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

Re: [idn] summary of reordering discussion




----- Original Message ----- 
From: "James Seng/Personal" <jseng@pobox.org.sg>
To: "Soobok Lee" <lsb@postel.co.kr>; "Bruce Thomson" <bthomson@fm-net.ne.jp>; <idn@ops.ietf.org>
Sent: Friday, November 09, 2001 2:06 PM
Subject: Re: [idn] summary of reordering discussion


> > Disagree. Significant portion of real CJK and otherscript
> > registration contains 6 or more characters. That will be
> > increased by introduction of new future broader and diverse
> applications of IDN.
> 
> You forgot that this group is doing "Internationalization", not
> "localization".  Therefore a draft merit should justify itself based on
> its applicability to the global community (ie all script) not a just
> particular locale (ie just some subset of script). No script should be
> and must not be more important over the other.
> 
You should not forget that 
han/hangeul are the disfavored script in ACE encoding.

> I believe this point was raised in London by Paul, on how reordering is
> going to work for other script. You have answer it a couple of times
> but, correct me if I am wrong, summary of what you wrote as I read it is
> "I dont know those scripts and I dont care".

I had suggested a threshold. It was not thoroughly discussed.
The size of script of han/hangul exceed 10000. Others are not!
The size of script affects the ACE encoding efficiency.
They are already suggested in my draft. no one cited that in this list yet.

> 
> > Most webhost/mailhost names are seldom typed in , but *clicked*
> through
> > from web pages or email clients. official long biz names are welcome
> in IDN.
> 
> Irrelevant. I still dont see why a bq--ewojasdje39a is better than
> bq--sdkjs993k to naked eye (if and when they do see it). You mean you
> are able to to remember bq--sdkjs993k?
> 
no need to remember, but just compare or transscript. short lables help.
What if the are displayed in narrow wap phones?
ACE IDN will be used by and exposed to  over 6 billions, not only to handful of us engineers, 
but also to children, elders, educated, non-educated from various backgrounds.

see this:

bq--wqeqwkejwkqeqwewq
bq--ewqeqweeqww

bq--qweuiwqueqwiueoqwueoqwueoqwueqwueoqwueqwewq
bq--weqweqweqweqweqweeqweqewqw

For long CJK names, compression ratio reach 30%, and for frequent names up to 40%.

Soobok Lee