naming conventions, was: Case sensitivity

This is a very interesting point, and comes closest to convincing me we
should go with strict case sensitivity.  I'm not quite sure how to
think about this though - will we have to define a new command set for
each language?  My best guess would be that translating all of Maxima's
commands into another language is a highly nontrivial task - we aren't
even aware of all that's in there in English.  Documentation and maybe
things like menus can be handled this way, but I'm dubious as to how
the command language itself could be made multilingual.

CY

--- Steve Haflich <smh@franz.com> wrote:
> I've been following the discussion as something of an outsider -- a
> Lisp implementor rather than a real Maxima user -- but I want to
> bring
> up an issue everyone has been missing in the tangle surrounding case
> sensitivity: I18N.
> 
> It is more and more common for language implementations to support
> I18N characters, typically in Unicode.  Some Lisps do already, and
> Java has used Unicode (usually encoded externally in UTF-8) from the
> outset.  (FSF Emacs also does, except that Mule encoding has some
> minor differences from Unicode.)  Memory and transfer bandwidth
> continues to cheapen, so it is clear that over time more and more of
> the computing universe will support 16-bit Unicode or better.  Maxima
> should not make decisions now without thinking what typical
> applications will look like in ten years, or else the solution won't
> last.  Non-Emglish speakers should be able to name their variables
> with native words.  I've even heard about this obscure tribe of
> non-English-speakers called "Mathematicians" who adopted the Greek
> alphabet when they need to record their thoughts, and some obscure
> sects of Mathematicians have been known to use the distinction
> between
> "small" and "capital" Greek letters.  (Unicode provides for this
> distinction.)  Case is not normally used in the CJK languages, but
> typographically, there are variants (e.g. the "half-size kana" in
> phonetic Japanese) that are considered separate characters.  And
> (Lord
> help us) there are some Eastern-European languages that have _three_
> cases: upper, lower, and title.
> 
> [This is an extract from an online file describing the downloadable
> Unicode database:
> 
>   CASE MAPPINGS
> 
>   In addition to uppercase and lowercase, because of the inclusion of
>   certain composite characters for compatibility, such as "01F1;LATIN
>   CAPITAL LETTER DZ", there is a third case, called titlecase, which
>   is used where the first character of a word is to be capitalized
>   (e.g. UPPERCASE, Titlecase, lowercase). An example of such a
>   character is "01F2;LATIN CAPITAL LETTER D WITH SMALL LETTER Z".
> 
> ]
> 
> Now, it would be nice if some day Maxima users on capable platforms
> could use Greek and their native language without prejudice.  Of
> course, for the forseeable future internal modules and library code
> should be restricted to ascii/Latin1, otherwise some platforms won't
> even be able to read it, much less compile and run it.  But that
> restriction shouldn't unnecessarily extend to what the end user types
> into his Maxima command loop, or into the application files he
> writes.
> If it cannot do so already (on a I18N-capable platform) little would
> be necessary to make Maxima capable of full Unicode support.
> Therefore, language designers should not think about case just in the
> context of the Latin1 character set.
> 
> Now, the proposals that Maxima be case-insensitive have a built-in
> and
> IMO unwise assumption about the I18N capabilities of the underlying
> Lisp implementation.  This is entirely homologous to the similar
> discussions abot case sensitivity in the Lisp community.  Most
> case-insensitive implementations treat car and CAR as the same
> symbols, but treat upper-case-lambda and lower-case-lambda as
> distinct
> characters, as well as the odd case-like character equivalences of
> Near-Eastern, Far-Eastern, and Hyper-Eastern languages.
> 
> Implementationally there is a good reason for this.  It is
> straightworward for the Lisp or Maxima implementation to add a little
> reader code to fold the 26 Latin1 letters into a single case.  It is
> hugely difficult to do the same thing for the entire set of Unicode,
> which aside from being a moving target, also embodies the risk of
> violating incomprehensible non-Northern-European cultural
> expectations.  It leads to a language that considers X and x to be
> the
> same, but insists that upper-case-lambda and lower-case-lambda are
> different.  This bothers me as inconsistent and surprising to
> non-ISO8859-1 speakers.
> 
> Therefore I suggest that any future language, in which the end-user
> programmer can invent variable names of his own choice, should be
> case
> sensitive, if only for consistency with I18N text.
> 
> [Editorial note: The above was mostly serious.  What follows is not.]
> 
> Anyone who still doesn't think that case matters is invited to view
> the following web page, which shows how functional calculus can
> diverge when case is ignored.  It contains two nearly-identical
> images, unretouched, that I found in separate places on the internet:
> 
> http://www.franz.com/~smh/nlambda.html
> _______________________________________________
> Maxima mailing list
> Maxima@www.math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima


__________________________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo
http://search.yahoo.com