Re: Difference between ADA and c++

[jsa@alexandria.organon.com (Jon S Anthony)]

In article <Pine.SGI.3.95.970826174002.24873B-100000@shellx.best.com> Brian Rogoff <bpr@shellx.best.com> writes:

> On 26 Aug 1997, Jon S Anthony wrote:
> > 
> > I don't think Brian meant that this would make them easier to read (he
> > said "superior") I think it would make them more difficult to read,
> > but in some sense "easier to write"...
> 
> Cool, we disagree :-). Certainly, if I thought that adding type inference 

Sacre Bleu!


> Jon, I'm interested in knowing exactly why you think it would make programs 
> more difficult to read, especially if the automatic instantiation was 
> optional.

Only in the sense that many large scale things are happening and
propagating behind the scenes.  Sort of like the old issue with macros
where you can end up not having even a clue as to what is happening or
what the actual result is.  Of course, as you say eleswhere, this is
simply a matter of abuse (vs. use).


> My experience has been that long instantiation sections tend to
> contribute nothing to my understanding of a program, much like most
> of the comments I read, and just obscure things. I certainly accept
> that your experience might be different, so I'm all ears.

This is an excellent point.  I think you said it elsewhere as "more
visible code does not imply more clear".


> Have you used Haskell or ML (SML or CAML)? These languages do type

I've looked into ML and now (from an earlier suggestion by you) OCAML.


> inference for everything, and I find the short code to be fairly
> readable.

Yes, but I'm not sure how well this translates for large scale system
maintenance and understanding.  When you get down to it, type
inference is more an efficiency thing (literally and for the writer as
well) than anything else (another example, Baker has an interesting
paper on how to piggy back storage inference on top of type inference
to increase GC efficiency).


> I think it is better to be mostly explicit, and so I favor the Ada
> approach in general, but a little type inference in the right place
> seems like a big win.

Sounds reasonable, but how/where do you draw the line?  Maybe that's
easier than I think??

/Jon
-- 
Jon Anthony
OMI, Belmont, MA 02178, 617.484.3383 
"Nightmares - Ha!  The way my life's been going lately,
 Who'd notice?"  -- Londo Mollari