Re: charset aliases

[ned.freed@INNOSOFT.COM]

> > The alisases records names that are used in some networking
> > environments, such as ISO communications, X, POSIX and others.

> The IANA charset registry is for Internet protocols, such as MIME, HTTP,
> HTML and XML. It is not appropriate to contaminate the IANA registry
> with names that are used in ISO communications, X, POSIX and others.

I agree. Just because we made the mistake of registering such things in the
past doesn't mean we should continue to do so.

> > The words I get from people is: never delete entries in registries.

> The HTML <font> tag is deprecated. There are better ways to do that now,
> and we are trying to get the Net to move in that direction.

> On the other hand, Unicode and 10646 do not remove or move characters
> once they have been assigned (with the unfortunate exception of the
> Korean mess).

> I prefer to see the charset aliases as being more similar to the former
> (HTML's <font>) than the latter, since there are alternatives (i.e. the
> preferred or canonical name).

I think this goes way too far. We have no way of knowing which of the aliases
are in use now. Attempts to scan some tiny subset of transactions in a
particular protocol are entirely futile, as usage of odd things often tends to
be concentrated in small groups of systems. We can deprecate the use of some
names, but like the <font> tag, the existance of these names needs to be
documented, not deleted. (I note in passing that any browser that doesn't
support <font> isn't terribly useful in practice.)

> I still think we should at least consider cleaning up the IANA charset
> registry.

This has already been done. Expect to see the revision available in the very
near future.

				Ned