From: Doug Ewell (dewell@adelphia.net)
Date: Thu Sep 21 2006 - 00:12:40 CDT
Guy Steele <Guy dot Steele at sun dot com> wrote:
>> (laughs) Programming languages don't get "deprecated" like last
>> month's fashions.
>
> Sure they do. PASCAL and COBOL are probably (informally) in a
> deprecated state by now. And when was the last time you used TRAC, or
> SNOBOL, or JOVIAL, or COMIT, or MADCAP, or JOSS, or FOCAL, or PPL, or
> ECL, or PL/I, or RPG, or MODULA-3 (let alone MODULA)? There are
> indeed fashions in programming languages---they just change over
> periods of decades rather than months or years.
I've spent a lot of time lately thinking of "deprecation" in a formal
sense, as defined by a standards organization. I agree that a great
many programming languages have fallen out of common use (or in some
cases, any use), but to me "deprecated" implies a big red DEPRECATED
stamp applied by someone with authority to do so.
> It may help to know that the word "deprecated" is a term of art within
> the programming language standards community, though it is usually
> applied to specific features rather than entire languages. For
> example, the "assigned GOTO" was formally deprecated in Fortran and
> eventually discarded.
Within a language, certain features can of course be deprecated (like
assigned GOTO). HTML is chock-full of them.
> That said, while I have plenty of other reasons to deprecate C and
> C++, I see no reason why C and C++ standards committees might not
> adopt extensions that would support the Unicode character model.
> There are already libraries defined for those languages to support
> "wide" (16-bit) characters.
I write plenty of C++ that supports Unicode, because I use MFC and the
CString class and the _UNICODE compiler directive. Vendor-specific?
You bet it is, but that isn't the point.
-- Doug Ewell Fullerton, California, USA http://users.adelphia.net/~dewell/ RFC 4645 * UTN #14
This archive was generated by hypermail 2.1.5 : Thu Sep 21 2006 - 00:21:05 CDT