Re: Top 10 Ada myths

[eachus@spectre.mitre.org (Robert I. Eachus)]

   (I said):
   > [...] (Of course, the
   > moment anything in this field is "finished," you find errors and bugs.
   > But Ada 95 has been much, much better than Ada 83 in this regard [...]

In article <Mats.Weber-2003952122500001@mlma11.matrix.ch> Mats.Weber@matrix.ch (Mats Weber) writes:

   > How do you know ? There isn't even a complete compiler around. I
   > think certain bugs in the Ada 95 definition will only come up as
   > the language gets implemented.

   I know how many "known bugs" there were in the 1983 RM 12 years
ago.  I could argue that the 1995 standard has been more widely
available than the Ada 83 standard was at this point in time, but I'll
assume that the degree of review has been equal.  By this time in
1983, several it's too late to fix its were known, and some problems
which were to give the ARG fits for years had been recognized.

   You could contend that there are some "known problems" with Ada 95
that will turn out to be thornier than they seem, but I don't see any.
And the only "too late to fix" that I know of are the rules for
deriving from types with constructors.  (Yes, yes, the Ada 95 design
is very nice and workable in this area, and it is much more powerful
than some of the alternatives considered.  But it confuses people, and
I expect to spend the next ten years explaining those rules.  However,
in Ada 83 there were several issues in that category by this time.)

--

					Robert I. Eachus

with Standard_Disclaimer;
use  Standard_Disclaimer;
function Message (Text: in Clever_Ideas) return Better_Ideas is...