Contextual rule for MODIFIER LETTER PRIME?

[Chris Wright]

John,

I have to ask, why would anyone actually do this? No-one would ever be able type the name, unless we made a mapping rule that maps apostrophe to <U+02B9> which I doubt would ever happen. The only way this name could reasonably be used is if it was sent in emails, or in hyperlinks on websites. So I guess then the next argument would be phishing? Someone wants to pretend they are oreilly.tld by pushing out hyperlinks with o'reilly.tld (where ' is <U+02B9>). I don't see this as any different to what can happen today, I can pretend I am bobscars.com by pushing out hyperlinks like bobs-cars.com or bobcars.com or even bobscar.com. As a registry operator we already deal with these problems, and have established technical and administrative policies for preventing the registration of, and dealing with the ones that get through.

Finally, why would a registry ever allow this name to be registered? In your specific example the <U+02B9> is being mixed with ASCII characters, and I would find it hard to believe a registry would define a 'language' that included that set of code points, more likely it would be defined in a language set such as Greek. Even then the registry would have to come up with sensible, basically what I would call 'context' or policy rules for where this code point could be used.

As for lower levels of the DNS hierarchy where it is unlikely any rules that we come up with are enforced, your example can already happen now with the drafts as is, because given that the current context rule for the MODIFIER LETTER PRIME is not applied at lookup, and assuming lower levels will not apply registration only rules, and let's face it they never will, we will be lucky if they even know they exist, the name will be created in a lower level zone file, and nothing will stop it from being looked up. Thus we are no worse off, also phishing becomes less of a problem down at this level, as the registrant is only hurting themselves (i.e. it is one administrative domain, so they are only stealing traffic/pretending to be, themselves).

I hope I have explained myself clearly

Chris


-----Original Message-----
From: idna-update-bounces at alvestrand.no [mailto:idna-update-bounces at alvestrand.no] On Behalf Of John C Klensin
Sent: Friday, 17 July 2009 7:46 AM
To: Michel SUIGNARD; Patrik Fältström
Cc: idna-update at alvestrand.no; Kenneth Whistler
Subject: RE: Contextual rule for MODIFIER LETTER PRIME?



--On Thursday, July 16, 2009 03:43 -0700 Michel SUIGNARD
<Michel at suignard.com> wrote:

> I agree with the position expressed here. As long as 02B9 is
> allowed/requested in some context, it is just as easy to allow
> it in the protocol. Context conditions are just too
> complicated for that character and are bound to be
> controversial. I am too enthused though because of its
> resemblance to apostrophe (0027), but on the other end it will
> make supporters of many writing systems very happy.

I can live with this, but assume that it will lead directly to
labels that contain
    O<U+02B9>Reilly
in order to simulate
    O'Reilly

The perceived availability of such labels in the protocol will
almost certainly cause demands to permit that sort of
arrangement in SLDs and TLDs, not just down in the tree where
enterprises make up their own rules.  Since operating systems
won't parse it as an apostrophe or single quote, maybe there
will be no significant damage, but it makes me very anxious
indeed.

   john

_______________________________________________
Idna-update mailing list
Idna-update at alvestrand.no
http://www.alvestrand.no/mailman/listinfo/idna-update