Re: Is Linux right for Embedded?

["Marin David Condic" <dont.bother.mcondic.auntie.spam@[acm.org>]

I don't think anyone will contend that there is never a time when Ada might
be less productive or a less attractive choice than C - it depends on the
project and the conditions surrounding it. (For example, some small SBC's
only have C available for them.) If you have a bunch of C "coders" who don't
understand principles of software engineering and/or are unwilling/unable to
do a good job of constructing software, you may end up in the situation you
describe. Chances are, you'll end up in trouble for hiring "coders" instead
of "software engineers" anyway.

As for emulators, debuggers, etc., consider that all of this stuff is
constantly changing with hardware and development tools anyway and has
little or nothing to do with the language of implementation. (BTW: The Gnat
Ada compiler has the gdb debugger with it, so its basically the same toolset
you'll see with gcc - a very popular C compiler that is used for a lot of
realtime/embedded work.) I suppose if you want to go looking for reasons why
Ada is going to cost you a bundle, you'll be able to imagine lots of things
that might never really be a problem. If you want to look for the best tool
for the job at hand and are willing to keep an open mind, you might find
reasons why a new language might be advantageous. People change languages
for different projects all the time - why is it A Good Thing(tm) if the
language is Java (or insert the Language Du Jour here) but A Bad Thing(tm)
if the language is Ada?

BTW: Just for future reference, its "Ada" (named after Agusta Ada Byron) not
"ADA" (as in Americans with Disabilities Act, et alia.)

As for hiring "coders" - my general rule is this: Hire smart people who
understand software development and are capable of learning new things. Give
them the right tools for the job and the right training to use those tools.
Keep them happy and keep them around. (Turnover is expensive!) You'll end up
with better products built more efficiently that way. If your job ads read
"I need Microsoft Visual C++/IBM-PC/Windows95 programmers..." what will you
do with them if Linux becomes the hot craze? Or if the architecture changes
to Sun/Unix? Or (God Forbid!) some *other* language becomes the hot item
this week? Makes more sense to me to find adaptable engineers, 'cause the
world of computers is *always* changing.

MDC
--
Marin David Condic
Senior Software Engineer
Pace Micro Technology Americas    www.pacemicro.com
Enabling the digital revolution
e-Mail:    marin.condic@pacemicro.com
Web:      http://www.mcondic.com/


"bgeer" <bgeer@xmission.xmission.com> wrote in message
news:9pfcps$p0l$1@xmission.xmission.com...
>
> Time to learning basic ADA isn't the issue.
>
> Dump several 10s of thousands of lines of code in hundreds of source
> source files, all poorly documented of course since ADA is
> self-documenting, on any but the best memorizers & you end up with
> long delays in productivity.
>
> Add to that time to figure out emulators & debuggers for which manuals
> can't be found, long since carted off by a former project member.
>
> Now tell me finding good, quick-to-produce ADA coders, is easier than
> finding good C coders.
>