Re: Embedded Keynote Speaker Mentions Ada

[Cesar Rabak <crabak@acm.org>]

stephane richard escreveu:
 > "Cesar Rabak" <crabak@acm.org> wrote in message
[snipped]

 >>
 >> Although I see you intended as joke, the truth is that given a
 >> technology backed up by good marketing and adequate (healthier)
 >> financed companies will be considered first.
 >>
 >> Also, as the software engineers will have to know much more than
 >> just a programming language (Database, GUIs, Modelling, etc.), the
 >> available knowledge by the average working force is considered the
 >> appropriate choice.
 >>
 >> I insist I do not second these attitudes!
 >>
 >
 > Agreed, you're totally right about that.  And that's why most project
 >  management software come with a "extend deadline or hire more
 > resource to compensate but no matter what it's gonna cost more than
 > planned" feature. ;-).

Yes. It is a recurrent theme on consulting jobs the very immature
estimating capability of most organizations.

 > Indeed it is meant as a joke, but even without knowledge per se, Ada
 >  would probably lessen the use of the afore mentionned feature just
 > by the fact that there is less "surprises" in the development
 > process.  Also by the fact that unlike C/C++ and most other
 > "fashionable" language by the time the Ada code compiles, it works
 > whereas, in other languages,  by the time the code compiles, it
 > executes, but works?

Boy, I wish we could have hard and statiscally clean data on this point!
Unfortunately, when you go after the facts you notice the code written
in Ada and that performed well in this aspect had also sound
engineering, goood test plans so difficult to cling to Ada technology
only. . .

 >
 >
 >>> 2. I think that if any manager or any other person capable of
 >>> making a "which language" to use decision, if they took 5 minutes
 >>> to do proper searches, would see the "real" benefits of using ada
 >>> both as a programming language and as a "economical" solution for
 >>> the lesser time to debug only.
 >>>
 >>
 >> True. But it is hard to make a business plan on these grounds
 >> because the focus today is not in production of code but
 >> integration of already built packages.
 >>
 >> Even if a system were to be built from scratch, it is acknoledged
 >> coding takes about 10% to 20% of all project resources.
 >>
 >
 > Indeed for the case of "built from scratch" projects, I would
 > evaluate more t=like 20% to 40% if required for coding if not relying
 > on existing code base or libraries (assuming no one in the world ever
 > did anything close to the intended project).  but I do get your point
 > however, and to me it's a sad but true reality.  To me, this is not
 > take software engineering and software development for their true
 > meaning that htey were intended for.

I agree to your comment on SW engineering. I wish I could propose the
packages and components people has to integrate would have better
quality if written in Ada, and _being_ _heard_!

 > Now everyone wants someone to build their solution in their basement
 > in the words of "martin david convic" <- right name?  everyone thinks
 > that the "garden variety programmer" is all they need.  If only they
 > could see the alternate reality that they are missing.

Again. Do you think it is possible to gather some data to prove them
wrong? Other than that, we are thought as bringing just more 'anedoctal
evidence'. . .

 >
 >
 >>> 3. The only thing stopping them, in my book is none of the above.
 >>>  The availability of Ada developers just isn't as big as the
 >>> "popular" languages.
 >>
 >> Yes. In fact I dare to say Ada technology is 'invisible' for a big
 >> part of the IT industry.
 >>
 >
 > Based on your knowledge and experience.  Could you explain the why of
 > this?

I may try :-) I feel there are too many factors. Some of them already
discussed recently in this NG and with some fine comments. In order I
can bring some light to the discussion, I will discuss only 'general' or
'business' oriented IT, taking out of discussion the specialized fields
(like defense or medicine).

First of all, we have the problem of legacy and the castle built upon
the sand. To be able to explain this, lets get a simple example:

Some years ago there was in the help of Object Ada a sort of Petzold's
book examples of Windows programming in Ada. Once you're in this
environment, most of the Ada help (for producing more solid code)
becomes less visible while the programmer has to struggle with the
subtleties of the Win32 API.

 From a management perspective the ROI in going Ada is not positive. The
same can be said if we were on the Database binding, etc. See posts on
the GPS 'bugs' and we see it right now on a real project!

Second, academia (the three amigos on UML, for example) abandoned the
use of Ada in favor of C++ bringing a perceived vision of the "way to
go" for less technically oriented people. Now, let me ask you: if you go
to a newstand where fine IT magazines are sold what are the odds a
manager sees CUJ or VisualBasic or perhaps a .Net mag, and what are the
chances Ada is ever mentioned? And to finally blow the wollf dead most
of the 'Research' and 'Advisory' firms when mention Ada send the message
it is a 'niche' language or an 'ageing' technology not to be considered
for new developements. . .


 > Like you I do realize that what you say is indeed true.  But I like
 > to know why things are the way they are.

In a nutshell I would say: because most of the strategic decisions are
not made only in the basis of the merits of the technology, upper
managment on most of organizations may not even have the skills in this
area and the incremental benefits Ada brings to a project are diluted
into the rest of problems IT industry is plaged with (too soft
requirements analysis, featuritis, etc.).

 > And from one professional to another, would you make that same
 > decision, knowing Ada as you do?

I find hard to explain to my peers I have reason, although if I could
manage a project without having to 'negociate' this issue I will prefer
to use Ada and invest in the proper ammount of training of the personnel
so they started to think differently and not using it as 'just another'
language.

 > My speculation on this is simple, they choose what they choose
 > because they don't know any better, and don't have time to know
 > better.  To me, these are unacceptable reasons to base a decision on,
 > especially as far as a development project goes.  Yet it seems to
 > happen all the time, why do think that is?

Because we are tied to a technical echosystem: major players 'help' you
if you go their way. Why we could not avoid SAP created ABAP instead to
use Ada? Or why is it Microsoft swears they have a 'security' mentallity
and cannot ever _think_ of rewriting some critical software is plagued
by buffer overflows?

Even in a non commercial arena, why can't we convince Open Source
enthusiats to write the system software in Ada and start to make a
difference in this realm (my perception of their account on bugs is that
they have a better response time, not intrinsical better designed in
security)?

[snipped]

 >
 > Indeed, I know there are tools, I have a project on my website for
 > each of these features, I have UML to Ada generator somewhere, I have
 > CORBA integration bindings (PolyORB is a good example and there is
 > more). Ada2HTML producers and the likes, the key is none of htese
 > work with the others in an integrated environment.  the tools exist,
 > now we just have to make them work together.

I'll consider this a call to work on our community! And I humbly add *it
is a must* we get able to do so in order to show our technology really
helps to do such things!

On closing, I think we could try to make better marketing of the
strenghts of Ada going to the place where good wrtitten software counts 
more: infrastructure and components.

Strategically I think we should think to reduce the efforts on the 
'wrapper' (the more known name for our bindings) libraries and go for 
writting some fine libraries for others which stand for their robustness 
and security records.

--
Cesar Rabak