[uf-new] The Process (was: hAudio case study)

[Manu Sporny]

Frances Berriman wrote:
> I imagine it will be valuable as a primer for those interested in
> creating some type of microformat (or wondering if microformats are
> indeed for them) and maybe how the process works.

Good, as that is one of the goals of the document :)

> Talking of... I note that one of your encountered problems was with
> the process itself.  Was it just the shifting/vague targets that were
> the main problem, or do you have others? 

Shifting/vague targets was the most frustrating part of the process. The
idea that Microformat authors now are being held to a different standard
than the ones that created hCard, hCalendar and hReview.

For example, once we had collected around 30 examples for hAudio, we
thought we had enough. After all, review-examples (hReview) only has 17
examples collected. However, some on the list kept going on about
collecting more examples before a decision could be made. After we had
50 examples, the same issue was brought up that we needed more examples.
Thus we went overboard and collected over 100 examples.

To us it seemed like this was an unfair argument that was thrown out
there whenever we disagreed with somebody that was "more senior" in the
community. The argument came across as "Well, you're just not analyzing
the correct sites - otherwise, you'd see my point... so why don't you go
and analyze more examples until you can prove us right". The person
making the previous statement doesn't have to do anything to defend
their viewpoint and the person creating the Microformat now has to spend
tens of hours collecting more examples.

In the end, it made hAudio better - but at the expense of frustrating
the Microformat authors.

The other issue seemed to deal with being surprised by things that you
couldn't do with Microformats, as the wiki doesn't necessarily focus on
what you can't do with a Microformat. For example: overlapping
Microformats is a problem, so is the lack of namespaces, global
identification on pages, and the scoping issue outlined in the hAudio
case study.

We felt that the community wasn't very upfront about these shortcomings
of Microformats. We didn't even know that the community understood that
Microformats had these shortcomings until we were 7 months into the
process. I think we as a community should be very honest about what one
can't do with Microformats.

In an attempt to address these issues, I've authored the following
document (with help from Martin McEvoy and Brian Suda):

http://microformats.org/wiki/how-to-start-a-new-microformat

The document attempts to outline The Process more clearly and in a
step-by-step manner. We found ourselves missing steps or not
understanding what needed to be accomplished at certain points in the
process.

Our general view of The Process when going through it the first time was
that it seemed to be being made up as we went along. There are several
community members that have their own take on each step of the process
and their advice sometimes conflicted with what other members of the
community advised us to do. In short - there didn't seem to be a
consensus on the process, which led to frustration when attempting to
get to the next step.

I think the best thing that the community could do at this point
regarding the creation of new Microformats is to smooth and refine The
Process. It is not very easy to grok until you've been through it... and
after you go through the New Microformat meat grinder, you really don't
want to do it again. :)

-- manu