Re: Layer 2 and "idn identities" (was: Re: [idn] what are the IDN identifiers?)

["xiang deng" <deng@cnnic.net.cn>]

On Tuesday, December 04, 2001 5:29 AM, Doug Ewell wrote:

> Does the solution work for all CJK code points?  Are there any "holes" in the 
> solution that will cause confusion and frustration for users?  (i.e. "Why 
> does it work for this character but not for that one?")

At first, you should know why the SC/TC conv was intruduced, it's to avoid 
the confusion and frustration for users while they input SC and TC. Reasons:
  Some Han characters have many variants, users can input whatever they like.
While there isn't a mechanism to make these variants to match, users will 
confuse and frustrate the situation.

If we don't concern other opinions, the simpliest solution is just do SC->TC
conversion or TC->SC conversion, So all characters which have variants will
work. But things aren't always be simply, we must concern other important 
factors and make the solution satisfy all Han character users.
 
> If the solution that is "on the desk" really does work for all CJK 
> characters, and does not violate the Law of Least Astonishment, and does not 
> involve any unreasonable tradeoffs, then it is a greatly improved system 
> compared to those that have been discussed here in the past.

From engineering viewpoint, I prefer the former solution. It solve the issue more
completely.

> > "satisfied users = $", you look down on users. It include youself while you
> > are a user.
> 
> My concern is that a long-term structural change is being proposed, and it 
> should be solid and stable, not geared toward next quarter's bottom line or 
> being first into a new market.
 
Correct. the long-term structure should involve the rooms to solve the issues 
which we can predict. If not, why not try to think other ways. We all face the 
serious pressure, but if there isn't a good solution, we'll be shame to answer 
the users' questions.

> You probably should know that as a software developer, I am deeply involved 
> in the requirements analysis process and have often been given credit for 
> putting the needs of the customer first.  I do care about users, and I have 
> seen what happens when users are promised functionality that is not actually 
> available, or not ready for prime time, or not completely thought out.  The 
> net result, after the initial excitement, is dissatisfaction and distrust.  
> Users are indeed important; they are the sole reason the product exists.  So 
> it is important to give them a top-quality solution.

Thanks you for your education about principles of software developing, though
I had learnt it again and again.
And thanks you for you care about users, if it includes CDN users.
So, agree with a top-quality solution. 

> > Before we publish the software, we had find the problem, we'll don't fix it
> > and try to salvage the situation in the next version. WHY not salvage the
> > situation in this version?
>
> Well, that's true, but it's kind of a twisting of what I meant.  What I meant 
> was that if we implement an inadequate CJK solution, or one that causes undue 
> hardship on the rest of the ACE and nameprep stages, NOW, it will only have 
> to be "fixed" with the next version.

If it's a realistic advice, Could you tell me how to "fix" it in the next version?
If the method haven't creativity and is same as the current, 
why not "fix" it in this version?

> > If a old patient will die for serious heart disease, there is no way to cure
> > completely.
> > If a poor nation fall into serious famine, there is no way to help them all.
> > If the answer to #1 is Yes and #2 is No, then the world will be hell of
> > human being.
> 
> Gee, what's wrong with this analogy?  In these cases, the answer to #1 is NO. 
>  That is, it is *not* necessary to solve the problem "completely" (i.e. cure 
> the patient "completely" or feed every last citizen) in order to effect a 
> significant improvement in the welfare of the patient or nation.

So please don't use a absolute attitude to analysis a real issue.
 
> By contrast, CJK users may well expect (demand) SC/TC equivalence to work for 
> all Han characters if it works for any, and so the partial solution may be 
> worse than none.

Maybe.
If we do nothing, everyone will frustrate while he use CDN.
If we do something, maybe 20% will frustrate.
So we should make a decision whether to do it or not.

Regards
Deng xiang