W3C home > Mailing lists > Public > public-iri@w3.org > May 2011

Re: requesting feedback regarding HTML5 and RFC 3987

From: Bjoern Hoehrmann <derhoermi@gmx.net>
Date: Fri, 27 May 2011 04:04:38 +0200
To: "Chris Weber" <chris@lookout.net>
Cc: <public-iri@w3.org>, "'Mark Nottingham'" <mnot@mnot.net>, "'Thomas Roessler'" <tlr@w3.org>
Message-ID: <9uvtt6plog98oejg96ipv9mfdn7oo2u634@hive.bjoern.hoehrmann.de>
* Chris Weber wrote:
>However, our understanding is that ISSUE-56 can be reopened if new
>information emerges, such as "IETF completing production of a
>document suitable as a formal reference".  And of course as chairs
>of the IRI WG we would like to deliver such a document.

My impression is that the Working Group largely lacks the resources to
do much to satisfy requirements of third parties beyond doing rudimen-
tary bug fixes. That includes a diverse set of active participants and
the necessary technical expertise as well as a good understanding of the
needs of any third parties. There are a number of problems with various
of the deliverables of the group, addressing those meets the needs of a
very broad community, I think that should be the primary concern here.

>Here is the minimum baseline that we understand is necessary in order
>to meet the needs of the W3C's HTML WG:
>
>1) The IRI specification will provide "MUST" language and normative
>   algorithms for parsing arbitrary Unicode strings as IRI against
>   an absolute base URL.

I don't understand what this means. It seems pretty clear to me, to give
some examples, that not all strings can be processed as if they're IRIs,
it is common for browsers for instance to reject certain strings as mal-
formed when you try to use them as resource identifier. Further, due to
bugs in widely deployed implementations some URI processing is specific
to the scheme that's employed, say, in `javascript:...#...` the `#` is
not considered to denote a fragment identifier in some contexts in some
implementations. I would most certainly disagree with the IRI WG as it's
currently chartered to include this kind of scheme specific rule.

>2) The IRI specification will define how to extract the hostname out
>   of an IRI for proper resolution and application of the same origin
>   policy.

This is already defined in the URI specification for all that I could
tell based on this brief description, and it would seem anything like a
"same origin policy" is certainly out of scope of the group.

>3) The IRI specification will define how base URI referencing would
>   be performed for hierarchical schemes.

I am not sure what "base URI referencing" is, but the URI specification
would seem to already address how to turn a relative reference into an
absolute one given a base reference. I do not see what "hierarchical"
schemes have to do with this.

>The chairs would like to request feedback from the group, especially
>those who are also participants in the HTML WG, about whether the
>those three deliverables would be sufficient to meet the needs of
>the W3C's HTML WG.

I am very sure the three points, even with my limited understanding of
them, would not be sufficient. As I understand it, the issue originated
with "There is all sorts of URI stuff that should be outsourced"; I've
not seen so much as a clear separation of what this working group might
work on versus what it would not. I gave an example above what I would
find out of scope at the moment.

Again, I would very much prefer if the working group would use the ex-
isting documents as a baseline, solicit bug reports, and then deal with
those, instead of trying to have some document referenced by "HTML5".
The people working on "HTML5" can and will solve what problems they have
and there is no need nor good reason to try to solve their problems for
them at this point. I would be more sympathetic to this if I regularily
saw people from Microsoft, Mozilla, Google, Apple, Opera, posting here,
but they are not.
-- 
Björn Höhrmann · mailto:bjoern@hoehrmann.de · http://bjoern.hoehrmann.de
Am Badedeich 7 · Telefon: +49(0)160/4415681 · http://www.bjoernsworld.de
25899 Dagebüll · PGP Pub. KeyID: 0xA4357E78 · http://www.websitedev.de/ 
Received on Friday, 27 May 2011 02:05:07 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 30 April 2012 19:52:01 GMT