Re: [xml-dev] version numbers and infosets

Michael Kay scripsit:

> You could do it without changing the definition of well-formedness by
> saying that the set of characters considered to be whitespace, and
> normalized as such, is a property of the encoding.

Fine and dandy for EBCDIC, but not so good for Latin-1 as used on mainframes,
where 0x85 = NEL.

Actually, not so good for EBCDIC either, because it means that each of the
dozens of EBCDIC code pages has to exist in two flavors, a native flavor
where 0x15 encodes U+0085, and an XML flavor where 0x15 encodes U+000A.
This is more or less what FTP software has to do, and it's ugly.

-- 
John Cowan                              <jcowan@reutershealth.com>
http://www.reutershealth.com            http://www.ccil.org/~cowan
                .e'osai ko sarji la lojban.
                Please support Lojban!          http://www.lojban.org