Re: PL/I Versus Ada (Was: Arianne ...)

[adam@irvine.com (Adam Beneschan)]

tonyk@sseos.lbl.gov (Tony Konashenok) writes:

 >I would add to this that a human being (assuming that a programmer is
 >human) also has certain capabilities of data abstraction [s]he uses in
 >writing code, and it had better be matched by the programming language.
 >What I assert here is that PL/I has the abstraction capacities roughly
 >equivalent to those of human thinking; furthermore, it operates the same
 >syntactic categories as most humans. Therefore, it is very intuitive.
 >If you look at meta-languages used here and there to facilitate the
 >description of logic flow, most of them are very much like PL/I. Not Ada.
 >Not C++. Oh well, BNF is indeed closer to Ada, but hardly anyone uses
 >it for intuitive explanations.
 >Even if one needs greater degree of abstraction, it is always possible to
 >formulate this abstraction once using the traditional syntactic constructs,
 >include it in your code, and then reuse it as you please. Just like writing
 >your own macros in assembly code. PL/I, by the way, has certain features
 >to help implement object-orientation. GENERIC comes to mind.

OK, now you've got me curious.  I looked a little at PL/I many years
ago, but I'm afraid the language may have changed too much in the last
15-20 years.  Could you explain in greater detail the assertions
you've just made?  In particular, I'd be interested to see:

(1) What the syntactic categories of PL/I are that correspond to human
    thinking, and how Ada falls short in that regard.

(2) What the abstraction capacities of PL/I are, and what about them
    makes them more equivalent to human thought than Ada constructs
    do. 

(3) How PL/I's GENERIC features facilitate object-orientation better
    than Ada's GENERIC features do.

(4) How Backus-Naur Form resembles Ada.

                                -- Adam