MVALID (was Re: M-Label or MVALID, and dangers with mappings?)

Pete Resnick presnick at qualcomm.com
Tue Apr 14 20:58:37 CEST 2009


On 4/14/09 at 7:52 AM -0700, Mark Davis wrote:

>Having arbitrary mappings, subject to "application context" just 
>makes a messy situation worse: if program 1 sees 
>href="http://ÖBB.at" and goes to http://öbb.at, but program 2 goes 
>to the different http://oebb.at, we have a horrible interoperability 
>problem.

First of all, you should have a look at the ripe mess your email 
client turned those IRIs into. (It chose "%" encoding, which is quite 
fun.) But leaving that aside:

Who said *ANYTHING* about *ARBITRARY* mappings? Straw man fallacy 
here. As far as I can tell, we *are* writing a mappings document and 
it *will* map "ÖBB" to "öbb". For program 2 in your example to choose 
not to implement that mapping would mean that either (a) it had 
discovered that more people using the Lower-Slobovian version of its 
web browser want to go to "oebb.at" instead of "öbb.at" and its users 
scream if it tries to send them to "öbb.at" or (b) they are stupid, 
its users will scream that they are not getting to the right web 
site, and it will be fixed. This is *all* about user expectations and 
*not* usurping them by insisting that there is one and only one way 
for a user to type a name, and all other ways that we haven't thought 
of in 2009 are forbidden to implement.

>I had thought that we were making some progress towards having a 
>uniform single mapping handling at least case and width, which users 
>really want (Cf the CJK presentation at the latest IETF meeting).

Argumentum ad populum. Nice. Do we get to cover all of the informal 
logical fallacies today?

First of all, on what basis are you claiming that *you know* what 
"users really want"? The CJK presentation was probably about the 
desires of the registry community, but I'm not at all sure what it 
says about the desires of the end users of most domain names. So lets 
just drop these histrionics about what "users really want." What 
"users really want" is web browsers that get them to the place they 
want to go when they type what they think will get them there. My 
guess is that most users don't know what they want at the protocol 
level.

And for that matter the CJK presentation started with the assumption 
that there was *NO* mapping. If we introduce a mapping as some of us 
have described here (one that is not necessarily written in stone for 
the rest of time), perhaps that would satisfy the concerns expressed 
in that presentation. In the last slide of that presentation, most of 
what I think of as the mapping step was put in a box labeled "BCP". 
That doesn't sound like part of the protcol proper to me.

>Mapping is not just a UI issue.

Another straw man. I never said that mapping was *just* a UI issue. 
But it is in large part a UI issue in this context.

pr
-- 
Pete Resnick <http://www.qualcomm.com/~presnick/>
Qualcomm Incorporated - Direct phone: (858)651-4478, Fax: (858)651-1102


More information about the Idna-update mailing list