Re: Is Linux right for Embedded?

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

I was rather thinking about the case where someone develops software for a
product and the product lives for a relatively short time and the *next*
product that comes along reuses little to none of the existing code because
it is sufficiently "new" to warrant a whole new development. You might see
this in some kinds of consumer electronics products and some PC types of
apps, where basically a whole new look&feel needs to be developed every year
to 18 months. In effect, the developers are building throw-away code.

Granted, part of the reason this may be done is that the language of
implementation makes it sufficiently hard to maintain, enhance or reuse, so
it becomes more cost effective to pitch it and start over. Some of the
reason it might get done is simply to guarantee a difference with every
release. Some of the reason may be because the company wants to avoid the
costs involved in building systems that will hang around for a long time or
have big reuse factors. (You produce little to no design artifacts and have
to do little to no configuration management and certainly don't need to
produce much in the way of programmer documentation or provide for a lot of
customer problem reporting. If you can build the thing all over again from
bottom dead center at less cost than doing a *good* job, maybe you don't
need to do a *good* job?)

I'll easily concede that Ada buys you a lot for long-lived systems or
developing reusable code or any of the conditions that may keep what you
build around for five years. I'd just offer that even when you don't have
long-lifespan concerns, Ada can make a lot of sense from a reliability and
time-to-market perspective as well.

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/


"Ted Dennison" <dennison@telepath.com> wrote in message
news:jeKu7.14722$ev2.24082@www.newsranger.com...
>
> I wasn't talking specificaly about maintenance. Just because you have to
> *develop* quickly doesn't mean that people won't be trying to use that
software
> 5 years from now. It also doesn't mean that developers won't be trying to
use
> your sources 5 years from now. In a short-cycle environment reuse of old
code
> from the previous generation is vital to a project's success. What you
don't
> want in that enviroment is to find yourself stuck with a language that was
> designed to promote *itself*, rather than reuse.
>