[clean-list] Clean on Unix

[Brian Rogoff]

On Tue, 13 Mar 2001, Martin Wierich wrote:
> Brian Rogoff wrote:
> [... about Unix as a second class citizen with respectto Clean ...]
>
> Your impression is quite right. Our policy is that Windows and MacOS are our
> "main" platforms and the Unixes are not, (although we have by far more Unix
> users than MacOS users). 

I would have guessed this, and I'd guess that a large chunk of those Unix
users are Linux users, who tend to be more technically adept than Mac and 
Windows users. I'd also guess that if you have by far more Unix than MacOS 
users that you'd shift your priorities; sadly (to me) that doesn't appear
to be in the cards, so I guess I can tell my manager that we shouldn't 
invest effort in Clean.

> As you might know this means, that you have less libraries (ObjectIO) and 
> tools (profilers, optimizing linkers, CleanIDE) for the Unixes and you will 
> have to wait longer for new releases. 

Exactly my concern. 

> This situation is not going to change. The differences will become even
> bigger: With Clean 2.0 there will be no dynamic linking for Unix, but
> there will be dynamic linking for our two main platforms.

Well, there's a reason that new OCaml users are sprouting up all over the
place. There's not as much support for MacOS with OCaml but almost no 
complaints either. Installation and use of OCaml on Unixes is simple. 
I've been using it at work for over a year and while I think I'd find 
Clean competitive as a language, as soon as you factor in the Unix issue 
Clean comes up very short in comparison.

> >     IMO this situation is unfortunate, since Unixes appear to
> > dominate the scientific and engineering software markets, which,
> > while small, are more likely to appreciate the benefits of a language
> > like Clean.
> 
> I agree. If you ask us for reasons then you will hear that it's more work to
> support Unix as much as the main platforms.

Given that there are so many more Unix users, perhaps with open source
Clean 2.0 a way can be found to take advantage of our numbers to offload 
some of the porting effort? 

To answer Martijn Vervoort, I see a port of Object IO to GTK and the
GNOME environment as making the most sense at this time. I can volunteer 
to do some of the grunt work, since I am not familiar enough with the 
library (guess why not ;-) to take a lead position. 

To answer Simon Peyton-Jones cheeky suggestion with a similarly cheeky
one, perhaps he can assist in getting Clean working on Unix, since he's 
done such a great job with GHC and surely has the expertise in Clean too
;-). 

-- Brian