Re: Clear Screen

[dewar@merv.cs.nyu.edu (Robert Dewar)]

Robert Eachus said

<<(And I lost on one in the formal vote--pragma Elaborate--but won
later when the Development Team actually checked out the issue.  My
guess is that in about 30% of the cases where pragma Elaborate
actually needed, it can't be replaced by Elaborate_Body or
Elaborate_All.  But I suspect the real reason it was left in Chapter
10 was that it only requires a few words there. ;-)

>>


Rober notes

The history here is that we found it was impossible to build the GNAT
runtime without the use of pragma Elaborate. I investigated carefully,
and sent the exact examples into the vr list, and when we looked into
them they were indeed reasonable.

Nevertheless the estimate of 30% is *way* too high. This is not a guess
it is based on actual observations of the number of times that pragma
Elaborate is used, and could be replaced by Elaborate_All in the GNAT
test suite.

Robert Eachus said

<<There are SOME Annex J features which will probably disappear next
time around: the allowed substitution characters and the chapter 13
stuff (with the possible exception of interrupt entries).  But some of
the stuff there, including package Ascii and the package renamings, will
probably be around in Ada 3x.

>>

That's exactly what people said about ALTER in the context of the COBOL 74
standard (they were sure it would be removed next time around). People who
make such statements greatly underestimate the force of compatibility
arguments from legacy code. I will bet anyone who cares to take the bet
that none of the Annex J features will disappear from the next version
of Ada. I note in advance that of course I will argue against their
removal regardless of whether anyone takes up this bet. We are talking
about a couple of minor features here that are trivial to implement. It
is inconceivable to me that the designers of the next version of Ada
would introduce gratuitious non-upwards compatibilities by trying to
remove them!