It must be the week for drive-by postings from me.
Just this morning, I tripped across an excellent write-up from ML that seems to be freely available that goes into exhaustive detail on the opportunity, how they see it evolving, who's affected, etc.
Although relatively recent (Oct 22), I could already spot a few areas where things had changed -- more of a testament to the fast-moving nature of this marketplace.
Although I can't really endorse any output from any analyst organization, if you're really interested in the topic, you might want to go here.

Chuck,
Could you add a link to the article?
Thanks,
Bill
Posted by: BIll | November 30, 2007 at 12:44 PM
Sorry, Bill, that the link isn't working for you.
The entire URL I am using is: http://cfcr.ml.com/GetDoc.aspx?e=50O3tGj8KWAObsJX5dDLPqFDXh203EpLzSE9jJo91g4nPwdDrXPbOGPKBhnyXTn9d1VdNOBGy7vXnmBgsYIT1Q%3d%3d&ctbDocIDs=10662725&v=1&m=HGI2%2b%2fw3tp5fuAMPKxFwWaldRDg%3d
I've tried it from a few machines, and it seems to work ok -- you may need to wait a bit for the PDF to download. Let me know if you still have problems, please.
Thanks
Posted by: Chuck Hollis | December 01, 2007 at 03:12 AM
The meta-metadata will exist, but we probably will not have to use it.
We, as humans, just can't make practical use of even metadata in data-sets as large as what is in use today and certainly in the near future.
AI is the only answer. Only an AI interface, who learns who we are and the context we want the data for, can interpret the metadata and meta-metadata to give meaningful results for human use.
Posted by: Jim Hatfield | December 03, 2007 at 12:48 PM
Yes, maybe, but --
AI is a wrapper term for wide range of technologies. I would offer that we probably have to be more specific as to the technologies, and -- more importantly -- how they will be productively used.
Given my age, I've been through the AI barrel more than once. And, let me tell you, so far the promise hasn't lived up to reality.
Maybe this is different.
Posted by: Chuck Hollis | December 05, 2007 at 02:04 PM