[uf-discuss] Microformat Encoding Thoughts

Thu Nov 3 19:48:15 PST 2005

These thoughts are specific to the citation microformat, but not
exclusive to it.

The citation microformat properties will be serving 'double duty' both
to describe the bibliography and the document itself. For example:
...
<body>
<div class="foo">...</div>
<div class="bar">...</div>
...
<ol class=""bibliography>
<li class="foo">...</li>
</ol>
</body>
...
So if the 'foo' property describes something like an author, then the
first 'foo' property is in reference to the document itself, and the
second 'foo' is in reference to that citation. The 'bar' property is
in reference only to the document itself. So if this were a blog post,
the 'foo' and 'bar' properties outside the bibliography refer to
things you wrote, where the 'foo' inside the 'bibliography' is in
reference to something else external.

My question for this is: Does this make sense, is it intuitive? is
there a reason this is a bad idea. (my one thought is what happens if
a page contains multiple documents you want to mark-up - the answer to
that would be a root node of something like class="document" wrapped
around the data, but it would have to be a unique name to avoid
collisions with existing classes. Also if a page contains multiple
documents, what would the expected output in a transformation to be? a
list of bibliography citation from across several documents?)

The second idea is something that came-up in hCard/hCal but was never
implemented. This might be a good place to explore this idea further.
It deals with the CSS 2.1 class selectors (not implemented in all
browsers). In a CSS class, you can use the hyphen to separate terms
and use a CSS selector to extract portions of that information. For
example:

E[lang|="en"]
Matches any E element whose "lang" attribute has a hyphen-separated
list of values beginning (from the left) with "en".

So that would find 'en-us' and 'en-gb' and 'en-foo-bar'.

The advantage to this form would be to allow for a more flexible
second-hyphen-term. The idea was kicked around in hCard to mark-up the
TEL attribute. So this would be valid class="tel-work" or
class="tel-home" or unknown x-params class="tel-voip" or
class="tel-skype".

When dealing with citations there is a plethora of possibilities for
types of authorship. To consolidate them we could look to a format
such as:
class="author" (the primary author), class="author-producer" (in
citations that support a producer property, it is encoded as such, in
more generic citation forms, this will be considered as just an
author) - i'm sure there will be a citation expert to correct me in
saying a producer is an author, but you get the general picture (i
hope). Or even class="author-secondary", class="title" and
class="title-alternate" and <span class="title-translated"
xml:lang="fr">...</span>

I'm sure there was good reason we dropped support for the CSS selector
route, but this microformat might be a good place to re-consider it.

Any thoughts from the community?

-brian

--
brian suda
http://suda.co.uk