New language features?

[Joe]

It seems to me that C could be improved by adding a few more features
to the standard. I read some of the recent thread "Criticism of the C
programming language ??????" which shows a lot of opinions for and
against adding language features.

What makes sense to me is to incorporate features that are commonly
available as compiler extensions (i.e. anonymous unions) into the
standard, because they are well-tested, and obviously have many actual
users interested. It also would not put a big burden on compiler
developers. Maybe the formal standards process is just too complicated
to allow the addition of a small number of minor additions? Maybe it
would be useful to have a well-defined, standardized set of language
extensions to make it easier to write portable code? Or, maybe it is
too hard to get everyone to agree, and I should be thankful at the
current level of C standardization?

 

[Eric Sosman]

It seems to me that C could be improved by adding a few more features
to the standard. [...]

It seems that way to everybody else in the world, too.
That way, me lad, lies PL/1.

 

[SM Ryan]

# would be useful to have a well-defined, standardized set of language
# extensions to make it easier to write portable code? Or, maybe it is
# too hard to get everyone to agree, and I should be thankful at the
# current level of C standardization?

(1) gcc source is available. You can modify it (or sponsor someone
to modify it) to incorporate the changes you want. You can then
release those changes to the gcc community and try to get them
incorporated in the compiler releases. That gets you leverage to
get ANSI to include them.

(2) Write a preprocessor to convert the language you want to the
language you have to use.

(3) Write a new language.

 

[cr88192]

It seems to me that C could be improved by adding a few more features
to the standard. I read some of the recent thread "Criticism of the C
programming language ??????" which shows a lot of opinions for and
against adding language features.

What makes sense to me is to incorporate features that are commonly
available as compiler extensions (i.e. anonymous unions) into the
standard, because they are well-tested, and obviously have many actual
users interested. It also would not put a big burden on compiler
developers. Maybe the formal standards process is just too complicated
to allow the addition of a small number of minor additions? Maybe it
would be useful to have a well-defined, standardized set of language
extensions to make it easier to write portable code? Or, maybe it is
too hard to get everyone to agree, and I should be thankful at the
current level of C standardization?

here is the problem:
there are no standards;
or, there is some long and drawn out process for making them.

now, a system for informal compiler extensions would be useful, especially
if some compilers could agree on whether or not to implement them.

a tradeoff could be possible, for example, something similar to the
old-style IETF RFC system, or similar...

Scheme has something similar (the SRFI system).
http://srfi.schemers.org/

one could look here for a possible organizational idea.


but, then is the question:
who would organize such a system?...;
and, how much would it actually matter?...;
....

unless some more major players in the compiler field were involved (such as
GCC, ...), this would be, at most, largely restricted to minor and hobbyist
implementations (my case is a case of hobbyism...).

this may not be all bad though, as even some semblance of standardization is
probably better than the existing practice of "word of mouth" compiler
extensions...

or, maybe similar exists already?...
I don't know...