Mark and Miles gave me plenty of ribbing about the name change from CaseShare to Catalyst. They said it was hard to find us and likened it to the obscurity that Autonomy enjoys/suffers from with their name choice.

Pictures may follow if my new acquaintance follows through on an email. The audience was divided into three columns and two sections: 20x14, 15x14, 20x14 in section 1; 8x14, 6x10, and 8x14 in section 2. So, my math says over 1000 people. Michael T, EVP Global Sales FAST, kicked it off and used some familiar sayings like "meat on the bone for the concepts you are developing" and we should "goal ourselves with joint innovation".

^{From The Wow Agency}

Actually, that's not true. Orange Man Group kicked it off and they were a replica of Blue Man Group down to the opening pipe drumming. Perhaps comedic, perhaps a harbinger of mimicry.

Then, a heady section of the future of search in soundbites from experts. The most interesting from Batelle was "search is a shared language between your business and your customers." Another (anonymous to me) was "search is both an opportunity and threat."

Batelle took the stage and reviewed the central theme of last year: "Search is the new interface." His reasoning is that hunt and poke of the DOS prompt gave way to the GUI which gave way to Google. His new theme is that search is a conversation with your customers. What are you trying to say? What do you hear?

Next up was Andrew McAffe from Harvard Business School. His theme was is Enterprise 2.0 a strategic advantage? McAffee is a professor and wanted to examine E2.0 in the context of academic pursuits. He used the VRIN model.

Valuable
Rare
Inimitable
Non Substitutable

Are these tools of wikis, blogs, folksonomies, search, and prediction markets (E2.0) capable of meeting the VRIN model?

Valuable: fosters collaboration, provides judgement (Surecki's Wisdom of Crowds poster child was thrown out - The Iowa Futures Market for politics), enables emergence - or self organization, and builds real knowledge management through the passive collection of tacit knowledge.

Rare: you can't easily buy your way into this, just building it doesn't ensure success, and these require sustained management attention and leadership. (Honestly, I was lost on his point here).

Inimitable: the software is easy to buy or acquire, "successful exploitation is fiendishly difficult" (Spark rollout anyone?). Best practices diffuse at glacial pace.

Non Substitutable: Non IT based approaches don't deliver. Alternative IT approaches like groupware have been unsuccessful.

References a 9X problem for new products (research needed). Suggests email may be to embedded. References TIVOs superiority but inability to displace the VCR.

^{From The Wow Agency}

Next was the panel on Search plus Business Intelligence. In this panel discussion, three panelists were lead through a fairly painless, wrote conversation about using search in the context of data marts (FAST's next product). One was Bill Inmon (widely considered the father of data warehousing), one was from Accenture, one was from FAST. Nothing much learned here accept FAST has some new RDBMS tools that understand ER and will draw the schema and let you participate in a QBE before ingesting into FAST (probably through JDBC).

Next came an analyst from the Economist Intelligence Unit - a company tied to the publication. She reviewed a comprehensive survey they did about E2.0 tools with C-level folks. These weren't the average companies, each was about 2.5 B in revenue. She found that a lot of them understood E2.0, but only a lite version, not one that focused on the network effect (more users == greater value). See presentation here.

Then, Tim O'Reilly took the stage. He covered his definition of Web 2.0 specifically pointing out a key tenant borrowed from Clayton Christensen: The Law of Convservation of Attractive Profits. CC's theory is very similar to matter theory in thermodynamics. If the top layer of the value chain becomes commoditized, profits shift to the next, unknown layer of the chain. So, after IBM's success in standardizing the PC market, profits shifted to software companies. The next point is certainly arguable, but as the software (OS, DB) became commoditized, profits shifted to the information brokers - Google. Now, profits may again be shifting to the next layer. This interesting thread bobbled it's way into a segue discussion with the CEO of Reeds Online and another FAST big shot. Reeds is using FAST in a fairly unremarkable way and this was bantered about between the three of them.

