Re: [idn] case folding

[James Seng <jseng@pobox.org.sg>]

Frank da Cruz wrote:
> Am I the only one who finds all this just the least bit scary?
> There is already so much that can go wrong when trying to make an
> Internet connection that I have to be skeptical about the ability
> of an uncountable number of clients and servers correctly
> implementing complex normalization and case folding procedures
> (according to a database that changes periodically) when the majority
> of Internet developers are still almost totally unaware of character
> sets to begin with!  (I can't prove this but I say it with some
> confidence.)

I find the discussion on NFKC and Compatibility characters useful but
I hope the Unicode do not design depending on where the normalization 
and canoniclaization rules occurs.

Having it on the client *is* scary. Consistency to maintain is unimaginable.
But that may be the the only technical feasible way to do it on deployment.
Whatever the case, it is best not to make any assumption.

Next, it would be nice if we have a consistent form to use and not KC for
some, and not KC for others. 

> Plus I wonder how it will all play out in the marketplace.  Given
> that "much" of the world's population will not be able to enter
> non-ASCII (or non-Latin-1) on their keyboards, how much success will
> a website enjoy in the world market when it names itself in a complex
> (or even non-ASCII) script?  (Of course it's another matter for
> websites designed only for local consumption, e.g. in Taiwan.)

While it is kind of out of topic, the question to ask is if the website
is *already* in Arabic or Chinese, why do they still need an English URL?

Or the person can *only* read/write Chinese, why should he learn English?
 
> Also as a practical matter, what about fonts?  Our Windows-based
> dialog boxes (as in the Web browser's Open dialog) use monospace
> fonts.  How will we open websites in our browsers -- even assuming we
> have an appropriate input method -- if the name can't even be
> displayed in the Open dialog's URL text box?  (In other words, who
> has a monospace font that includes anything beyond Latin, Cyrillic,
> Greek, and Hebrew?)

Moot. 

-James Seng