Again, with the Westlaw Next

Pricing don't mean thing if it don't got that Algorithm:

As it was explained at this meeting in Eagan (though they may change it, I suppose!), there is no charge at all until you look at a result from the list. So the charge is not based in the Westlaw Next model on the size of the database searched, unlike the model we are used to.

What you understand is absolutely the current model -- in transactional or per minute pricing, they price by the size of the database unless your contract has "flat pricing" for certain databases -- meaning that no matter how long anybody searches on your password, the price is the same-- and even that flat price depends on the size of the database. That is, in Massachusetts Cases, for instance, I might pay $100/month (I have no idea!) for a flat fee at my firm. But for the All States cases database I might instead have to pay $1,000/month. (Again, I have no idea how realistic these numbers are, but it certainly changes radically on the size of the database, even for a flat fee).

The folks in Eagan explained that the Westlaw Next search plan is to look in ALL databases for ALL relevant documents, using an algorithm that starts with the terms the user enters. If the algorith recognizes boolean connectors, it will implement a terms & connectors search. If not, it runs a natural language search. There is a glitch where you might try to run a phrase search in Terms & Connectors, because it does not recognize the quotes, right now, and it does not deal with a single term search, which it always runs as natural language, even if you wish to run it as a Terms & Connectors search.

But the search is enhanced, no matter whether it is recognized as T&C or natural language in these ways. The search engine looks into the databases for similar or related terms in:

1) Secondary sources (treatises, law reviews, indexes, including "see also" entries to develop additional terms;

2) Key Cite, especially using the headnote feature to see which cases cite a case for what reason, to find related cases;

3) Key Number system, to look up the topic & key number for the query;

4) The web of related publications -- statutes, regulations, etc. that either cite the case found or use the term or if you are finding a statute or reg, the cases, law reviews, ALRs, encyclopedia entries that discuss; it all hangs together & this system pulls it together & brings it back.

5) User information; This last is very interesting. The system gathers information when users are printing, printing with citation, printing with reference & linking between the various documents as they do these actions. They add a metadata tag that either adds a term or links the documents to each other. So when you find one document, it pulls the other documents that previous searchers have indicated might be of interest.

I have to say that West needs to get the sample passwords out to the academic librarians so they can see what is being talked about. The folks who saw it at Legal Tech were much less than wowed. I don't know what happened at Legal Tech but it's too bad because it's a very important advance in legal research and deserved to be noticed more by those who attended. The mainstream legal press (N.Y. Times & ABA Journal) have addressed this much more as another ho-hum story about Westlaw reacting against recent changes by Google and the market hotting up with Lexis and Bloomberg. They are really not understanding what they are seeing. This was in production for five years, way before Google's Legal Search Button became public, and before Bloomberg became a major player. The academics are the future for this product, while the law firm market is certainly going to fund it, they need to get the law schools to buy in as well. