- From: Martin J. Dürst <duerst@it.aoyama.ac.jp>
- Date: Tue, 31 Mar 2009 19:32:13 +0900
- To: Mark Nottingham <mnot@mnot.net>
- CC: HTTP Working Group <ietf-http-wg@w3.org>, Apps Discuss <discuss@apps.ietf.org>, public-iri@w3.org
[copied to public-iri@w3.org, which should be used for all public discussion related to IRIs] On 2009/03/30 18:08, Mark Nottingham wrote: > Various folks from the IETF and W3C's HTML5 WG got together in San > Francisco last week to discuss the parts of that work. > > Rough minutes are at: > http://esw.w3.org/topic/IETF_HTML5_Meeting_March_2009 From there: >>>> - HTML5's URI section (DanConnolly and Larry Masinter to work on this; e.g. HTML WG action 68) - We discussed IDN and URI/IRI (international domain names vs. HTML5/W3C use of IRI). Changes to IRI would impact specs like Atom. Larry advocated revising this spec, others were less enthusiastic. It would be a big undertaking, and it wasn't clear that Martin Dürst was available. - Rob Sayre suggested the name "Hypertext References". This was met with wide approval. - Action Item: Dan to reformat the document as an Internet Draft >>>> Some comments: - My ability is limited, but not zero. Just recently, I had a major blackout period due to upgrading of notebook/OS/emailer/... Not quite out of it yet. (At least, I'm now able to write my name correctly in email after all these years of working hard for internationalization on the Web :-(. - I managed to absorb a major 'variant' of IRIs for the XML folks under the name 'Legacy Extended IRIs' (see http://tools.ietf.org/html/draft-duerst-iri-bis-05#section-7). Much of that text was originally from the XML side (thanks to Norm Walsh, Richard Tobin, Henry Thomson), but I tweaked it quite a bit for tone, and we went back and forth to try and make sure we had all the differences covered. I think the exercise was successful in the sense that it was satisfactory for the parties involved (XML specs community and IRI spec authorship), and paper (or these days electrons) is patient anyway. I'm still not sure what kind of reaction there will be from a wider community (hint, hint: feedback welcome!) - The motivation for doing the above were about as follows: - It's better to have everything in place, so people can look it up in one go. - It's better to have it under the same name, and not send the wrong message with names (the originally proposed name for LEIRIs was "human readable identifiers", which was misleading in many ways) - It's better to allow it but warn against it than to ignore it in silence. - The IRI spec (some pre RFC 3987 drafts) allowed spaces and some other strictly ASCII non-URI characters, which was the reason they got allowed is some XML specs, so part of the reason was that the IRI spec also bore some responsibility. - Last time I looked at the discrepancies between the URI/IRI specs (RFC 3986/7) and the HTML practice (as far as documented), my impression was that some parts of it could rather easily be absorbed/buffered in the IRI spec, but for some others, the URI spec would be more appropriate. As an example, I think what HTML browsers do with buggy %-encoding sequences has nothing to do with the first I in IRI, which stands for Internationalization. So much for the moment. Regards, Martin.
Received on Tuesday, 31 March 2009 10:34:22 UTC