This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.

Bug 3050 - Linking mechanism
Summary: Linking mechanism
Status: CLOSED FIXED
Alias: None
Product: ITS
Classification: Unclassified
Component: ITS tagset (show other bugs)
Version: WorkingDraft
Hardware: PC Windows XP
: P2 normal
Target Milestone: LastCall20May
Assignee: Felix Sasaki
QA Contact: Felix Sasaki
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2006-03-27 06:11 UTC by Felix Sasaki
Modified: 2006-07-24 10:32 UTC (History)
0 users

See Also:


Attachments

Description Felix Sasaki 2006-03-27 06:11:29 UTC
Background: http://www.w3.org/2006/03/22-i18nits-minutes.html#item04

Purpose: A mechanism to link from an XML document to an external file containing ITS global rules.
Proposal: Use an element ITS link:

<its:link href="someRules.xml"/>
Comment 1 Sebastian Rahtz 2006-03-27 11:25:58 UTC
since the rules elements are the same in the external file as the instance file, the rules could be referenced using XInclude very easily. WHy invent a new mechanism?
Comment 2 Felix Sasaki 2006-03-27 11:55:27 UTC
(In reply to comment #1)
> since the rules elements are the same in the external file as the instance
> file, the rules could be referenced using XInclude very easily. WHy invent a
> new mechanism?
> 

Because people might want to implement ITS without having an XInclude processor available. However, most of them will have in their programming language a function to access a document, which takes as an argument the document Uri. Hence, the @href attribute.
Comment 3 Yves Savourel 2006-03-27 15:29:18 UTC
Then maybe using XLink's simple link would be better: not re-inventing something and keeping it processable by simple tools. I'm not very familiar with XLink but it seems we should be able to do:

<its:link xlink:href="someRules.xml"/>

Comment 4 Yves Savourel 2006-03-27 17:55:01 UTC
<its:rulesLink> sounds good to me.

One question: Where does it go? stand-alone (wherever the host format allows it)? or as the first entry in a <its:documentRules>?

For being sure we look at all the possibilities:

- This link could be also done or as a standalone attribute: <myDoc its:link="href">

- Or as a PI (like for CSS): <?its-rules href="file.xml" ?>

The PI has the advantage of working everywhere (always valid even in DTD-based formats).

Comment 5 Felix Sasaki 2006-03-27 22:00:28 UTC
(In reply to comment #4)
> <its:rulesLink> sounds good to me.
> 
> One question: Where does it go? stand-alone (wherever the host format allows
> it)? or as the first entry in a <its:documentRules>?

I think it would not make sense in <its:documentRules>. Rather standalone.

> 
> For being sure we look at all the possibilities:
> 
> - This link could be also done or as a standalone attribute: <myDoc
> its:link="href">

I would prefer XLink, as Sebastian proposed at http://www.w3.org/Bugs/Public/show_bug.cgi?id=3050#c3 .

> 
> - Or as a PI (like for CSS): <?its-rules href="file.xml" ?>
> 
> The PI has the advantage of working everywhere (always valid even in DTD-based
> formats).

What do you mean by "valid"? Processable? That would be the case with XLink too.

> 

(In reply to comment #3)
> Then maybe using XLink's simple link would be better: not re-inventing
> something and keeping it processable by simple tools. I'm not very familiar
> with XLink but it seems we should be able to do:
> 
> <its:link xlink:href="someRules.xml"/>
> 

++1
Comment 6 Yves Savourel 2006-03-27 23:02:06 UTC
>> The PI has the advantage of working everywhere 
>> (always valid even in DTD-based formats).
>
> What do you mean by "valid"? Processable? That 
> would be the case with XLink too.

You're right. What I meant to say was different:
Using a PI would allow DTD-based formats to use the ITS-link function even when the users cannot modify the DTD. The DTD does not to know anything at all of ITS, while with the link element (or attribute) it needs to know at least about that.

-ys
Comment 7 Felix Sasaki 2006-03-28 06:12:18 UTC
(In reply to comment #6)
> >> The PI has the advantage of working everywhere 
> >> (always valid even in DTD-based formats).
> >
> > What do you mean by "valid"? Processable? That 
> > would be the case with XLink too.
> 
> You're right. What I meant to say was different:
> Using a PI would allow DTD-based formats to use the ITS-link function even when
> the users cannot modify the DTD. The DTD does not to know anything at all of
> ITS, while with the link element (or attribute) it needs to know at least about
> that.
> 
> -ys
> 

Ah, o.k., I see. The disadvantage is just the other way round: you cannot validate a PI. Hence, if somebody types
<its:rulesLink xlink:hrefs="someUri"/>
it is an error since "hrefs" should be "href". However, if you type
<?its-rules hrefs="file.xml" ?>
nobody will find the error.

I am wondering what is more important here: The easy of validation, or being on disturbance to validation ...
Comment 8 Sebastian Rahtz 2006-03-28 18:13:22 UTC
> I am wondering what is more important here: The easy of validation, or being on
> disturbance to validation ...

My view is that people who want minimal disturbance can
specify the rules files some other way (in their
processing tools); people who want decent XML
validation should be encouraged to do it well. But I am not rabid on the subject
Comment 9 Yves Savourel 2006-03-28 18:34:47 UTC
I'm fine with <its:rulesLink xlink:href="someUri"/>.
I just wanted to be sure we had looked at all possibilities.
Comment 10 Yves Savourel 2006-04-07 20:12:22 UTC
Resolution: The link is to be an xlink:href attribute in <its:rules>.
(See http://www.w3.org/2006/03/29-i18nits-minutes.html#item08)
Comment 11 Felix Sasaki 2006-07-24 10:32:46 UTC
Closed, no further action necessary.