[uf-discuss] HTML5, Microformats and RDFa

Tue Aug 26 13:01:45 PDT 2008

On 25 Aug 2008, at 19:47, Manu Sporny wrote:
> There have been several threads discussing Microformats, RDFa and 
> HTML5
> that are occurring on the WHATWG mailing list. The discussion 
> relates to
> whether or not HTML5 should depend on the Microformats community to
> solve HTML5's semantic markup issues, or if both Microformats and RDFa
> should be considered for semantic web markup issues.

I've been out of touch with HTML5 development for a bit, but the way 
you describe this paragraph is somewhat alarming.
We, the microformats community, absolutely *should not* be relied on 
the fill every gap in HTML. That they would not specify minority 
concerns in the HTML language is perfectly understandable, but the 
Microformats Community is itself not designed to do that either. This 
community, with this development process, is completely inappropriate 
for filling every single extended use for HTML that people might have.
HOWEVER, there may just be misinterpretation here. Perhaps rather than 
intending to depend on our specific community, the intention is that 
the gaps be filled with ‘microformat-like patterns’. Patterns, class- 
patterns, ‘posh’… whatever you want to call it. Microformats.org does 
not own the class attribute and anyone working on techniques that are 
incompatible with our process can do so.
It seems to me the case is not about ‘microformats.org’, but instead 
about the capabilities of the class attribute itself. Is it just that 
the word ‘microformats’ is being used as a generic catch-all for 
semantic class name patterns?
It seems quite reasonable that the HTML working group be considering 
the use case of ‘extended semantic description in HTML’ and 
considering its existing capabilities (which are proving very capable 
in the specific case of microformats), rather than a use case of 
‘support RDFa in HTML’, which is just one solution.
I think Scott is correct in that you may need to reframe your 
argument. Any push to have RDFa made a part of HTML5 should be focused 
on the capabilities of RDFa compared to the class attribute, not the 
(often intentional) limitations of one particular user of the class 
attribute (us).
Ben