Backwards compatibility (was: Re: Wwhich RFCs the new work would obsolete, vs update or leave alone)

Mark Davis mark.davis at
Tue Mar 18 21:47:10 CET 2008

It does not have to be formally required in the protocol. I like the way the
current documents separate out any notion of preprocessing so that it is not
in the base protocol. And for many implementations it will be sufficient;
they will not require any preprocessing.

However, for many implementations the use of preprocessing will be
absolutely required in practice, and it had better be the same across them.

The main point is that preprocessing is not limited to UIs; there are a wide
variety of mechanical processes that interpret textual URLs - not only
browsers in href attributes, but many other programs do it in flowing text,
such as email clients, word processors, even spreadsheets. People often
paste what they have typed into an address bar into that text, and they will
expect the same degree of case-insensitivity and variant insensitivity that
has been present in IDNA2003.

I've said this before on this list, but for those joining later, here's a


On Tue, Mar 18, 2008 at 1:21 PM, Paul Hoffman <phoffman at> wrote:

> At 11:25 AM -0700 3/18/08, Mark Davis wrote:
> >For more information on possible preprocessing, see the rough draft
> >at
> The steps there look interesting. However, I'm *very* concerned with
> putting this forward with the current wording in the document, given
> that it goes against one of the core changes in IDNA200x from the
> design team, namely that pre-processing is optional. The fact that
> you mention the possible need for post-processing makes that even
> more poignant.
> If you mean this proposal to be optional, please reword it a bit more
> carefully for that. If you mean it to be required, then that might
> cause a reset of the charter.
> _______________________________________________
> Idna-update mailing list
> Idna-update at

-------------- next part --------------
An HTML attachment was scrubbed...

More information about the Idna-update mailing list