Microformats

[Repost, so the subject is correct!]

The Privacy Policy
  http://groupserver.org/about/policies/privacy
is my first use of microformats
  http://microformats.org/

There are two benefits with microformats. First, it provides a consistent set
of CSS classes, which we can use to mark up information about people,
organisations and events. The data about me, at the bottom of the privacy
policy, uses the hCard set of microformats. Second, the data can be
automatically extracted. For example, the Operator extension for Firefox
  http://www.kaply.com/weblog/operator/
can turn the data into a vCard, and it can automatically link to the Google Map
for OnlineGroups.Net HQ. There is also a microformats bookmarket that supports
Safari, Firefox, IE 6 and IE 7
  http://leftlogic.com/lounge/articles/microformats_bookmarklet
and plans for supporting microformats in Firefox 3
  http://wiki.mozilla.org/User:Mkaply:Fx-Docs:Microformats

When I refactor the profile pages, I will probably use microformats to provide
semantic information about the members.

I have been looking more closely on how we embed semantic information in the
pages produced by GroupServer, after being prompted by two recent articles on
“A List Apart” about RDFa (see References below). In this post I will summarise
what GroupServer currently does with microformats, outline my proposal for how
to extend our work, and briefly discuss what I do not like about RDFa.

Currently we use the hCard microformat to embed semantic information about
people into our pages. For example, the name at the top of a post on the web is
within an hCard:
  <span class="author vcard">
    <strong>From:</strong>
    <a class="fn url"
      title="Profile of Michael JasonSmith"
      href="/p/mpj17">Michael JasonSmith</a>
  </span>
The semantic information is indicated by the value of the class attributes of
the elements. For example, *vcard* says that the element contains information
about a person. The *fn* says that the contents is a full-name, while the *url*
indicates that the link is the homepage for that person. The profile page
itself is also a vcard, but longer. This information can be used by a search
engine, like Yahoo!, or by the browser itself, such as with the Operator add-on
for Firefox. At the very least, it keeps the profile properties and class
attributes consistently named.

The hAtom microformat (see References below) is designed to allow semantic
information about blog posts to be added to a page. Our posts on the web are
very similar to blog posts: they have an author, content, a published date, and
a permalink. Indeed, we already produce ATOM feeds from out posts. By adding
hAtom microformat code to pages produced by GroupServer we can make them easier
to index for search engines. We also get the bonus of being more consistent
with our CSS class-names, which has been really useful to me in the past. The
change should be painless as we do not use the class-names in question, so
nothing is likely to break and testing can be kept to a minimum. It is also
quite easy, as posts are produced by one template.

An alternative to microformats is to use some form of RDF (Resource Description
Framework). RDFa uses XML-attributes (hence the “a”) to indicate the semantic
meaning of data in the page — much like microformats do with the value of HTML
class-attributes. While far more extensible and generalised than microformats,
the main flaw that I see with RDFa is the pages are not XHTML, they are
XHTML+RDFa. This may throw some browsers, and I cannot afford the time to test.
In addition, the markup itself is more complex than the class attribute used by
microformats.

GroupServer will get the benefit of semantic markup more cheaply using
micorformats than with RDFa. This does not prevent RDFa from being used in the
future, as both can be added to the same page as they do not conflict. However,
I suggest we leave RDFa until we need to support semantic information that is
too complex for microformats to handle.

*References*

  Introduction to RDFa
  <http://www.alistapart.com/articles/introduction-to-rdfa/>

  Introduction to RDFa II
  <http://www.alistapart.com/articles/introduction-to-rdfa-ii/>

  RDFa Syntax
  <http://www.w3.org/TR/rdfa-syntax/>

  hCard Microformat
  <http://microformats.org/wiki/hcard>

  Leveraging the Data Web
  <http://developer.yahoo.com/searchmonkey/smguide/semantic_web.html>

  Mike’s Musings » Operator
  <http://www.kaply.com/weblog/operator/>

  hAtom 0.1
  <http://microformats.org/wiki/hatom>

  All posts: GroupServer.Org
  <http://groupserver.org/s/search.atom?t=0&p=1>

Thanks for your comments, Steve.

Google, as far as I can tell from Googling Google, does not explicitly
say what microformats Google Googles. They say they support hCard,
hReview and hProduct
<http://www.google.com/support/webmasters/bin/answer.py?hl=en&answer=146897>. I
presume they support hAtom, because Blogger uses it
<http://groups.google.com/group/bloggerDev/msg/f4f869737b18a967>. However,
Google does  not provide anything as nice as the list of supported formats on
Yahoo!
<http://developer.yahoo.com/searchmonkey/smguide/supported-microformats.html>.

As of June 18, Microsoft Bing was not using Microformats
<http://searchengineland.com/microsofts-search-engine-optimization-advice-for-bing-21152>
<http://www.bing.com/community/forums/p/647235/9537206.aspx#9537206> but
they may in the future.