kent@xanth.UUCP (Kent Paul Dolan) (08/24/87)
In article <1043@bsu-cs.UUCP> neubauer@bsu-cs.UUCP (Paul Neubauer) writes: >We don't really need to worry about it [diacriticals in other languages] >for C programs, since the characters needed for that are already known. >Paul Neubauer UUCP: {ihnp4,seismo}!{iuvax,pur-ee}!bsu-cs!neubauer Paul, your posting was fine except for this one lapse. For some number of years we have been getting away with assuming that English is the universal programming language. Since it is not proven that programming has the same criticality as, say, piloting of ships and airplanes, where such a rule makes sense, I expect programming in the native tongue to become more widespread as programming becomes a more worldwide activity. In particular, for C, the use of "meaningful identifiers" must imply "meaningful in the tongue of the reader". We can probably get away with english keywords, keywords have such conventional meanings they are almost divorced from their common English or other language meaning anyway, but I would expect to see the alphabet in which C identifiers are written to vary to match the needs of the language group using C. This of course implies a lot of new problems in code portability. It is time to face these problems, either by changing the de facto English predominance in coding to be de jure, or else by providing for extended alphabet standards and compliant compilers. Gives a whole new meaning to the idea of program translators, too! Kent, the man from xanth.
karl@haddock.ISC.COM (Karl Heuer) (08/25/87)
In article <2242@xanth.UUCP> kent@xanth.UUCP (Kent Paul Dolan) writes: >I would expect to see the alphabet in which C identifiers are written to vary >to match the needs of the language group using C. Probably. Such programs cannot be strictly conforming, but if they are not expected to be ported outside the locale of origin (this includes not only foreign countries with nonstandard letters, but also VMS sites in the USA), that's tolerable. (And it's not difficult to write a transliterating program, anyway.) A compiler can admit such identifiers and still be conforming, so the implementor has no good reason not to. (Maybe someone needs to design a language along the lines of APL, with no natural-language bias at all.) Karl W. Z. Heuer (ima!haddock!karl or karl@haddock.isc.com), The Walking Lint