Michael Feldman wrote:
>
> Someone recently mentioned a piece on GNAT in Dr. Dobbs, December
> issue. I'm having trouble finding it here in the stores. Anyone got
> a copy?

I have a copy of the article here.  December is the Object Oriented
Programming focus issue.  The article is by Gavin Smyth, who I do not
know.  While I understand the editorial difficulties of getting
comprehensive material published, and am glad to see an Ada article in
DDJ, I am somewhat disappointed with what was published.  No offense to
the author, I just think Ada deserves better.  Since I couldn't do
better, I hesitated to give my observations, but maybe I'm the
demographic we should be after.  (C and to a lesser degree C++
programmer interested in software engineering issues and predictable
development processes).

For what it's worth, here's the top things that struck me about the
article:

- Never mentions that Ada 95 is an ISO standardized, object oriented
language (this is the OOP issue, after all)

- Refers readers to a comparison of C++, Ada, and Modula-3 from 1992 --
before Ada 95 was standardized.  Don't know what that says, but it
probably isn't relevant or flattering.

- Focuses on DJGPP and makes extensive use of DJGPP features
(portability???)

- Says you ought to get the DJGPP binary file compressor -- GNU execs
tend to be bulky (resources???)

- "For brevity, I have not bothered to make the code very efficient..."
-- I know the answer, but "Ada's not efficient" jumps into my mind

- Discusses, at some length, how Ada 95 tasking doesn't work under dos.
(In it's favor, does later say there isn't such a problem under other
OS's, but still struck me that he discussed the use of Ada 83 tasking
facilities here -- if Ada 95's so good, why's he using Ada 83?)

I know the answer and the truth of all these issues and understand that
it may, in fact, make great sense to use Ada 83's tasking facilities
under DOS.

All I kept thinking as I read the article, though, was "couldn't we have
found a more flattering environment to write about Ada 95 for the OOP
issue?".  If I'm a member of the list and believe in Ada and was struck
with these thoughts, how many C/C++ heads just skipped on to the next
article?

Again, I'm not trying to insult or flame Gavin Smyth.  Everything in the
article was accurate and truthful to the best of my knowledge.  I was
just disappointed that for the first piece on GNAT in DDJ, it didn't
send the message that Ada's relevant, modern, object oriented, stable,
portable, _and_ reliable.

One man's point of view...
Scott

--
Scott Renfro              Lawton      Suffering is inevitable.
[log in to unmask]         OK        Misery is optional.