W3C home > Mailing lists > Public > www-archive@w3.org > February 2007

Re: How to link W3C Technical Reports

From: Ian B. Jacobs <ij@w3.org>
Date: Tue, 27 Feb 2007 21:53:28 -0600
To: Bjoern Hoehrmann <derhoermi@gmx.net>
Cc: w3t-comm@w3.org, www-archive@w3.org
Message-Id: <1172634808.32264.31.camel@localhost>
On Wed, 2007-02-28 at 04:09 +0100, Bjoern Hoehrmann wrote:
> * Ian B. Jacobs wrote:
> >This has happened to me in practice. I don't think there's any
> >perfect solution (until the day we can do redirects with
> >fragids as Bert Bos has suggested [1]).
> >
> >One idea: redirect the URI to section 8.15 to a "changes" chapter. In
> >that chapter, explain in prose what changed and where to turn. You might
> >even have a fragid in the changes chapter so that user agents that
> >re-append the fragid on the redirected URI land right on the explanation
> >they seek.
> 
> Odiferous Style Sheets 2.0 will be a great technology, but it will take
> several years from publication of the specification to widespread de-
> ployment of client software. In the meantime most people will continue
> to write OSS 1.0 style sheets. During that time readers of my OSS 1.0
> tutorial will not appreciate if they follow links to the OSS 1.0 spec
> that turn out to be links to the changes section of the OSS 2.0 spec.

They should use the dated URIs then.

> It's nice to know what will change, but to get their job done using OSS
> 1.0 they really need the OSS 1.0 specification.

It will still be available. And perhaps the 2.0 specification can
point back to it to be sure people following the latest version URI
can find it.

> Even if section 8.15 is kept intact in OSS 2.0 but with new features
> added, my readers will get confused about it, they will try new 2.0
> features in their 1.0 clients and it won't work as they expect.

Hmm, I think you are making the scenario harder than it has to be.
Sounds like the 2.0 text just needs to be clear.

>  I know
> because I once spend an afternoon trying to make the HTML 3.0 <banner>
> element work in pre-HTML 3.2 browsers. My book had it, my editor had
> it, I saw it in tutorials, it must work! But didn't. So I'm afraid this
> is not a viable solution.
>
> I need a URL that always resolves to the latest edition of a particular
> Recommendation. Why is there no such URL for current Technical Reports?

What do you mean by 'latest edition'? Do you mean the latest of a
particular major revision number (2.x) or do you mean the latest and
greatest (say, 3.0 if that's the most recently published one)? 

We propose both URIs here:
  http://www.w3.org/2005/05/tr-versions#tracking

Some Recs have started to use this.

> Isn't the type of link I want to make much more common than linking to
> "the latest edition of this Recommendation, or a PER of it, or a draft
> for the next version of the technology, or a placeholder document, or
> something else entirely, depending on the phase of the moon and how the
> group managing the technology feels like"?
> 
> There is a simple solution of course, once a shortname resolves to a
> REC, it will always resolve to the latest edition of that particular
> Recommendation. Shortnames for anything else, like "latest publication
> in the area of Odiferous Style Sheets" or "latest Odiferous Style Sheets
> Recommendation" either do not exist or, if you die without them, are
> much less prominent and redirect to some other shortname. I proposed
> just that two years ago when we discussed this in the context of SMIL2.
> 
> If that is not possible I will probably remove all links to shortnames
> in the specifications I edit. As an example, I cannot have a bib entry
> like 
> 
>   [XML Events] 
>     XML Events, S. McCarron, S. Pemberton, and T.V. Raman, Editors.
>     World Wide Web Consortium, October 2003. This version of the XML
>     Events Recommendation is http://www.w3.org/TR/20... The latest ver- 
>     sion of XML Events is available at http://www.w3.org/TR/xml-events.
> 
> in DOM Level 3 Events because what is available at the latter URL is a
> "XML Events 2" Working Draft, which to me is by no means the latest ver-
> sion of "XML Events" and much less of the Recommendation I ment to re-
> ference. Similarily, "The latest version of XML 1.0 is available at
> http://www.w3.org/TR/REC-xml" must be removed also because tomorrow you
> might find a XML 2.0 Working Draft there, or a placeholder page that
> explains that in XML 2.0 DTDs are in one spec and the rest in another.

I think you should use dated URIs when you want to refer people to 
something that never moves (e.g., for conformance), and you should
use latest version URIs when you want people to find something newer,
and explain to them on arrival what has changed and where to find
what they need. The semantics of the latest version URI need to be
clear, which is why we propose 2 in the versioning document above.
I've heard people desire even more (e.g., groups may wish to know
the latest version of a particular minor revision thread), but for
now we have only suggested two.

 _ Ian

-- 
Ian Jacobs (ij@w3.org)   http://www.w3.org/People/Jacobs
Tel:                     +1 718 260-9447

Received on Wednesday, 28 February 2007 03:53:36 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 7 November 2012 14:18:02 GMT