Re: Incorporating 9X into Ada courses

[cis.ohio-state.edu!zaphod.mps.ohio-state.edu!uwm.edu!msuinfo!uchinews!iit max!thssamj@ucbvax.Berkeley.EDU (Aditya M. Jani)]

>*  Avoid the three technical problems facing new Ada users.

<I hope that one of these three problems is correctly recognized to be
<the non-existence of a powerful and flexible Ada implementation for the
<IBM-PC that is in seriously competitive with corresponding C/C++ offerings
<in any of the realms of price, optimization, resource hunger, interface
<capability, or support.

<--ben


>*  Avoid the three technical problems facing new Ada users.

<please explain to all of us, thank you.

<Regards,

<Peter Hermann



 I am forwarding the reply by Mr Fred Francl, who would be teaching
 the Ada course at IIT Chicago this summer.
 I will forward any replies to him.
---------------------
forwarded message
--------------------- 


Compiler problems have plagued Ada since at least 1984, when our group
used to measure compile time in fractions of a day. We used the now 
infamous ALS on a VAX 780. But as Paulkovich points out in the May/June
Ada Letters, C/C++ compilers will always be smaller and cheaper because 
they do so much less.

My concerns are very narrow for the purposes of my IIT Ada course. For 
academia, the GNU Ada/Ed translater being released this year is good 
enough. And its price (free) is certainly all right. It won't do you
much good on typical industry problems, or even advanced Ada courses.
But my focus is on introductory Ada work in defining the three technical 
problems FOR NEW ADA USERS.

I readily admit that any two Ada instructors will define wildly different
sets of "three technical problems." To try to at least stay in good company
I have taken two from Michael Feldman (Communications of the ACM, Nov 92, 
page 58). He cites "excessive reliance on global variables" and "deeply
nested procedures" as "unfortunate habits...learned from the monolithic
style of Pascal." As a software manager and, more recently, as a consultant
I have seen these problems over and over in the work of new Ada programmers.
The other big problem I have seen is "AdaTRAN."  To me these are the major 
defects in the products of new Ada programmers who are still treating Ada
as "just another programming language."

Your top three candidates are undoubtedly different because we have had
different experiences. If you have the time I would appreciate hearing 
what yours are. 

Sincerely,

Fred Francl