Re: Getting back to the Chinese example
> Date: Mon, 28 Oct 1996 11:51:08 -0800
> From: avine@dakota-76.Eng (Andrea Vine [CONTRACTOR])
> OK, David's example of a preamble style in English vs. Chinese seems to point to a particular view of the
CLASS name, that is, it is the individual user defining the name. Not the language, not the locale, not the
international standards groups.
> So, regardless of what country or language a person has defined, they type:
> <H1 CLASS=FOOBAR>
> When their HTML parser sees CLASS=FOOBAR, it looks in defined paths for a style labeled FOOBAR (here is one
point of contention, does it look for an exact match, including case, for FOOBAR? Or does it look for
Foobar, foobar, FoOBaR, etc. I recognize that this is not yet resolved...). If it does not find FOOBAR, be
it Chinese, Swahili, or Esperanto, that style cannot be used, therefore it defaults to some standard style it
would normally use.
I hope **style** is just one of the "uses" intended for CLASS names. There are far richer
applications that are possible with semantically tagged text. e.g.I'd love to see, abstracts,
preambles, citations etc. indexed separately from full text indexes of the web.
> Now, recognizing that case-conversion is a problem, I am trying to understand the other bones of
contention. In my interpretation, I would say the matching should be exact internal encoding, no conversions
no composition/decomposition. However the individual user has typed it is exactly what should be matched.
Exact matching without case conversion seems to be the simplest implementation
alternative. Recommending composed characters for western European languages
and decomposed characters for the rest of the world scripts may be the most
portable interpretation for names.