Re: Why isn't Ada as popular as C? [Was: C is better than Ada because...]

Re: Why isn't Ada as popular as C? [Was: C is better than Ada because...]

[Tim Behrendsen]

On Wed, 31 Jul 1996, John F. Bode wrote:
> In article <01bb7da5$ef97cf00$96ee6fcf@timhome2>, you write:

> Ada hasn't caught on outside of DOD-based applications for several
reasons.
> First and foremost, Ada's problem domain is large, mission-critical
> systems.  Using Ada to write a word processor is perceived (wrongly, IMO)
> to be overkill (although, with good design, you would get a very robust
> word processor).  It hasn't caught on for personal computer applications
> because the hardware has only recently caught up with the demands Ada
> development places on the host environment.  Hell, we had a mini
> (admittedly ancient) that could barely handle the load imposed by Ada's
> development environment.  Besides, running Ada on top of something like
DOS
> is like tying an F-14 engine to a Piper Cub.

Back in the Unix early days, it had the reputation for being a very slow
operating system, and its proponents would always come back with, "Hey!
Unix is not slow!  Get a *real* computer."  Meanwhile, of course, the PC
folks would laugh at them and get back to work.

Now, in hindsite, *both* were right.  Unix *was* slower than other
operating systems that were either 1) written in assembly, or 2) not as
feature-filled.  Nowadays, when computer power is much cheaper, the
overhead of the operating system is not as much of an issue, and
we don't worry about it anymore (much).

Hearing your statements above, this sounds a *lot* like the old Unix
arguments.  Is it possible that Ada really *is* slower, but the world
will ultimately figure out that Ada (or some other, call it, more robust
language) will be worth the overhead?

> Sociological factors play a role as well.  Ada is not a hacker's language
> -- you don't get the thrill of being truly obscure with Ada code.
> Programming in Ada requires a bit more forethought than programming in C
--
> you can't just rush in and bang out code.  Rapid Application Development
in
> Ada is a bit of an oxymoron, and RAD is the way many people are going.
> Finally, because so many people are familiar with C, there's not a lot of
> push to convert to Ada.  It's easier to move to C++ or Java, because they
> inherit much of their structure from C.  But then, C didn't catch on
> overnight either.  Remember, Real Programmers use FORTRAN.

I agree. Compatibility is everything, which is why Windows took over
the world, despite very real flaws in the design.

Based on your statements above, it makes me wonder if were in the
same state that we were in when Unix came on the scene; the world
was not quite ready for portable operating systems, and right now
the world is not quite ready for languages that aren't "close to the
metal" (Not to make assumptions about Ada that may not be necessarily
true).

Of course, when you look at C++ and the amount of dynamic memory
allocation it does, performance has begun to rear its head for that
language, too.

> I agree that no language will overcome poor design, but Ada provides more
> tools to support good design than C does.  Ada's typing rules are much,
> much stronger than C's.  Data abstraction is easier to accomplish.
> Encapsulation is easier to accomplish.  And, like C, Ada does allow you
to
> get down to the hardware level, letting you implement bit fields to your
> heart's content.  The fact that it isn't more popular is a shame, but not
> surprising.

Actually, you refer to the "environment" quite a bit; what is different
environmentally speaking compared to a C development environment?

Tim Behrendsen (tim@airshields.com)

Re: Why isn't Ada as popular as C? [Was: C is better than Ada because...]

[Mark McKinney]

1. Software metrics people help me out. I suspect that error rates are not 
linear with project size. If this is the case then as software grows more 
complex it will be more important to use a language that assists you in 
finding potential errors. Ada is a tool to be used to reduce errors. 

2. As advances in processor speed begin to slow down and the trend gears 
toward systems with multiple cheap processors a language that has tasking 
constructs will prove more valuable as well. Ada has laguage construct 
that define separate tasks.

3. C seems to be an unnecessary level between Assembly and a high level 
language. Like microcode was eliminated in RISC processors will eventually 
be eliminated as a method of developing complex software(True most of us 
may be retired by then). A larger leap at each level of abstraction in 
systems is needed to keep the layering from getting to deep(keep this in 
mind when developing objects) as a good balance between depth and width of 
design is required. IMHO Ada provides higher level abstractions of low 
level constructs than c.

4. C++ programming seems to require simultaneous programming at a low 
level(C) and at a high level(++). The OO features added in Ada95 seem to 
fit better with the language giving it better balance. 

5. In the long run. It will prove to cost more to develop 100 RAD systems 
that never quit do the job. Than one upgradable modifiable system that 
performs reliably throughout it's lifecycle.

Ada will continue to gain popularity particularly when the as of systems 
becomes more important. 

These opinions are mine and those who agree with them them.

                 Mark McKinney
                 mckmark@mail.concentric.net