robin

and Ada, etc


They merely continued the trend set in FORTRAN.

That's a load of nonsense. You know it. I know it. We all know it.

Have you been listening?
Can you hear the guffaws?


In some instances where that reserved word has not been used
as a keyword, yes.
Try replacing a reserved word inadvertently used as a variable
when the program already uses that same reserved word
as a keyword.

But that's beside the point.
RESERVED WORDS WASTE PROGRAMMER'S TIME.


You haven't been listening.
RESERVED KWYWORDS WASTE PROGRAMMER TIME.


You wish that were so, because your favorite language,Ada,
has reserved words.

The fact remains that a language with reserved
is a nuisance.

You would wish that to be so, since your Ada has reserved words.

robin

But you do.
Otherwise you wouldn't continually repeat over and over
parrot-fashion, the same tired tirade.

Several people in this forum have posted otherwise.
You are out-of-touch with the real world.


Others in this forum have posted otherwise.

Others have posted otherwise in this forum

It is an important issue.

robin

That's because it is a general purpose language,
encompassing commercial, scientific, system programming,
and computer science, etc.
The lack of reserved words is an important component of the
design, and is integral to that design.


Indeed.

adaworks

I am quite content to acknowledge that some small
number of software developers may regard reserved
words as a nuisance. This does not seem to represent
the view of the larger community of software professionals.

It certainly does not represent the view of those who labor
long and hard to design new languages. These people are
neither lazy nor stupid. They are serious students of
programming language design. For the most part, they
have abandoned the PL/I approach.

I really don't care that much one way or the other. Whether
a language has reserved words or not is unimportant in
deciding the larger issues such as whether a given language
is well-suited to the solution of real problems. It is also
unimportant in determining whether a language is able to
support new ideas in programming language design.

If PL/I can evolve to support modern programming ideas
and concepts without reserved words, so be it. I realize
you believe that PL/I has evolved, but most of the items
you have listed are rather pedestrian when viewed in the
context of other programming language designs.

As I see it, PL/I is still not up-to-date when compared to
most of the new programming language designs. This may
irk you, but that is not my problem. It is a problem for
those who want to promote PL/I for a larger programming
audience. The language will either evolve or continue to
exist in a tiny, and shrinking backwater.

Every older technology can be praised in one way or another
for its virtues. I have a friend who still prefers to hunt with
his muzzle-loading shotgun and he can cite wonderful reasons
for doing so. He is a skilled marksman and can bring down,
in a single shot, more pheasants than a lessor marksman with
a more modern shotgun. Some people still prefer to program
in old-time BASIC, and they can justify this decision with
excellent argument.

You, and many others, can continue to write PL/I code and remain
quite happy with it. However, it remains a procedural language,
and, for all its virtues (and there are many), it will not be noticed
by new software professionals as long as it fails to correspond to
current views on software.

Many of the languages that do support those current views are also
flawed. C++ is one of the worst, but it has a huge following. Therefore,
popularity is not a measure of value. Java is also flawed in many
ways. Other languages, less well-known, are being developed to
accomodate those flaws. Java is evolving to compensate for some
of its inadequacies.

The larger ideas are not directly related to any given language. Rather,
language implementations are an attempt to incorporate those ideas
into a language design. For example, aspect-oriented programming
is a larger idea, but it is really difficult to create or modify a language
to support the idea.

Type-safety is another large idea. A lot of languages have been designed
to try to satisfy this concept. PL/I, simply falls short. But so do a lot of
other languages.

So, you can rant away about whether a language should have reserved
words or not, but that is a non-issue in the larger scheme of things. It
is an issue for you and a handful of others, but has little relevance to
the dialogue in the mainstream computer science community. The goal
is to design languages that satisfy contemporary ideas that support the
creation of well-formed software, reserved words or not.

The criteria for well-formed software require properties in a programming
language currently not present in PL/I. It is not impossible to add those
properties so the language can evolve into a modern programming language.
However, there needs to the will for such progress to occur.

At present, I see little evidence that PL/I enthusiasts have the will to examine
their favorite language in the context of contemporary computer science thought.

Richard

robin

It does, in spite of your silly assertions that it does not.

Naturally.
They are lazy.
BTW, there are relatively few of them (language designers).

Because they are lazy and/or stupid and/or arrogant.

You are the only one saying that.
It is a false statement.

You keep saying that, no doubt because you think that
if you repeat it often enough,
everyong will eventually believe you.
But they won't.
Already you seem to have forgotten again (since yesterday)
that you ignored strong data typing of PL/I that I have
repeatedly told you about.

Still the "everything must be modern" zealot?

Do you agree that you made a mistake about strong data
types that are in PL/I ?

[nonsense about guns deleted]


You mean, your views on software.
Because PL/I isn't like your favorite language, Ada.

Well, go and use them.

We don't need more of them [flawed languages].

And PL/I has that, in case you hadn't noticed.

You are being downright dishonest here.
I have drawn your attention to what's available in PL/I.
You chose to ignore it.

You'd like it to be a non-issue, because your favorite language, Ada, has reserved words.


It is an issue for everyone except you.

It is an important part of the design of any useful language.

Again, you are being dishonest.
You have not even bothered to find out what is in PL/I.
I have given you examples. But you persist in quoting
texts and manuals that are 25 years out-of date,
and have refused to look at any recent manual of those
recommended.
Either look at them, or shut up.


I see little evidence that talking to you is any different from talking to a brick wall.

It is quite obvious that you are out of touch with contemporary
computing.

henrik.sorensen

After having watched the energy put into these threads, I am left wondering
what it would take to divert some of that energy into actually helping me
creating a compiler and prove Richards arguments wrong, regarding
development of PL/I.

Please get involved
pl1gcc.sourceforge.net

Henrik

tim c.

Following up to "robin" <robin_v@bigpond.com> :

And where that number is greater or equal to 1 the condition is true.
--
Tim C.

tom linden

I admire your enthusiam, but as i have said before if you are not writing
the compiler in PL/I you are wasting your energies.


--
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/

henrik.sorensen

what do you mean by writing the compiler in PL/I?
- and would you mind expanding on why you thing it is 'wasting energies'?

I think this is also how Peter Flass decided to build his PL/I compiler.
Is your compiler written in PL/I ?

When using GCC, there are a number of advantages in writing the code in C,
and not in PL/I, but I guess when a first version of the compiler is made,
it doesnt really matter how the syntax trees are built.
But of course using flex and bison, also in principle means writing the
supporting code in C. That said, I have some plans for letting flex and
bison generate PL/I code, so maybe when the PL/I compiler matures, it will
be possible to write the compiler in PL/I.

Henrik
pl1gcc.sourceforge.net

tom linden

A hand-coded state machine lexer for PL/I doesn't need to be more than a
few hundred lines of code (now I have to go count ... ok its about 500) and
is highly extensible and maintainable. Now I haven't used flex and bison,
but
my experience with lex and yacc 20 some years ago left me with a pretty low opinion of these tools


Yes, and the first bootstrap was accomplished by writing an intermediate
language interpreter. To make this easier, the compiler was written in a
fairly
small subset of PL/I only requiring approximately 80 operators to compile the compiler

Actually the PL/I we did for Tru64 used vpo as the backend, an earlier
version of which, is what gcc uses for the backend. So what we did there
was to take an existing code generator (written in PL/I of course) and
retarget it for the abstract machine to generate the rtl's.


Maybe not as important as how they are designed.

--
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/