[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [idn] opting out of SC/TC equivalence



You are correct in either/or not both as current ASCII
world.  If we go IDN, a good webpage can display 
IDN or send it out after [nameprep] it.  If you do not 
let TC/SC mapping into [nameprep], all the
 internationalized <Chinese>.com can only use the 
current cut and paste feature with ASCII version, 
which is uniform for everyone.  If you
do let TC/SC into [nameprep] it can have limited 
TC/SC rate, which is good enough for a lot of people. 
This is up to TSCON's definition listing.  Users will
treat technology with a little salt in it.  When a user 
application decide to translate a URL before it is 
displayed or sent out, cut and paste can have a good
 chance to be [nameprep].  Isn't this IDN's goal?  Or
are you thinking of changing everything into UTF-8?

If we follow John's keyword searching scheme, then 
<Chinese>.com has to go through another level of 
guessing stage.  Then it will be impossible for 
<Chinese>.com to get cut and paste feature.  Because  
this becomes an AI technique issue not a font 
display issue any more.  We are dealing with identifiers
not artistics expresstions of an address.  Chinese 
has a lot of aritistry, but hostname is not a place for 
that TC/SC perfection.  

Liana

On Fri, 31 Aug 2001 08:42:15 +0200 Harald Tveit Alvestrand
<harald@alvestrand.no> writes:
> to my mind, it seems that you can have keyword search with 
> near-perfect 
> TC/SC equivalence, OR cut-and-paste URLs and email addresses without 
> TC/SC 
> equivalence.
> 
> Not both in the same part of the user interface at the same time.
> 
> --On 30. august 2001 14:55 -0700 liana.ydisg@juno.com wrote:
> 
> > To the keyword type of solutions, it is good for name
> > registration only.  It is slow and painful, a good toy to
> > study with, but not good to use as day to day tool,
> > especially unthinkable if I want cut and paste type of
> > URL implementation.  So, I am agaist keyword search
> > type of e-mail addresses, but still interested in
> > any solutions to guess how it can be implemented in
> > English.
> 
>