Clipboard (was: Re: [idn] RE: [idn] Debunking the ACE myth)

[Martin Duerst <duerst@w3.org>]

At 22:10 01/07/23 -0700, Yves Arrouye wrote:
> > You're going to rely on the dq-- or whatever prefix to find
> > names in free
> > text? Now I like IDNA even less than I did before.
>
>FWIW a lot of MUAs do this today to let you follow a link in plain text
>e-mail. They recognize URIs, even www.*, ftp.*, ..., host names, and turn
>them into links. Nothing new.

Yes, nothing new, and it works well enough to be more of a help than a
nuissance, to a human user who can apply final judgement. But it does
not work well enough to build a protocol on it, such as the 'clipboard
protocol' that is implicitly proposed.

If we really want to go down
that path, I would expect IDNA to contain very exact and explicit
details about how exactly that would work. Currently, clipboards
are covered in a sentence or so; to have any chance
to make this work, at least a page or so may be necessary. Also, please
note that detecting ACE isn't the only thing necessary. We also have
to be able to detect arbitrary IDNs in free text. Assume we have
application A, which is a new application and therefore works
only with UTF-8, throughout, except for the calls to the DNS
(which are most probably done via an UTF-8+nameprep resolver
interface). Now this application gets some free text that contains
some IDN. Some of this free text is then cut or copied into
the clipboard. To adhere to IDNA, the application has to know
which parts of the free text are IDNs. Can it really do that?
If yes, how exactly? Can those who think it should work please
write up a draft?

Also, please think about the following:
How many 'clipboard protocols' has the IETF already defined?
Is the IETF involved in/does it have experience in defining
'clipboard protocols'? Does the IDNA draft require any other
protocol to use ACE for domain names? Does the IDNA draft
describe how any other protocols (except DNS itself) deal
with IDNs?

So one solution may be to just not speak about clipboards
at all.


Regards,    Martin.