Re: XLink 1.1: Animation

* Grosso, Paul wrote:
>XLink has no knowledge of animation or animatability.
>In general, XLink as a spec cannot--and should not have
>to--comment on every use other specs may make of it.

In order to use XLink with SMIL Animation this needs to be defined. If
the XML Core Working Group thinks this should be addressed in some other
specification, please coordinate with the relevant working group; should
they agree to include the relevant information, I'm happy to accept not
making such changes to XLink 1.1.

Likewise, if the XLink 1.1 draft clearly points out that XLink is not
appropriate for use in compound document environments as a broad range
of semantics are undefined at this point, and W3C does not intend to
remove these problems any time soon, I'm happy aswell.

If you say that general purpose XLink processing is not possible and
XLink implementations should apply XLink semantics only to elements for
which it is known how the XLink attributes apply to them and affect the
semantics of the document, that should address my concern also.

>Does this adequately address your comment, or do you wish for
>the XML Core WG to record your feelings on this matter as an 
>official objection when we request CR?

Yes, I think the XML Core Working Group made it very clear by now that
XLink 1.1 is not and is not going to be a specification that meets the
requirements to be reasonably integrated into compound document formats
or implemented in rich web clients that support a broad range of other
W3C technologies like CSS, SMIL, Scripting, DOM Events, and so on.

This is not news, as you point out, XLink already has a long and painful
history of not meeting basic requirements, it won't be used in formats
like XHTML and where formats like SVG adopted XLink, this in practise
relies on illegal markup and undefined semantics throughout, and is now
widely regarded as mistake. It's unfortunate the HTML Working Group
apparently never got around to send the comments they wanted to make.

It's also unfortunate that the XML Core Working Group missed the
opportunity to make this clear, e.g. Anne van Kesteren's comments in

received meaningful attention from the Working Group only after explicit
requests to formally address comments. Had the Working Group made it
clear in response to Anne that the XML Core Working Group is going to
reject any and all requests to make any non-trivial change but those
listed in the Requirements document, I would not have wasted my time
sending comments on this useless experiment, or tried as hard to
compromise as I did now.

In fact, the SotD section of the Last Call Working Draft is clear that
the intent of XLink 1.1 is to make it more useful in environments that
already use it. The changes I've requested are aimed to make exactly
that -- even though I was unconvinced that XLink 1.1 is of any use.

The set of changes that are part of the Requirements document are all
really errata, hardly any content ever respected the requirement to
specify the xlink:type attribute for simple links, IRIs were already
allowed, the schemas are editorial additions, and XLink 1.0's failure
to meet basic quality requirements to define and properly constrain
extensibility and extensions is simply an error.

I agree with Simon St.Laurent and George Chavchanidze

XLink is dead, dead-by-design, and our time would be better spent on
other things. The XML Core Working Group could, for example, make an
effort to initiate processes by which all the changes accumulated over
the many years in the 14 Recommendations it is chartered to maintain
become normative and republish the Recommendations as required by the
W3C Process. Indeed, non-normative errata hidden in the many errata
documents have caused considerable confusion and interoperability
problems over the years, yet only XML 1.0 ever received the benefit
of normative corrections.

In fact, it would already be helpful if comments on XLink 1.1 were
addressed even though "Working groups are not obligated to respond
formally to comments made on every working draft" as the Working Group
told me when I wondered what happened with Anne's comment and

It took me some time but I now realize that George raised exactly the
same issue that Jonathan Watt raised much later

>Nothing in our XLink 1.1 Requirements Document [1]
>relates to this issue, and nothing we did to XLink 1.0
>to create XLink 1.1 relates to this issue, therefore
>the XML Core WG has determined this is not an XLink 1.1
>Last Call issue.

Yes, I saw your suggestion in

that the XML Core Working Group simply omits substantive XLink 1.1
review comments from the XLink 1.1 LC DoC document in response to the
issue above. Now, I personally don't understand this approach, I tend to
work on the assumption that we all want the best possible specifications
and the process to be fun for everyone, and I'm fairly sure the process
would be more fun for Jonathan if he didn't have to raise this already
known issue, and XLink 1.1 would be better if this issue got addressed,
but then I do this in my spare time, so maybe that's the difference.

As you could guess by now, I don't think the XLink 1.1 draft should earn
The Director's endorsement, there is no point in increasing the version
number of XLink; if you really must, put those changes from your so-
important requirements document on the errata page and accept that XLink
is dead. In any case, please make an effort to address issues such as
those raised by George and Jonathan in a more meaningful way than to
point out that your charter doesn't require to work on them, this will
help a lot until we manage to get rid of the last XLink artifacts and
rescind the "not without its critics" XLink 1.0 Recommendation.

Björn Höhrmann · ·
Weinh. Str. 22 · Telefon: +49(0)621/4309674 ·
68309 Mannheim · PGP Pub. KeyID: 0xA4357E78 · 

Received on Wednesday, 25 January 2006 23:00:56 UTC