Re: Three simple questions

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

In article <8sn71i$8a21@news.cis.okstate.edu>,
  dstarner98@aasaa.ofe.org wrote:
>
> Anyway, I can't see Ada not using KC normalization on the
characters. NJ
> (an encoding digraph) should not be different from the ASCII
NJ, for example.


It may be helpful to remember that the Ada RM has NOTHING to
say about how programs are represented. There is no necessary
connection between the enumeration literal 'A' in the type
Standard.Character, and the way the letter A is represented
in source programs. In typical compilers there may be a strong
relationship in practice (quite useful for example if you
want to use type String to read in an Ada source program),
but this is an implementation choice, it is not something
that is required or even suggested by the RM.

I mention this, because I have found that there are wide
spread misconceptions in this area.

As an example, a compiler that required all keywords to be
input in bold would be a conforming compiler, that's true
even if the method of achieving bold was to have some strange
escape sequence in the sources. Similarly a compiler on IBM
main frames (contrary to widespread misinformed rumour, such
things still exist, and people buy new ones all the time) might
well represent sources in EBCDIC, and that would be just fine.


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