Re: Future with Ada

[Richard Riehle <richard@adaworks.com>]

Brian Rogoff wrote:

> Well, duh! These people must be pretty dumb. Overconstrain the problem and
> there are no solutions. Did they think about hiring people who don't know
> Ada? Ada is a fairly easy language to learn IMO, *far* easier than C++,
> especially if you're willing to omit concurrency.

We did train a core group within their company.   As I understand it, some
of their software is still written in Ada.   However, any new work is being
done in C++, for the reasons mentioned earlier.

As a resident of the aforementioned valley, there are a number of reasons

> I never saw a job solicitation from an SV company here, so it seems that
> your clients neglected this newsgroup. Doesn't seem like they tried too
> hard.

Au contraire.  The advertised in the SJ Mercury News, posted to this very
web site and posted to the Team-Ada Listserv.   Also, I reconmended some
people, but none of them was interested at the time.


> Hmmph. I took a job programming in OCaml, far less widely used than Ada.
> Why people worry about this is beyond me. I wouldn't want a job where
> the language du-jour was the primary considertaion. Been there, done that
> (with Java) and as Alex Stepanov said, it doesn't smell so good.

I have been consistent in my support for Ada for a long time.  My company,
AdaWorks, has been focusing almost entirely on Ada for about fourteen
years.   We kept going through lean times as well as through those that
were more lively.   Currently, I am engaged in an effort to increase interest
about Ada among a specific and carefully targeted group of future influencers.
I hope the fruits of this effort will result in more decisions to use the language
for more systems.

In my view,  Ada is the still the most appropriate language development
environment for DoD military software and any decision to use something
else does not represent careful  evaluation of the relative merits of the
competing alternatives.   I may be accused of C++ bashing, so be it.  I
really don't care.   I regard the choice of C++ for military software as
just plain bad.    It is a dangerous choice, and those who make it often
fail to realize just how bad it is when compared to the current Ada
standard.

> Hire willing programmers who don't know Ada and train them. It is *not*
> that hard! A good programmer will not be hung up on surface syntax and
> will be able to take much of their knowledge of C/C++/Java over to Ada.

Good advice.  However, it ignores the realities of the software marketplace,
the current state of awareness among managers, and the substantial anti-Ada
bias that pervades the Silicon Valley software community.   It is a bias that
originates, not in actual knowledge, but in a collection of minsinformation
that conveys a false view of Ada as a programming language.  This false
impression goes back a long way, beginning in the early eighties when the
DoD mandate required Ada for all new software even though the compilers
were not at production quality and the programmers were not adequately
trained to understand how to use it correctly.  We are still working to
overcome this history.

> Much of the interesting Ada work we see now is in the open source
> community. That's great, but there needs to be some more activity in the
> commercial world too.

Agreed.  And we also need to start using Ada for DoD weapons systems
again.   Of course, one of the best things that could happen for Ada is for
someone to create a suite of commercial software that competes with the
stuff that comes from Redmond, using Ada.   Sell lots of copies.  Make it
wildly popular.   Later announce that it was programmed in Ada, using
CLAW or GtkAda.

For example, the publishers of the new XP series of software have just
made it attractive, with their oppressive new licensing policies, for
someone to put a competing integrated office suite on the market, one
that runs on all operating systems (using something like GtkAda).   I
just participated in a project using GtkAda that, although quite specific
to a particular application, demonstates how easily one can get out from
under the domination of a software Mafia.

It would take funding, creative people, and extraordinary management,
but the current state of Ada technology is such that it is possible.

On a scale less grand,  we need some original products written in Ada
that attract a commercial following.  With all the brainpower that exists
in this forum, I find it odd that there are no entrepreneurs with the chutzpah
to take on this kind of project, no one with the imagination to conceive a
new product or improved version of an existing product.  When Ada succeeds
in the commercial marketplace, it will be because Ada enthusiasts have
started to build viable commercial products in the language.   Once that
happens.  Others will follow and corporations will take Ada seriously.

> FWIW, I've discussed Ada quite a bit with my manager and he acknowledges
> that it's a lot better than C, C++ or Java, from his POV. We use OCaml but
> if we had to do some C level work I'd have no problem doing it in Ada.
> There are a *lot* of managers and programmers who are willing to try new
> things. Thank goodness (really Schonberg, Dewar, and the rest of ACT)
> that there is a high quality, free Ada compiler so that prospective users
> can kick the tires instead of just reading!

I agree.   The work of ACT has made all the difference, even for those who
publish commercial compilers such as ICC, DDC-I, and OCS.   But not
everything can be free.   The ACT business model is unique and creative,
but few businesses can survive on that model.  RR Software sells a powerful
tool, CLAW, that is pretty much unrivaled for its completeness, the level
of craftsmanship that goes into it, and the dedication one sees in Ada
stalwart, Randy Brukhardt.    We need to be willing to support, with our
checkbooks, those who keep Ada alive through their day-to-day efforts
with commercially available products.

We also need to keep others in our community informed about what is
going on.   The only newsletter I receive on a regular basis from an Ada
company is from DDC-I.     Joyce Tokar and Jennifer Sanchez do a great
job of summarizing their current successes, sharing a little bit of technical
information, and informing about the current state of the community.   No
one else takes the trouble to do this.    As far as the Ada community is
concerned, nothing is going on.   There is no news of new projects, follow-on
work,  contracts won for Ada software,  new products released that
use Ada,   new tools, new compilers, etc.   There is an Ada consortium,
but it is a tightly-knit group that seems to assume no one outside their
immediate group is interested in what is happening with Ada.   They
never announce anything, never publish anything, never post anything
to this newsgroup, and never send anything like a newsletter to those
who continue to support Ada.

If Ada is not visible, even to its supporters, imagine how little visibility
it has among those who might benefit from it if they only knew it still was
an option.

Richard Riehle