Re: [Caml-list] syntax change (was: camlp4o problem)

[Markus Mottl <markus@oefai.at>]

On Mon, 04 Feb 2002, Daniel de Rauglaudre wrote:
> What do you call "standard"? Yes, my revised syntax is "standard": it
> is "standard" Camlp4: the sources of Camlp4 are in the revised syntax.

I don't think there is much point to argue: "standard" is what the
OCaml-reference manual says on the language definition and what the
compilers accept as default. Quite unsurprisingly, this is what >99%
of all OCaml-developers are using...

> I compile the whole compiler with -pp camlp4o, the otherlibs included
> without any problem. I am pretty sure that it would work for your code
> with minor changes which would remain backward compatible. Try it out.

Sure. But programming is also a social activity (well, at least sometimes
I'd hope this were the case ;).

This means that it is out of the question to let everybody write in
whatever style he thinks is right just because some preprocessor does
conversions automagically. Just think of teaching OCaml: would you
really want to have every teacher use his favourite style? - That's an
easy way to drive off prospective users.

> The problem is that if you changed your syntax, you cannot "downgrade"
> your compiler to the previous version.

Nobody changes production code to a new language syntax just for the
fun of it. In any case, I am not sure that this is really a problem:
unless you use new, possibly broken semantic features, you'd only have
to fix potential bugs in the preprocessor to make your sources work with
the old compiler again, wouldn't you?

> > I disagree here. People need a "soft kick" to change.
> 
> No: people will accept a change if they need some.

This is, I fear, not the way the "masses" behave.

> But I am quite sure that many people think this: "yes there is a
> problem of syntax in the language, but not so important".

The current OCaml-syntax is definitely not as broken as the ones of most
(if not all) mainstream languages. In fact, one may get to like its
quirks :-)

> I often heard the argument: "it is only syntax!". A certain "Markus
> Mottl" telling a few mails ago: "Though semantics is usually
> considered the more interesting part of languages..."

This is not a contradiction, IMHO. Semantics is not only more interesting
(= potentially more rewarding), it is also much more difficult to improve
(= higher costs). If you perform a cost/benefit analysis, there may be
a point where the ratio is better for syntactic improvements. I am not
sure whether we are already there. AFAIK, there are still somewhat large
semantic extensions in the development pipeline.

Just to make this clear: being a "normal syntax" user, I am not
pushing for any specific syntactic changes. I am just asking whether
the OCaml-team considers major syntax improvements in their development
strategy at all. It's not a problem for me directly if things stay as
they are. It may, however, benefit the growth of OCaml if there is such
a strategy.

Regards,
Markus Mottl

-- 
Markus Mottl                                             markus@oefai.at
Austrian Research Institute
for Artificial Intelligence                  http://www.oefai.at/~markus
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr