Re: Three simple questions

[Tucker Taft <stt@averstar.com>]

[NOTE: as you can see from the number of responses,
there is no such thing as a "simple" question about
a programming language...]

Frank Christiny wrote:
> 
> Friends,
> 
> Having arrived recently into the green Adian world, I am trying to
> assimilate all there is to know about this powerful, unfairly
> maligned computer language.  Most questions I have been able to
> answer from the FAQs here and there but there are three (so far) that
> have gone unanswered:
> 
> 1.  What is the correct pronunciation of Dr. Ichbiah's name?
>     Is the "ch" pronounced as in "Loch" or as in "each"?  Is the
>     final "h" is silent?

Eesh-bee-uh is how I pronounce it.  Jean will have to give his own
official version.  By the way, at one of the early Ada conferences,
someone claimed that "Ichbiah" was Swedish for "iceberg" and that
is why 90% of the complexity of Ada 83 was hidden below the surface.

Of course, with Ada 95 we tried to raise the iceberg a bit, and maybe
even smooth off a few of those dangerous irregularities.

> 
> 2.  What is the rationale for case-insensitivity in the language?
>     I am beginning to think it is mostly for historical reasons
>     (some machines didn't have two case letters way back then, etc.).
>     Are there any real technical reasons for why this is so even
>     after Ada95, besides preserving backwards compatibilty?

This is an explicit choice, and one that I think I would personally
make the same way again.  I have been teaching a "Programming Languages"
course recently, and it is surprising how frequently students, even those
conversant in C, are surprised that in some contexts, "Begin" and "BEGIN"
mean two different things.

One thing that some compilers do is warn the programmer if they
use different upper/lower casing in a reference than in the declaration.
E.g., if you declare something as "XYZ" but have a reference like "xyz"
then you would get a warning.  That seems like a friendly feature, and 
might even be worth enshrining in a language rule, or perhaps at least
a pragma.

> 
> 3.  How can one block-comment in Ada?
>     It can become cumbersome sometimes, adding all those "--"s in
>     front of a large block of code one wants to "test out" on the fly.
>     Again, are there any good technical-engineering reasons why this
>     feature was left out?

From my perspective, being a big fan of "grep" (the Unix version of
a file "search" command), I am always pleased that every line that
has been commented out starts with "--".  It can be quite distressing
to think that a line is executable, and then discover that it is
actually buried in a commented-out block.

Note that most programmer-friendly text editors have an ability to insert
something in front of every line of a block, or to remove same.

Finally, in languages that do provide block comments, I find they
rarely work right if the block being commented-out already contains
comments.  If you tend to comment at least once every 10 or so lines,
a block of code doesn't have to be very big to require significant
fiddling to allow block commenting to work.  By contrast, you can
always stick "-- " in front of every line, even if the line already
contains a comment.

Both of these latter two answers reflect in part Ada's philosophy of
favoring the "reader" over the "writer."  And since in most cases,
even for our own code, we end up being a more frequent reader than
a writer, this philosophy generally increases the overall productivity 
of the original programmer as well as some subsequent program 
reviewer/maintainer/enhancer.

> 
> Regards,
> --
> Frank Christiny                       fchris@pdq.net
> Sr. Software Engineer                 Lockheed Martin Space Operations
> Houston, Texas, USA                   http://freeweb.pdq.net/fchris/

-- 
-Tucker Taft   stt@averstar.com   http://www.averstar.com/~stt/
Technical Director, Commercial Division, AverStar (formerly Intermetrics)
(http://www.averstar.com/services/IT_consulting.html)  Burlington, MA  USA