Next, the most interesting part of the morning session happened. Michael Schrage (MIT Media Labs) hosted a difficult, but engaging conversation with Matt Brown from Forrester and Bjorn Olstad from FAST. Michael challenged the benefits of search and W2.0. He was full of inflamatory and pithy quotes like "your interface looks like a pig's breakfast" or "search is a link ghetto" or "you expect to craft all your needs around my narcissitic intentions on your blank page" or finally, suggesting that "folksonomies are merely changing the angle of attack" providing "an illusion of choice" but still subjecting the end user to filtering, not just by experts. Outside of watching the speakers clearly be thrown off guard, it was clear that FAST is inserting new language into the culture of "retrieval centric" vs "transaction centric" architectures.

Lunch provided by Accenture: Pork empanada, ceasar salad with grilled beef, chorizo quessadilla, minnestrone, and some iced tea. I sat at the table with a guy from Google, a guy from IBM, and the new VP of sales at Grokker. I'm not sure we should do business with the latter since he has a very low opinion of his customers (he thought I was from FAST).

I went to the customer Video Taping. I met with Jay Famico. He did the search implementation for the corporate web site. He's going to provide contacts to his team for live search. Also met Chris Riopel from Inxight and he seemed like a nicer guy than Grokker.

Then scheduling confusion ensued. The tracks got out of sync, and I became a conference info snacker - catching parts of Forresters research on W2.0, an uninspired presentation from ING about replacing SQL as the front end of the data mart, and then an introduction to PSP with the architect from Utah. I then listened to Steve Dimarco who does implementations for the government. He's bright and suggested lots of best practices. Afterwards we discussed the file system as the meta data storage and the void left by Web DAV. I'll follow up with him in email.

Met Jeanne (AM from FAST) and Thomas Molbach (Dan and Daryl's boss). We discussed more challenges in depth and he had some suggestions I'll share with my team. We were interrupted by the Zantas product manager, but I stayed quiet. Neither Jeanne nor Thomas would say much after he left despite some prying.

Things learned: Archivas (FAST customer that replaced RDBMS in backup solution with Search) was acquired by Hitachi Data Systems. ADI is not joining the panel.

Chris Anderson from Wired kicked off Day three with an examination of the Long Tail. The Long Tail evolves because the web has new abilities to avoid the constraints of scarcity and can eliminate some inefficiencies. In addition there are economic incentives for promoting the long tail since new media costs more closer to it's official release date, so WalMart can afford two weeks of losses and make up for it on cross-sales, but NetFlix and Blockbuster don't have a lot of cross selling options. As Chris puts it, far too often we "are seduced by marketing and buzz to herd towards new releases." Search helps the long tail by finding massive variety and making it accessible.

Chris Weitz from Bearingpoint took the stage next and talked about "The Very Long Tails of Information." Chris focused on the enterprise. He mentioned on more than one occassion, "in the enterprise, information does not want to be free, it wants to hide." Chris sees Search replacing middleware as "the new default command line" and full of buzz and promise in the industry. Chris calls Vista "The Search OS."

Next began the technical tracks including a presentation from FAST spin off, Comperio. In the section called "Mining for Gold," Comperio showed how they harvest and correlate information found on the web.

I spoke next.

Lunch was a boxed sandwich with potato salad, apple, rice krispie treat, and potato chips.

After, I went to the Multimedia Miner presentation from FAST. This has improved and seems to be an obvious choice for us. Then, several oddly structured technical sessions including a tools section for ESP Deploy, SFERanquilizer, Acuity, and Clarity. ESP Deploy is an XML config ontop of the FAST installer for convenience. SFERanquilizer is a drop in on the admin node to show why documents got ranked a certain way. Acuity is a fucntional test tool and may replace the PQE. Clarity is in use already.

Finally, I attended the session on Lingustic Studio and Fact Extraction. Linguistic Stuido is a 32-bit application that puts all language and syntax features in one console. Fact Extraction is a revamped matcher that does mark up in documents. Both look very interesting.

Courtesy of taka_tazawa

Courtesy of taka_tazawa

• http://www.torreypine.org/
• http://www.sail-technology.com/
• Conservation of Attractive Profits
• Permabit
• 24.com (South African Portal - Charles Talbot, Jonathan Muir)
• 9x Problem
• FF07 blog
• Long Tail Blog