Re: SmartEiffel open Edition - what would be the goals?

Patrick Mauritz wrote:
> I just looked at the changelog of se2, and there seem to be lots of
> changes (esp. in the compiler core) that look too useful to discard.

Anyone considering forking 1.1 should certainly look at the changelog to
see the enormous amount of work that has been done since then.

> as for your comment about "simplicity", the main issue with "complex"
> languages these days seems to be the class library ... finding your
> way through the library is quite a big job...

With Eiffel it's not just a matter of find your way through the
libraries; you also have to reconcile:

   1. which Eiffel compiler versions the library supports
   2. which back-end C compilers the library supports
   3. which other Eiffel libraries are dependencies, and what
       Eiffel compilers and C compilers they support
   4. which C libraries are dependencies, and what versions
       of those libraries are supported

Ease of use promotes use (as Apple Computer used to advertise). For
example, I'm sure that ePosix would be more widely used if it didn't
also depend on Gobo.

I'd like to see more tightly-focused, easily-usable libraries. A library
that wraps C should JUST wrap C; it shouldn't depend on other Eiffel
libraries. Pure Eiffel layers can be built on top of that. If people
want to make C streams look like Gobo streams, that can be done at the
Eiffel level with an adaptor - the C wrapper doesn't need to "know it
all".

Ease of installation, learning and use pays off.

Regards,
Roger
-- 
Roger Browne <[email protected]>