Microformats for Thinglinks?
I had an interesting discussion about Thinglinks with Joi Ito. He suggested using microformats, a set of open data format standards, for structuring microcontent about things.
Microformats are essentially about describing the relationships between resources on the web and adding structure to web documents. For example, in a blog post one can use a simple XFN type 'rel' attribute on a "a href" tag to indicate some additional metadata on an object. Something like this:
How would microformats work with thinglinks? Some things that come to mind:
1. Microformat link types help to understand the various functions of a thinglink.
2. Could the hCard be used for describing a Thing?
hCard properties:
fn, n (family-name, given-name, additional-name, honorific-prefix, honorific-suffix, nickname), sort-string
url, email (type), tel (type)
adr (post-office-box, extended-address, street-address, locality, region, postal-code, country-name, type), label
geo (latitude, longitude), tz
photo, logo, sound, bday
title, role, org (organization-name, organization-unit)
category, note
class, key, mailer, uid, rev
Really quick example of an hCard for things:
- name of a thing
- maker (desinger, producer, etc.)
- address (where it can be found)
- geo (latitude, longitude)
- photo, logo, sound, date of manufacture
- physical properties (size, material, etc.)
- thinglink
- category, note
- class, key words, rev
3. Microformats could be used for generating HTML documents from a thinglink database.
It would also be practical if one could just fill in some basic information about a thing and have ready-made microformat tags.
4. Could thinglinks be used as as unique ids within microformat tags?
---
Some more questions:
- How would you actually put an hCard into a blog post about a thing?
- Who are the principal users of microformats today, and where can I find examples of bloggers using microformats?
- How could craft bloggers benefit from using microformats?
Ulla,
Certainly, this is one potential use of thinglinks. There are loads of potential uses though. I think the important thing at this stage though is to clearly define exactly what a thinglink is. This means determining the format, deciding exactly what it refers to (for example, whether it refers to an item, like a serial number or a class of items like a UPC or ISBN), and deciding how thinglinks are registered.
Every extra piece of information you add makes things more complicated. For example, the geographical location of an item changes quite a lot, especially if it is being traded. Another issue is privacy. The owner of an item does not necessarily want to be known to anyone beyond the person they bought it off.
So I say, keep the schema as simple as possible at the outset, and let application developers build meaning and complexity around it.
A connection that may be useful - David Weinberger has a chapter about unique identifiers in his forthcoming book - http://www.hyperorg.com/blogger/mtarchive/the_synchonicity_of_intertwing.html
Posted by: Antoin O Lachtnain | December 26, 2005 at 07:20 AM