Mail Archives: geda-user/2015/07/08/09:51:08
On Wed, 8 Jul 2015, John Doty wrote:
>
> On Jul 8, 2015, at 6:50 AM, Bob Paddock (graceindustries AT gmail DOT com) [via geda-user AT delorie DOT com] <geda-user AT delorie DOT com> wrote:
<snip>
>
>> If there is not a language to that problem
>> domain, maybe it needs created.
>
> *Lots* of high-level languages would work well for the high-level parts. However, most would have the problem of unfamiliarity that Scheme has.
And there is another factor to this. Does anyone get paid for hacking
gEDA? If not, what's the motivation?
For me, and probably for many other "open source guys", the main
motivation is having fun. There are usually many more projects out there
to work on. All of them are useful and important. If one project doesn't
reward with fun, for whatever reason, I just switch to another. The
reason sometimes is related to the language and how efficient I can be
using the given language. This how the language becomes the barrier: if I
am struggling to get even the simplest things working with seemingly zero
benefits, it's no fun.
If is sit down to hack gnetlist and I instantly bump into scheme and then
I struggle getting the simplest things done... If a "copy an exsiting
backend and modify the text format it prints" kind of should-be-20-minutes
job spans accross a weekend... You can blame me, but I did not have fun
at all.
Probably it's not scheme's fault. Maybe I'm jut not intelligent
enough to enjoy scheme.
Anyway, my point is still this: in theory any serious language could be
good enough for netlisting. In theory some languages would offer features
that make some parts of the job easier. In practice I don't really believe
any particular language could make such a difference that we could pick it
and say "wow, this is clearly the best language for netlisting". I don't
code python or ruby, but I'm pretty sure the same task would have taken me
at most 1/3 of my time if gnetlist happened to be coded in them.
This is a totally different aspect: if we want contributors, and want
people to join, we can't ignore the fun factor. It's not a technical
property of the language. It is not how people should choose a language.
It is about how people actually do, and examples from the past decade
don't make scheme's position very strong.
Don't misunderstand me, I have several little pet projects where I go for
something totally unusual. For example my favorite programming language is
AWK, which is probably not any more popular than scheme is. I do have some
projects which are heavily based on AWK, or are mixtures of C and AWK. I
just admint that I won't get any contributors (or even users...) in those
projects.
Regards,
Igor2
- Raw text -