Re: C/C++ programmer giving Ada95 a chance -- writing an emulator.

[Robert Dewar <robert_dewar@my-deja.com>]

In article <38E3D0A5.4F93106F@Raytheon.com>,
  "Samuel T. Harris" <samuel_t_harris@Raytheon.com> wrote:

> It is interesting how two documents from two different
> sets of folks concerning basically the same things
> can be so very different in style.

Note that it is not so much a difference in authors style here,
as a very deliberate design choice of favoring formalism and
supposed precision over readability.

I am not sure everyone on WG9 and the reviewers group really
understood what a very big change was being made. The above
note is *exactly* what concerned me when I tried to argue for
retaining the original style.

I suspect that if the original Ada 83 RM had been written in
this style we simply would not have achieved the pretty much
unique status among languages that practicing programmers use
the defining document as a reference.

Luckily we benefit from history here, Ada programmers from 83
days are so used to operating in this mode that they continue
the habit, even though it is harder these days.

Is the Ada 95 RM more precise? I am not 100% convinced at all.
Looking at the set of AI's that are coming in, we are seeing the
same kinds of minor gaps and errors that happened with the
Ada 83 RM, I would be hard put to say there was a big
qualitative difference in the scope of AI's considered in the
two cases. Bob (Duff) what do you think on that score? [I
think Bob's input here is useful since he knows both RM's
well and both sets of AI's, and if biased at all would be
biased to thinking the 95 AI's are less significant (*)

(*) in fact, I never noticed Bob's viewpoints being perceptibly
affected by bias of any kind :-) even though he is constantly
warning that as an author he may be biased!



Sent via Deja.com http://www.deja.com/
Before you buy.