[uf-discuss] Some (newbie) questions on microformats

[stephen mulcahy]

Hi,

I'm coming across an increasing number of references to microformats and
already do some work in the area of the Semantic Web so I'm trying to
see the big picture of how it all fits together.

I've read a few introductory articles and faqs but still have some
questions about how microformats fit together and work. When I first
came across microformats I had the (mistaken) impression that they were
a way for anyone to arbitrarily mark-up their data. After some more
reading and a little contempltation I come to the (obvious)
understanding that that couldn't really work - I mean, what good is a
formatting your data in a format that only you can understand (I guess
it might make some sense for large organisations but apart from that ...)?

So my current understanding of microformats is that they are a new
approach to adding meaning to the web by (lightly) tagging existing
content (in XHTML) to add a semantic dimension to documents. The barrier
to entry is pretty low because in a little of cases you can tag your
existing content by simply adding some class attributes to your
document, right? This contrasts with the semantic web where you need to
take your data in its existing formats and create RDF from it.

In terms of standards - is
http://microformats.org/wiki/Main_Page#Specifications a definitive list
of the microformats in use? If someone wants to introduce a new one is
the approach described in -
http://microformats.org/wiki/faq#Q:_When_should_I_use_a_microformat.3F_What_are_they_for.3F.22
the best way to approach things?

I have a problem in mind but I'm not clear on whether microformats is
the answer or not - but I guess its what initially prompted the search
that brought me to microformats and it fits the notion in
http://microformats.org/wiki/process of their having to be a "problem"
to be solved. The problem in this case seems to me, to lend itself to a
microformat. As a sysadmin for a small IT company, I regularly have to
buy new equipment. Let's take laptops as a example.

Typically, I come up with a set of criteria for the laptop I want e.g.
(the following are some example criteria, not to be taken too seriously)

1. Must be a HP or an IBM
2. Must have a minimum of 512MB of RAM
3. Must have a Core Duo 2 processor.
4. Must have a minimum of 40GB of hard-drive
5. Must have a wifi card that supports 802.11a, 802.11b and 802.11g.
6. Must have a dedicated (not embedded graphics card).
7. Does it support Linux?

So I take this info and I surf to the various manufacturers websites and
  try to narrow their list of models down to the list that meets my
criteria (maybe ordered by price).

I guess in most cases this info is already on manufacturers websites,
but its certainly not amenable to scraping and parsing semantically (and
maybe its not in the interests of the manfacturers to provide the
information in a format that lets me easily compare them to other
manufacturers) but it strikes me that if they did .. it would be really,
really easy for me to go to all the major manufacturers websites, suck
them their microformatted data and then analyse it off line - I see
something like an openoffice datapilot table (microsoft excel pivot
table) where I can click various filters to match my criteria above and
sort the output according to something like price and voila, my choices
are obvious - is there a microformat that lends itself to this sort of
thing already.

Is this the kind of scenario that microformats could meet or am I way
off of the mark? I guess even if manufacturers didn't want to
participate in this, there are lots of sites out there that review
laptops - if you could get them to sign-up to this the microformatted
information would become available quickly (all it would take is one
person to review a laptop).

Thoughts and comments welcome ..

-stephen

-- 
stephen at skynet.ie                             http://www.skynet.ie/~stephen/