Dear all,
The W3C HTML Working Group is developing a new version of HTML,
and proposals are on the table with implications for whether
DCMI would be able to apply solutions it has promoted for use
with HTML4 and XHTML in a new HTML5 environment [1].
The following statement, which puts DCMI's HTML-related work
into an historical context and articulates a position on the
issues as we currently understand them, is signed by the authors
of DCMI's current HTML-related specifications and by the
co-moderators of the DCMI Architecture Forum.
I am posting this message now to dc-architecture in order to
invite a close reading by interested members of this forum and
would particularly appreciate quick feedback if anyone spots
errors of fact.
Discussion of this issue has recently come thick and fast, with
new developments and contributions not only on the mailing list
of the W3C HTML working group [2,3], but also the mailing lists
of the W3C Linking Open Data project [4,5,6,7] many of which
have been Cc'd to the mailing list of the W3C Semantic Web
Interest Group [e.g., 8].
Unless anyone raises red flags, I plan to submit this post early
next week to all of the relevant mailing lists. Recipients Cc'd
to this message are invited to join the dc-architecture list [9]
if they wish to respond to the dc-architecture group as a whole.
Many thanks,
Tom Baker
[1] http://lists.w3.org/Archives/Public/public-html/2010Jan/0573.html
[2] http://lists.w3.org/Archives/Public/public-html/2010Feb/0683.html
[3] http://lists.w3.org/Archives/Public/public-html/2010Feb/0794.html
[4] http://lists.w3.org/Archives/Public/public-lod/2010Feb/0207.html
[5] http://lists.w3.org/Archives/Public/public-lod/2010Feb/0215.html
[6] http://lists.w3.org/Archives/Public/public-lod/2010Feb/0218.html
[7] http://lists.w3.org/Archives/Public/public-lod/2010Feb/0220.html
[8] http://lists.w3.org/Archives/Public/semantic-web/2010Feb/0214.html
[9] http://www.jiscmail.ac.uk/lists/dc-architecture.html
--------------------
[Text to be posted:]
Over the years, DCMI has developed several specifications
of relevance to the discussions around HTML5:
-- RFC 2731 of 1999 ("Encoding Dublin Core Metadata in HTML") [1],
now obsoleted by RFC 5791 ("RFC 2731 Is Obsolete") of 2010 [2].
-- "Expressing Dublin Core in HTML/XHTML meta and link elements" of 2003 [3],
currently a "superseded recommended" as explained in [4].
-- "Expressing Dublin Core metadata using HTML/XHTML meta and link
elements" (a.k.a. DC-HTML) of 2008 [5], currently a DCMI Recommendation.
The 2008 DC-HTML Recommendation makes use of the XHTML meta data
profile, which is associated with a Profile Transformation as
defined by GRDDL.
DCMI acknowledges that all technology is transitional, and that,
in particular, syntax for embedding metadata in HTML is likely
to evolve significantly over the next few years.
The intention of DCMI is to help metadata implementers use
metadata effectively in the Web environment. While
DCMI-specific solutions have been developed over the years to
meet pressing needs, DCMI's intention is not to promote
syntactic solutions that are specific to "Dublin Core", but
rather to enable the use of Dublin Core terms in metadata that
draws on other compatible namespaces, such as FOAF, in the
context of application profiles.
Increasingly, such syntax solutions are being developed outside
of DCMI. Since the finalization of RDFa as a W3C Recommendation
in 2008, DCMI also promotes RDFa for use in "Dublin Core"
metadata as an alternative to DC-HTML/2008.
In this spirit, DCMI does not consider itself to be committed
to the profile attribute in a way that would preclude migration
to alternative mechanisms in HTML5. DCMI is open to any forms of
markup that will meet the needs of the Dublin Core community.
In our understanding, the need of the Dublin Core community is
for a mechanism to encode RDF triples using <meta> and <link>
elements which:
-- supports both literal and URI objects
-- includes an abbreviation mechanism for property URIs, not
just for property URIs owned by DCMI, but for property URIs
generally.
The 2008 DC-HTML specification provides such a mechanism. RDFa
provides such a mechanism through CURIEs. If RDFa were
available in HTML5, DCMI would therefore be happy to promote its
use with HTML5.
If the new microdata proposal were to provide a mechanism for
achieving the same effect, then DCMI would be happy to promote
its use as well. Based on an imperfect understanding of the
proposal on the table, DCMI does have some concern that
requiring the use of full URIs for qualifying the provenance of
properties sets a high bar for implementers, who could resort to
pragmatic solutions that would depend on out-of-band information
or other forms of "context" to relate the properties used
specifically to DCMI metadata terms and could therefore, in
practice, introduce ambiguity about the precise terms referenced.
DCMI is interested in promoting any generic, standards-based
solutions which ensure that its properties are precisely
identified and can be referenced and used without ambiguity.
Best regards,
Tom Baker (CIO of DCMI), Pete Johnston, Mikael Nilsson, Andy Powell
on behalf of DCMI
[1] http://tools.ietf.org/html/rfc2731
[2] http://tools.ietf.org/html/rfc5791
[3] http://dublincore.org/documents/dcq-html/
[4] http://dublincore.org/documents/dc-html-notes/
[5] http://dublincore.org/documents/dc-html/
[6] http://dublincore.org/documents/dc-ds-xml/
--
Thomas Baker <[log in to unmask]>
|