Re: Internationalized CLASS attributes
Bert Bos wrote:
>> The next version of HTML will have a CLASS attribute on (nearly) all
>> elements, as described in several documents (, , , ). The
>> intention is to allow authors to attach semantic information to
>> So it appears that case-conversions are language dependent. That's
>> why, for example, there is setlocale() in POSIX. But is it practical
>> to make the case rules for CLASS dependent on the language? Where
>> would you get the language from?
>> Or do we change the interpretation of CLASS, and say that it is just a
>> code (class=xyz12, class=p-89x), that doesn't have to be
>> human-readable? In that case ASCII is all we need.
>> What do people think?
I assume that most "code" will be generated by software, rather than by a person
and a text editor. If there is some form of interface or tool between the
information provider and the "code" they generate, and a similar isolating layer
between the end-user and the underlying code, then I don't mind what language or
code is used to name the class, so long as there is a meaningful, well-known
mapping between the value used to name the class and the word, phrase or
icon the user uses to access the information in the entity.
Given the choice, I'd leave it in ascii.
Martin Bryan wrote:
> What we really need is something, like the RFCs relating to REV and REL,
> which suggest a set of useful class names that could be applied by anyone,
> irrespective of their country of origin. Admittedly most people would not
> then be able to use their native language to name such transportable
> classes, but the up-side would be that they would be able to identify
> information of the class they require without having to search for all the
> possible names for the class.
If standard classes are to be really useful, we need some kind of language,
country, and culture-independant semantic description of each one. Even in the
restricted information systems I play with, trying to define genuinely useful
classes which mean something to both the info provider and the end user is a
nightmare. I like the idea of standard class names, but I'm not sure how
realistic it is to expect to be able to define a standard set which would have
any kind of universal applicability, irrespective of the linguistic or cultural