Print

Print


Actually, the decision on mixed-case GIs was taken many years ago.  The
Metalanguage Working group discussed several possibilities, and settled
on the one you see in P3 and elsewhere.
 
The problem you describe arises because many pieces of software that
normalise SGML texts do so by uppercasing all GIs. Michael and have
already discussed this problem briefly: we considered, inter alia, an
option on the pizza chef to generate an upper-case only version of the
DTD (but should it uppercase attribute values as well?). I think a
better solution might be to produce a simple translation program of the
kind you suggest here.  Of course, the best solution would be for SGML
editors and similar software not to mess around with the case used by
the input document.
 
Lou
 
 
 
|
|The obvious benefit is: the tags are better readable compared to lower
|case letters all the way. The only problem I see (and have) now: If
|somebody wants to convert his or her SGML teilite text to XML, it seems
|to be necessary to write a line of code for each tag:
|
|In Perl:
|(1)
|$input =~ s/<RESPSTMT/<respStmt/g;
|$input =~ s/<PUBLICATIONSTMT/<publicationStmt/g;
|....
|
|If the teixlite tags would be only lower case, one line would be enough to
|normalize all tags:
|
|(2)
|$input =~ s/(<.+?) /lc($1)/e;
|
|Do I overlook a simpler solution for the conversion in case (1)?
|
|Is it more useful to accommodate to the reading of tags or to the
|automatic conversion?
|
|Thanks,
|Fotis Jannidis
|
|
|________________________________________
|Zeitschrift fuer Computerphilologie
|Dr. Fotis Jannidis
|http://computerphilologie.uni-muenchen.de
|
 
 ----------------------------------------------------------------
 Lou Burnard                           http://users.ox.ac.uk/~lou
 ----------------------------------------------------------------