More grouping discussion, was: Re: [uf-new] First draft of hAudio proposal

Manu Sporny msporny at digitalbazaar.com
Sun May 6 19:13:03 PDT 2007


Rudy Desjardins wrote:
> It seems like some people are still missing/ignoring a few key points
> with regards to the problem(s) which some kind of a grouping or
> collection pattern are intended to address:

I have to agree with Rudy - some of you are missing the point of the
grouping discussion. Perhaps it should be called the "relationships"
discussion. The problem statement and examples clearly outline why just
coming up with a solution for "lists" is not sufficient:

"It is useful to understand the relationship between objects on a
website. A blogger may want to describe several different objects on a
web page and group them explicitly. It is important that the structure
of the page not affect this grouping as network relationships are often
not hierarchical (HTML is always hierarchical)." [1]

If your answer to the problem is "just use <ol> and <ul>", then you are
only understanding half of the problem.

> 1) There exists a need to express a grouping relationship for data which
> is sparse. Please see the grouping-examples
> <http://microformats.org/wiki/grouping-examples> and
> grouping-brainstorming
> <http://microformats.org/wiki/grouping-brainstorming> wiki pages for
> descriptions of these types of data sets.

Again, Rudy is correct - the "ul/ol" solution does not solve the problem
of sparse grouping.

> 2) There exists a need to express an N-N grouping relationship, as
> opposed to simply 1-N which is all a heirarchical structure implicitly
> allows.

The "ul/ol" solution doesn't solve N-N relationships either.

> of the problem space we're trying to solve with the grouping 'pattern',
> although additional real-world examples are still being compiled and
> analyzed to support this fact.

I find the fact that we are having to gather examples to prove that
sparse sets and N-N relationships exist on the Internet a bit
ridiculous. The Microformats process seems a tad bureaucratic at this
point. The notion of sparse  sets is fundamental to human organization
and knowledge. Take this wikipedia article, for example:

http://en.wikipedia.org/wiki/Sculpture

There are several sparse groups listed on the page... many N-N
relationships. Here are the sparse sets:

* Indian sculptures
* Chinese sculptures
* European sculptures
* pre-1700s sculptures
* full-figure sculptures

The N-N relationships:

* full-figure sculptures + Indian sculptures
* pre-1700s sculptures + Chinese sculptures

Almost the entirety of Wikipedia is a lesson in sparse sets and N-N
relationships. So, do we count this as 1 example... or do we count
Wikipedia as 1,272,193 examples (give or take 400,000)?

Remember, we're not secretly marking up any of this information. It's
all there on the page and humans can understand it quite clearly. The
only entity having trouble recognizing the grouping of the concepts are
the computers. Microformats grouping is supposed to help solve ALL of
these problems, not just provide a half-baked solution.

Is the real issue that people don't think that 48 music examples plus
the hundreds of thousands of Wikipedia entries aren't enough proof of
sparse sets and N-N relationships?

Or is the real issue that people don't like the current grouping problem
solutions?

Those aren't rhetorical questions - just want to know where to
concentrate our efforts.

-- manu

[1] http://microformats.org/wiki/grouping-examples#The_Problem



More information about the microformats-new mailing list