Re: [dnsext] we need help to make names the same, was draft-yao-dnsext-identical-resolution-02 comment

[Alex Bligh <[e-mail]>]

--On 16 February 2011 12:14:30 -0500 Andrew Sullivan <[e-mail]> 
wrote:

> On Wed, Feb 16, 2011 at 07:43:30AM +0000, Alex Bligh wrote:
>> (*) I think it's really about orthographic representation rather than
>> spelling: we aren't really talking about jail.com and gaol.com, are
>> we?
>
> As co-chair but without having discussed with Olafur: We are talking
> about whatever zone administrators want to be equivalent, I think.

Yes. You are quite correct. What I meant to say was treating this
as equivalencies in "spelling" misses the point; most of the use cases
now are about orthographic representation (I think). However *if*
we solve this, we need, as you say, to solve the general equivalence
problem, as ...

> We are not well enough versed in the
> linguistic differences between two different spellings "in English"
> and the two scripts that are used to write Chinese for us to begin to
> make decisions on the basis of that sort of distinction.

+1

> This is why I, at least, have been adamant that we need a problem
> statement before we start trying to solve the problem.  It would be a
> bad thing to add a bunch of complications that don't actually address
> the real problems.

I'm not sure I follow that. If we don't understand the existing known
potential users (Chinese, Greek etc.), how can we pretend to understand
the requirements of users we don't know? Surely, the /only/ logical
thing to do is to solve "the general case", where an arbitrary number
of sets of labels each are equivalent inside that set.

> This is also why I have been so unhappy that we
> have attracted little interest from the applications area crowd, who
> will have to use anything we end up delivering.  If we can't tell
> whether what we are talking about looks like a real problem to (say)
> web browser writers, how will we know whether what we propose to do is
> going to have any effect?

+1

-- 
Alex Bligh
_______________________________________________
dnsext mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/dnsext