URI profiles [was RE: [uf-discuss] Comments from IBM/Lotus repabout Microformats]

Wed Dec 13 10:21:09 PST 2006

[sorry about that. Some keypress automatically SENT that last email,
which isn't what I meant to do.]

Scott Reynen wrote:
> So profile URIs are described here:
> 
> http://microformats.org/wiki/profile-uris
> 
> where it says:
> 
> "it is ACCEPTED that each microformat should have a profile URI."
> 
> I agree it would help to make that more clear, but if you're
> suggesting we change that "should" to a "must," I'd ask you what  
> practical benefit you expect publishers would gain from such a  
> change.  We're trying to avoid solving hypothetical problems here,  
> and I don't see a practical problem profile URIs solve yet, as I  
> haven't noticed anyone using class="vcard" to designate their  
> Valentine's Day cards or anything else other than hCard.  If you're  
> interested in seeing wider adoption of profile URIs, I'd recommend  
> work on filling in the XMDPs for every microformat, because it  
> wouldn't make much sense to require publishers to point to profiles  
> which don't exist.

Versioning is one practical problem, although the example Steve Marshall
brought up with hListing would suggest that you are absolutely right: we
should create XMDP profiles for every microformat. Another valid problem
is disambiguation between semantic html and "official" microformats. If
it doesn't have a microformats.org profile URI, it isn't a microformat.

I think however, that we can get profiles into usage even without the
XMDPs, as the provide value in the wild even if the URI doesn't resolve,
in the same way that rel="tag" works or namespace references use a URI
without requiring the URI resolve.

<head profile='http://gmpg.org/xmdp/samplehtmlprofile.html'>

That works for versioning and disambiguation regardless of whether or
not the URI actually resolves.

So, how about standardizing the URI for "official" microformats so that
they can be used in disambiguation and versioning?  That would also let
us know where we can start working on the XMDPs.

http://microformats.org/wiki/profiles/hcard
http://microformats.org/wiki/profiles/hlisting
http://microformats.org/wiki/profiles/xfn
http://microformats.org/wiki/profiles/hcite
Etc.

I don't know that this is the right addressing scheme, but it seems to
me that something like this makes sense. Then we can start using profile
URIs in the wild while simultaneously building those profiles on the
wiki.

Note that standardizing a URI scheme like the above would be useful even
if we don't change the policy on requiring profiles for compliant
microformats. It would make it easier to start working on the XDMPs, for
example.

After thinking about it, I would prefer something like this:
http://microformats.org/wiki/profiles/hcard
http://microformats.org/wiki/profiles/xfn
http://microformats.org/wiki/profiles/draft/hlisting
http://microformats.org/wiki/profiles/draft/hcite

With restricted access to the non-draft profiles.

Thoughts on how to organize official profiles?

-j

--
Joe Andrieu
joe at andrieu.net
+1 (805) 705-8651