12 November 2010

universal folksonomy

This is probably not entirely new. Maybe even not new at all. In fact I've probably just joined legions of people out there who are wondering why ..

There isn't yet a simple plain-text standard for meta-data, that can be applied universally across all services and objects - from web pages and web resources, to documents and files, from books and articles to audio and video files - applied basically, to all digital resources.

Then building on top of all this tagging - all this human-computation and folksonomy - building flexible tools that mine and collate the meta-data from across all possible services, communities, spaces and contexts. Making these tools expandable, so adding another service/context/data-source is as simple as possible.

So, how is this different?
  1. we use it in the same way, but consistently across all services/contexts/collections

  2. we use it even when no provision is made for it - wherever we find space for plain text that can be co-opted to meta-data - like in comment fields, sidewiki entries, bookmark notes, mp3 id3 comment fields, etc

  3. then we build software to collect them from both the formalised meta-data services and the informal - and shape flexible plugins to collect them from wherever the standard can be recognised
what sort of meta-data?
  1. First, the simplest forms of meta-data - tags/labels - flat, informal ontologies - collections of keywords - like those used within services across the so called web2.0.
    And when using services that aren't designed to use tags/labels - co-opting plain-text fields by, say, the following:
    tags = blog, tagging, ideas, software, socialSoftware

  2. Second, simple key value pairs, like author=JaysenNaidoo; date=20101113; license=gpl3

  3. Keeping the set of standards as open as possible - but defining them properly in public. Adding any other std that can be represented in plain-text and gains enough ground - depicting hierarchies, semantics, etc.

.. possibly more on this later.

some almost related links:
Actually some of these suggest practical tagging standards as well:

2 comments :

jaysen said...

when using other plaintext fields to enter metadata, titling your text as "metadata" might be a good idea.

jaysen said...

metadata:

author=JaysenNaidoo; date=20101113; license=gpl3

tags= folksonomy, humanComputation, ideas, metadata, socialsoftware, software, tagging