[TagCommons-WG] Rel-tag

Nitin Borwankar nitin at borwankar.com
Thu Mar 8 12:06:12 PST 2007


Marja Koivunen wrote:

>I think sometimes it makes sense for the author to attach the metadata 
>directly on a document it refers to but I'm not sure if I understand all 
>the aspects and goals of the microformat use case here? For instance, is 
>is important that the tags are embedded in the same document or just 
>that the author of the document can provide that information as part of 
>the authoring?
>
>  
>
Hi Marja,

I am sure Kevin will correct me if I am wrong, but the 'original intent' 
of the creators of microformats was close to the following

a) To create a *machine readable markup format* that was also
b) *easily human readable*, as well as
c) *valid XHTML* which would enable
d) *web spiders to extract structured data* inside unstructured text on 
a web page, thereby enabling
e) bloggers to easily and *widely disseminate things like contact info, 
resumes, event info* .... and
f) *allow aggregators to build directories* of this data 
g)* without having to custom scrape each web page*

It was narrow-need-driven, intentionally lightweight and did not purport 
to solve all connected problems so it is not surprising that it does not 
fulfil the needs of other use cases. The question remains open whether 
it can be extended to meet other needs and whether this fits in with the 
interests and current agenda of the the microformats community.  Even if 
the microformats effort does not dovetail cleanly with the eventual 
direction of the TagCommons effort, it would be in all our interests not 
to duplicate work and also not to build unnecessarily competing formats. 

It is not clear that all microformats define metadata, just the rel=tag 
format probably does.
The ability to loose-couple page metadata to the page rather than embed 
it in the page seems more extensible and powerful.
It seems feasible that the rel=tag approach can be extended to allow a 
pointer to tag metadata rather than the tag metadata itself. 
(rel=tagbase maybe?).
Then other use cases could be accomodated such as allowing taggers other 
than the page author to add tags/annotations ... etc.
This also allows both efforts to evolve independently without setpping 
on each other toes and requiring one particular syntax or other in the 
metadata.
The kind (RDF, ...) of metadata store could be declared in another 
attribute, and multiple such tagbases would be permitted for a site, 
page etc.
Tagbase annotating a page could also exist without explicitly being 
mentioned in the re=tagbase tag, but these others would, possibly,  be 
ones not "blessed"
by the author - 'unauthorzed tagographies' if you wish.

Thoughts?

Nitin Borwankar

-- 
Nitin Borwankar

http://greener.com
Find, Learn, Act .... Greener, the search engine for the planet

http://tagschema.com
Implementation of tag database applications

nitin at borwankar.com
510-872-7066




More information about the Wg mailing list