Re: The future of Ada is at risk

[Gary Scott <garylscott@sbcglobal.net>]

Jeffrey R. Carter wrote:

> Phaedrus wrote:
> 
>> Okay, I guess we all agree, Ada is a heck of a language and we all 
>> love it. (Or else why would we be here?)  But the implied question 
>> that started this thread was, well, what's WRONG with the ol' gal?  
>> Instead of making this an Ada-love-fest, how about doing some honest 
>> soul-searching about what's missing, wrong, or just too hard for the 
>> programming world in general?   The current lack of a DOD mandate 
>> isn't the reason for Ada's diminishing market share, the market share 
>> is the result of a real or imagined problem with the language, 
>> environment or something else.  So I'd like to propose that we put all 
>> of your Mensa-caliber gray matter to work and determine what's wrong 
>> or lacking.
> 
> 
> Ada is a SW-engineering language. It enforces SW-engineering principles. 
> 98% of developers are coders, not SW engineers. It should not be a 
> surprise that they don't choose a SW-engineering language.
> 
> We don't let construction workers design bridges or choose the materials 
> for them, but we regularly let the construction workers design and 
> choose the materials for significant SW projects. This is probably 
> because there is significant liability for incompetent civil engineers, 
> and none for incompetent SW engineers.
> 
>> 1.  It's relatively intimidating to learn.
>>
>>    Not that it needs to be, but between a needlessly complicated LRM 
>> and academia-bound (-and-hopefully-gagged) textbooks, a neophyte 
>> shouldn't dive in without a full scuba set.  It gets deep, fast.  And 
>> if you didn't get on the Ada-train in the mid 80's, then the learning 
>> curve is STEEP.  Nice folks like Dean Gonzales did their best to 
>> create annotated LRM's in the 80's, should that effort get going again?
> 
> 
> A controlled study by the US Military Academy found Ada to be a better 
> 1st language than Pascal, which was designed as a teaching language.
> 
>> 2.  More emphasis on cleverness than usefulness on rugged, reusable code.
>>
>>    Have you noticed how many folks would rather use a bunch of tasks, 
>> or do umpteen levels of inheritance, when something simple would work 
>> just fine? These ivory-tower things are real nice, but isn't it more 
>> important that it be able to be reused in LESS time than it took to 
>> write the darn thing?  Has the KISS rule been forgotten?  Honest, it's 
>> okay that package Hammer isn't completely safe for objects of type 
>> Toe, if you document it.  And don't even get me started on people who 
>> make their own types (AKA "Bobs_Integer") when the predefined type 
>> would work just fine.
> 
> 
> SW engineers can create useful abstractions, making things simpler. 
> Coders can't, and tend to create designs that make things more complex 
> than they need to be.
> 
> And don't get me started on people who use Integer when an 
> application-specific type would be a much better choice.

On the contrary, this more often than not obfuscates, particularly for 
future maintainance purposes.

> 
>> It really is quicker and easier to create small C apps for Windows 
>> (The prevalent environment) than Ada.  Why?  What's are your favorite 
>> gripes? What could be made easier?  Since a lot of large projects 
>> start life as little projects and little demos,  if we make it easier 
>> to make little projects then maybe the big projects will come our way.
> 
> 
> My experience is that it's easier to get something running in C than in 
> Ada. Getting something that seems to be correct tends to take longer in 
> C than in Ada, and getting something where I'm confident that it's 
> correct takes even longer.
> 
> Ada sometimes seems to avoid taking advantage of its opportunities. Ada 
> 95, with tasking improvements, came out just before Windows 95, the 1st 
> widely used OS with multitasking support. I suggested that we promote 
> Ada 95 as the language of choice for Windows 95, but that didn't happen. 
> Now multiprocessors are becoming common, and a language with good, 
> high-level support for parallelism should be the language of choice for 
> such systems. Maybe we can take advantage of this opportunity.
> 
> A standard, portable, Ada-oriented, easy to use for most basic tasks, 
> non-GPL GUI library would go a long way to making Ada more attractive. 
> One of Java's big attractions was its portable GUI library. It wasn't 
> perfect, but that wasn't a concern for most users.
> 


-- 

Gary Scott
mailto:garylscott@sbcglobal dot net

Fortran Library:  http://www.fortranlib.com

Support the Original G95 Project:  http://www.g95.org
-OR-
Support the GNU GFortran Project:  http://gcc.gnu.org/fortran/index.html

If you want to do the impossible, don't hire an expert because he knows 
it can't be done.

-- Henry Ford