Draft to Specification (was: [uf-discuss] More than three years)

[Toby A Inkster]

Yay! Finally the sort of discussion I was hoping to kick-start.

I think 'adr' and 'geo' should certainly be considered candidates for  
promoting from drafts to specs - these have been stable for ages and  
don't seem to have any issues raised (at least none which don't  
effect hCard, or microformats in general). There have been plenty of  
ideas proposed for extensions to them, but the current drafts  
certainly address 80% of use cases. These can be deferred to a future  
iteration which would address 80% of the remaining 20% of use cases.  
(This would also address the anomaly that hCalendar, a full  
specification, recommends that event locations may be marked up with  
adr and geo, each of which are only drafts.)

And hAudio and figure are probably stable enough to go on the  
"official drafts" list.

The Microformats process is extremely helpful in the early stages of  
drafting a spec, taking the authors/editors through the process of  
researching relevant examples, looking at existing standards,  
narrowing down requirements to a simple, usable and deliverable set,  
and building a draft vocabulary; but after that, the process leaves  
you dangling: there is no defined process for knowing when to freeze  
a spec, when to start asking for implementations, creating test  
suites, publishing as a formal spec once you've got a few  
implementations that pass the tests, restarting the effort for a v1.1  
iteration, etc. [That was a very long sentence. I apologise to  
readers: my writing style tends to favour run-ons.] The process  
itself should be considered a product of the microformats community;  
but the process does not yet cover 80% of use cases, so needs more work.

-- 
Toby A Inkster
<mailto:mail at tobyinkster.co.uk>
<http://tobyinkster.co.uk>