Monday, February 19, 2007

Why should we believe that using metadata will help J. Random User get in touch with her Buddha nature?

"Metacrap: Putting the torch to seven straw-men of the meta-utopia" is a short read on why metadata may not work in the open Internet: http://www.well.com/~doctorow/metacrap.htm

For a foretaste, check out the brief chapter headings:
* 0. ToC
o 0.1 Version History
* 1. Introduction
* 2. The problems
o 2.1 People lie
o 2.2 People are lazy
o 2.3 People are stupid
o 2.4 Mission: Impossible -- know thyself
o 2.5 Schemas aren't neutral
o 2.6 Metrics influence results
o 2.7 There's more than one way to describe something
* 3. Reliable metadata

There is nothing like the open Internet to sound out ideas. My own opinion goes more towards that metadata is a tool because information is locked away from powerful search engines which necessitates organizing it with tools such as metadata, and analysis from data experts. But this model may not really help in terms of the absolutely creative nature of human endeavor. Methinks it comes more from the realm of library science and pedagogy where finding something in a physical realm and instruction is the goal, which has not caught up to the power to SEARCH, and view others' metadata such as relationships (tagging, authority - author, etc).

Still in a nation focused on UFOs and celebrities, maybe this is what is required to indulge the great unwashed jungle, and security protects the scientific and verifiable.

I would feel much better if I knew that all drug and health-related companies shared their information on research and discoveries for example.

Maybe someday instead of this Age of the Information Cold War, we will have strategies for open information presentation and a way to nix the harassment from the identity thieves.

My personal model included metrics to influence the model on purpose - but that is "mixing metadata with data" - strictly the wrong direction according to current science.

Combining the link stats (by data mining) should provide more information. It's how Page Rank score/ hit rate combine with IP address and other long term use tracking - thus relevance is displayed that makes a difference. People are lazy and many really don't get tech - so I am thinking of a way to assemble the page to include the Page Rank score/ hit rate, to help in usability.

What I want to my frequently used websites (such as the MSIM iSchool) is for it to present ALL the data specific to me. In other words I pretty much use the same exact pages over and over again - I don't want to have to search or click or anything - I just want what I use to be presented first and foremost. The other data that the 'outfit' (the UW in this case, or a business or religious or other org) wants to present to me or makes available is also interesting but less so.

Including metrics with the metadata allows information scientists to tweak their presentation model by showing it's relevance - in effect putting in metrics to view it's context to the greater whole. But not as metadata - as other information mapped to such things as sniffing my IP address.

What I mean by Medieval Information Age is that we have lots and lots of closed information and closed data systems - I hope that these will open up but I know I will not see this happen in my life time or even in the next lifetime, Greenstone not withstanding.

Meta data is like the "thin client" of the Information World - it allows lots and lots of information to be relayed very quickly, in categories.

Will we look back on this as the Medieval Information Age? I view information as a contextual circle - I know nothing.

No comments: