This is a fast paced panel discussion.  Just going to jot down some notes.  If you all have questions about the post's content, comment and I will fill in the detail later.

 

One distinction: del.icio.us vs wikipedia = indiv space vs shared space

Del.icio.us tags= a document description & associated behaviors or additional context around it

 

del.icio.us- why tag, what to tag, how to tag? Key idea- people tag for different reasons, some to help themselves find stuff again, sometimes to help other people find their stuff, sometimes to help organize groups, behaviours, or context around a given concept. 

tags on flickr to create groups of photos

 

Audience question: can we create meta- meta tags to share tag info acros across different services?  Response: the purpose of the tags is not always the same and therefore it is difficult to normalize tags across different services. 

 

how to give users feedback on how to effectively tag?  what feedback loops can be provided to boost learning?  Flickr- no such thing as a bad tag, Wikipedia- community polices innappropriate tags, Del- shows your tags, top/popular tags, intersection of tags w/others. Del.icio.us- don't interfere too much by imposing restrictions as users won't be able to find the stuff they tag.

 

Shirky- user/time are new axes for classification and this creates a need for tags and taggers to understand context. 

 

Del- expert-generated taxonomies don't pay attention to what people are trying to do.  Del tags are for people to help them find stuff, not to necessarily describe things in the abstracts

Jimmy Wales, the founder of wikipedia is up on stage talking about how wikipedia and how it fulfills the original promise of the Net

Problems w/Wikipedia's model

• Quality control

• Author fatigue

Solution to these problems: wikicities.com.  Solves author fatigue problem by creating a community that is topic-focused buyt not dependent on any one author.  Solves quality control thru the wiki peer-review process.

Organization by the Community: free form wiki software lets the community decide how to enforce community mores. 

Conclusion- wikis are a major "social innovation" according to Wales

I was looking around at the ETech crowd during the break.  I would guess that @ least 75% of the attendees are white males.   I wonder if this mirrors O'Reilly's print readership demographics. 

George Dyson, who is an excellent speaker/writer on the history of technology and science innovation, is talking about how Von Neumann created a great environment for research at the Institute for Advanced Studies in Princeton. 

He is recounting the problems that Dr. Kurt Goedel faced trying to emigrate back into the USA during the early 1940's.  Goedel was one of the greatest mathematicians of the 20th century but he was also a very paranoid fellow, apparently. 

In any event, Goedel's work led to Turing which led to Von Neumann's work on computing.  Von Neuman is credited as one of the creators of modern computing. 

George has clearly spent a lot of time digging thru various archives to piece together this fascinating bit of historical brain candy.  One amusing little tidbit is the fact that the Institute's kitchen staff complained that the engineers building the first computers at IAS consumed too much tea and sugar.  Early example of the alpha-geek diet of caffeine and sucrose!

Increasing customization/personalization is a key focus for Peter Norvig, head of Google Labs.  Presumably, this strategy is designed to create lock-in for Google users (after all, switching between search engines is pretty easy and that low switching cost is, imho, Google's achilles heel.  It is also a weakness I think Y! and, more importantly, MSFT are going to target as they try to arrest Google's growth.

To read a really insightful article about Google's future and the presumed importance of creating user lock-in, see Charles Ferguson's excellent piece in Technology Review

Flake is showing us all the new widgets that Y! has in pre-beta:

• better movie recommendations

• Y! search development network

• Y!Q- provides related search on the fly.  sort of similar to what we just saw from Bezos/A9 but seems less flexible.  Offers contextual clustgered search

• Y! ticker beta

• Y! desktop search

• Y! FareChase beta

An overview of Yahoo Research Labs:  30+ scientists/engineers working on R & D.  Focus on machine learning, collective intelligence, scientific computing, and text mining.  Try to build a hedged portfolio of activities focused on short, medium, and long-term projects.  Can see some of their research oline- only show what is not likely to become a product.

Tech Buzz Game: looking at how to use markets as predictors for future events (ie HSX, Terorism futures market, etc).  Y! and O'Reilly jointly introducing the Tech Buzz Game  TBG is an artificial market where users can predict future tech trends.  Relative value is determined by volume of search results.  Players need to time their market predictions to earn TBG play $$$ and prizes. 

Looks sort of cool.  Certainly could offer some interesting insight to what users seem most interested in re: rolling out future products/services.  I wonder how easy it will be to game this game…. 

Also showing dynamic pari-mutuel auctions.  Did not catch what they were but Flake assured us this Y! innovation will change auctions and gamblings.

I recently arrived @ O'Reilly's ETech conference.  I missed the web services mash-up earlier today but caught the end of the kilted Ben Hammersley's presentation on Atom, RSS, etc.  Very informative.  The session concluded w/a not very practical but quite interesting discussion about the philosophical problems of ontologies.  Ben himself seems like an interesting fellow- he wants to do solo 600 mile ski trip to the North Pole.  Reminds me of my own Antarctic climbing ambitions. 

Also, I finally met Susan Mernit in person.  She and I had been swapping emails and calls for a while but had not f2f.  Nice to meet you Susan.