know best.
That said, I'm somewhat surprised about the "we don't want scheme"
emphasis expressed in this mailing list.
Look at a in some ways similar project: http://www.lilypond.org/
They use c/c++/python/scheme with scripting in scheme only, and they
are successful at it.
1, So first point is: even if I'm not at ease with with or like
scheme, I do see that one can make great things with it.
2. Their users are not programmers, they are musicians and the like.
Why can they use scheme to such a great extent without any complaint
whatsoever, and not I (even though I "know how to program") ?
3, the experience of lilypond devs is that c/c++ is much harder to use
than scheme, and I get a impression that if they would start over
they would skip the c++ part in favour of scheme.
///
So looking at libgeda, does it really matter what language it is
written in (since most of you would not touch that code at all),
isn't the problem that you are missing a binding to "your" language ?
And speaking of binding, is it good to have binding to mult. languages
or is it better to focus on a single scripting language ?
Lilypond gains strenght by not having mult. bindings, every script is
in the same language.
If I'd write libgeda, I'd use c and scripting in perl - but I know
that that wouldn't suit everybody, or would all of you be fine with
that ?
Somehow this comes to my mind:
http://www.userfriendly.org/cartoons/archives/99mar/19990319.html
http://www.userfriendly.org/cartoons/archives/99mar/19990320.html
and why not:
http://www.userfriendly.org/cartoons/archives/99mar/19990325.html
Regards,
/Karl Hammar
-----------------------------------------------------------------------
Aspö Data
Lilla Aspö 148
S-742 94 Östhammar
Sweden
+46 173 140 57
- Raw text -