W3C home > Mailing lists > Public > public-iri@w3.org > February 2004

RE: IRIs and bidi

From: Michel Suignard <michelsu@windows.microsoft.com>
Date: Fri, 6 Feb 2004 10:04:48 -0800
Message-ID: <84DD35E3DD87D5489AC42A59926DABE905B5F778@WIN-MSG-10.wingroup.windeploy.ntdev.microsoft.com>
To: "Jony Rosenne" <rosennej@qsm.co.il>
Cc: "Unicore" <unicore@unicode.org>, <public-iri@w3.org>

Jony, there seems to be some miscommunication or maybe I wasn't clear in
my messages. RTL IRIs are allowed, they just have to be rendered in a
LTR embedding level, and furthermore their rendering should have no
dependency on the surrounding text, and shouldn't interfere with
surrounding text either.

It is perfectly valid to enter a Bidi IRI in a RTL environment in
logical order. There are only 2 caveats:
- some restrictions have been put on IRI fragment concerning the mix of
RTL and LTR characters, similarly to the ones put on IDN
- once an IRI is recognized as such by the text processing engine, it
must render according to the rules expressed in the IRI spec.

If an IRI (or IRI fragment) is not recognized as an IRI by the syntax
analyzer (like having an invalid syntax), the rendering would default
back to whatever the bidi rendering of the original text is (plain text
or higher protocol).


> -----Original Message-----
> From: unicore-bounce@unicode.org 
> [mailto:unicore-bounce@unicode.org] On Behalf Of Jony Rosenne
> Sent: Friday, February 06, 2004 9:35 AM
> To: Unicore; public-iri@w3.org
> Subject: Re: IRIs and bidi
> Martin,
> The point is that I disagree with "the overall LTR direction 
> required for IRIs". In an RTL environment, be it Hebrew or 
> Arabic or any other RTL script, RTL IRIs should be allowed.
> You are right, I meant:
> Since it is entirely RTL, in an RTL environment I would just 
> type it in plain logical order.
> A person or a child who only knows his own language should be 
> able to use the internet, even if his native script is Hebrew 
> or Arabic. We have many people here who cannot read English, 
> and I'm certain the same is true for the Arab countries, Iran 
> and Pakistan.
> Jony
> > -----Original Message-----
> > From: Martin Duerst [mailto:duerst@w3.org]
> > Sent: Friday, February 06, 2004 6:40 PM
> > To: Jony Rosenne; public-iri@w3.org
> > Subject: RE: [bidi] Re: IRIs and bidi: Addition regarding 
> higher-level 
> > protocols
> > 
> > 
> > Hello Jony,
> > 
> > At 09:49 04/02/06 +0200, Jony Rosenne wrote:
> >  > Once we have internationalized TLDs, it it is conceivable, 
> > desirable and  > unavoidable to have RTL bidi IRIs.
> > 
> > The rest of your mail didn't show up on my (Japanese) mailer.
> > I went to the archive at
> > http://lists.w3.org/Archives/Public/public-iri/2004Feb/0002.ht
> ml
> to look at it.
> Using the usual 'upper-case for RTL' notation, what I saw was 
> (in a very crude transliteration):
> I can only identify the first and the last component, which I 
> think got placed right. The rest got messed up in my opinion. 
> Following the IRI spec, it should look like:
> This is based on the assumption that '/', '.', and ':' are 
> all weak, and that even the overall LTR direction required 
> for IRIs doesn't affect the total reordering if everything is RTL.
> Could you please clarify how you input this example, and how 
> it shows up at your end, or how you think it should show up?
> The above example of course is currently not allowed, and 
> would have to be rewritten to show up as:
>     http://MISM/LI.LSMM.DZVA
> Regards,   Martin.
Received on Friday, 6 February 2004 13:05:28 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:14:30 UTC