Fwd: [GAP Forum] openmath package on windows

[Yevgen Muntyan]

On Mon, November 6, 2006 8:39 am, Nilo de Roock wrote:
> Thanks, I think that should do the trick. But cgywin is like getting a 2nd
> pc with Linux, or installing vmware with some linux. It's not what I was
> looking for. More sort of some tips on how to get the code compiled on
> windows and included in gap.

GAP on windows *is* a cygwin application, it's built on/with cygwin and uses
cygwin1.dll (it's in the bin/ folder next to the GAP executable). You need
full-blown cygwin installation only if you want functionality provided
by third-party libraries ("not by OS"), or if you need third-party
applications. It's likely not the case with that "gpipe" program.

> If GAP supports windows then why is not every package available on
> windows?
> Why are packages accepted with documentation saying "neither tried nor
> tested on windows"? I don't think there is a technical reason for why this
> package doesn't work on windows except that the developer didn't want to
> develop for Windows.

Imagine a person who is working on unix-like system. He won't do win32 stuff
if it's not really needed. And it's not needed until someone needs it. And
even then, win32 is hard, it's way different from unix. And he needs to
*buy* windows system for many dollars, and he needs to spend lot of time to
setup development environment, and so on. (This was a scenario describing
why some package wouldn't be ported on windows, things are often better than
that)

> These packages belong in an incubator and not in a
> final distribution imho.

My personal opinion, not of those who decide what belongs to distribution
and what doesn't: if a package works on unix, and if it does useful job,
then it should be distributed even if it doesn't work and will never work on
windows. GAP is not about absolute portability (though GAP developers did
great job to ensure maximal portability), it's about results. Groups,
algorithms, problems, stuff like that :)

Best regards,
Yevgen