Re: Delays for implementations of Fortran features

[Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP>]

On 2014-01-19, Nasser M. Abbasi <nma@12000.org> wrote:
> On 1/19/2014 5:54 AM, Marius Amado-Alves wrote:
>
>> But, there seems to be an important difference between the two
>>open source communities: GNAT is quickly maintained w.r.t. to the
>>language revision, whereas GFORTRAN is not (as reported above).
>>
>
> May be that is because the new Fortran standard now has in it
> everything but the kitchen sink? Modern Fortran has become
> more complicated than C++, and one has to feel sorry for the
> compiler engineers working on gfortran trying to catch up with
> all those new "modern" features being added to Fortran every
> few years.
>

At some point, a language reaches a good set of functionality for it's
intended purposes and going beyond that can make the language more
difficult to use than it should be. Perhaps the Fortran community now
have the language they need in their day to day work.

Ada didn't get it fully right with Ada 83, but Ada 95 is a good sweet
spot for me. If the Ada language maintainers are not careful, then they
are in danger of making the same mistake with Ada and produce something
more complex than needed.

Sometimes I wonder (and this is a general comment on language committees,
not something directed to specific language committees like Fortran and
Ada) if the temptation is to continue adding features into a language,
even if the language doesn't need it, just so the language committee can
continue in it's job, even if adding those features were to make the
language more complex than it needs to be.

[And yes, I know many of the language committees are staffed by people
who are not directly paid for the work, but being able to say you are a
active member of a language committee is a good thing to be able to place
on your CV for your paid work.]

> In the good old days, a computer language was simple and could be
> learned in few days from cover to cover. Now it takes years
> and one still can't learn everything in it.
>

I _really_ like Wirth's approach with his active effort to reduce a
language to it's required core elements (I'm thinking about the Oberon
variants here).

Simon.

PS: I just wish Oberon didn't have upper case keywords which need to be
upper case because they are case sensitive keywords.

-- 
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
[Note: email address not currently working as the system is physically moving]
Microsoft: Bringing you 1980s technology to a 21st century world