Re: Increased Interest In Ada?

[Cesar Rabak <csrabak@uol.com.br>]

Marin David Condic wrote:
> > I disagree, but see below.
> >
> You disagree? Does this mean you know of a good Ada compiler targeted to an
> inexpensive SBC that would fit the description? Or are you saying that
> porting an Ada compiler to some SBC would be no big deal? In either case,
> I'd definitely like to challenge you to verify this through demonstration.
> :-)

OK. I'll elaborate!

I don't know of any good Ada compiler targeted to inexpensive SBC or
LCDS.

As already posted elsewhere in this thread, I second the position that
perhaps porting is not the principal problem, but rather to assemble all
the pieces, including a targeted tutorial (with exercises), etc.

So perhaps we agree in the factual observation, and my disagreeing is
more in the sizing of the opportunity.

> >
> Well the world has become a lot bigger than 8 bit microcontrollers. 

Agreed. But you see 32 bits processors abundant in kits designed for
instructional use?

>I am
> currently working with a box that has a MIPS processor and almost the whole
> system on a single chip. So a 32-bit processor able to control some
> physical/electrical devices from a single board at an inexpensive price is
> not at all out of the question. The problem is: Which One? 

This is an interesting question! It had to be inexpensive with abundant
(and perhaps free) documentation available, and if possible a chip which
people would feel it is worthwhile to expend time on it.

>If you are
> familiar with embedded systems, I'm sure you know that there just aren't
> thousands of Ada ports out there for popular boards/development kits. You
> need the compiler, plus a good, powerful linker & cross-target debugger
> along with probably some available libraries, bootstrap code, descent
> documentation of everything, etc. Saying "Well GNAT has a port to chip X
> available somewhere on the net..." is interesting, but if you don't have all
> the pieces pulled together into a nicely integrated package that works
> reliably, it wouldn't make a good student environment. (It's hard enough for
> the pros to figure out how to get this sort of thing to work - how much
> harder would it be for the neophytes? :-)

Second in full.

> 
> I'm always interested in hearing ideas on this topic if you have any.
> Thanks.

I think a way to reduce the 'initial' cost of a such project would be to
detect the 20% of board/kits which have the 80% of the "market" and have
the ports (with all the provisos above mentioned) funded.

On a second but necessarily concurrent front, we need to create a mind
share in users/instructors/industry about the use of Ada for embedded
systems. Otherwise, I feel the whole exercise will doomed.

Cesar