W3C home > Mailing lists > Public > public-i18n-its@w3.org > July to September 2006

RE: Exploring the change from Ref to Uri

From: Yves Savourel <ysavourel@translate.com>
Date: Wed, 26 Jul 2006 07:51:44 -0600
To: "'Richard Ishida'" <ishida@w3.org>, "'Martin Duerst'" <duerst@it.aoyama.ac.jp>, "'Felix Sasaki'" <fsasaki@w3.org>
Cc: <public-i18n-its@w3.org>
Message-ID: <001001c6b0ba$a1eff0e0$9b05a8c0@Breizh>

Hi everyone,

OK, I see now that my early thought that changing to xyzURI was possibly a good thing was misguided :)
Felix' and Martin's arguments seem convincing. I also think that Felix is right about the name describing the function rather than
the type (after we do recommend something similar for naming elements... <strong> not <bold>).

With that in mind:

'Ref' clearly indicate a reference. Whether the actually reference is expressed as an IRI or something else is secondary (still
important, but secondary). Actually (as far as I know) ITS stops at providing the value it does not follow it. 

'Pointer' clearly points to something. Yes, 'ref' and 'pointer' are a bit difficult to distinguish at first, as Richard says, but
I'm not sure we could find other term(s) that would make things clearer.

So, I would propose nothing about the names (we were just 'exploring' the possibilities of a change not proposing one yet) and
address the i18nCore issue (#3494) more directly by trying to make the text more specific (which is already an action item for Felix

One last note:

At least this issue had the merit to bring up IRIs vs. URIs. We do have one attribute called 'uri' in the <its:ns> element. It has
the same type of termInfoRef: do we need to do something explicit about talking about IRIs instead of URIs in this case (and in the


-----Original Message-----
From: Richard Ishida [mailto:ishida@w3.org] 
Sent: Wednesday, July 26, 2006 5:29 AM
To: 'Martin Duerst'; 'Felix Sasaki'; 'Yves Savourel'
Cc: public-i18n-its@w3.org
Subject: RE: Exploring the change from Ref to Uri

Let's pull back a bit.  I think part of the problem is that it is possible (even we during our calls sometimes) to get confused by
the differences between things like termInfoPointer, termInfoRef, and termInfoRefPointer.
As I understand it, anything ending with Pointer has an Xpath expression as its value, and anything ending with Ref expects to point
to an IRI.

In some ways, using IRI and Path, or Href and Location, or some such, might help keep the distinctions clearer (though there may
well be better terms).
It's not intuitively apparent that there's a difference between a ref and a pointer.

Btw, Felix, these other vocabularies you looked at, what did they use, apart from HTML and XLINK?  Note that in xmlspec termRef's
value is an IDREF, not a URI, so I'm thinking maybe that following your line or argument this would be a good reason to avoid Ref
too. (?)

Received on Wednesday, 26 July 2006 13:52:02 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:04:11 UTC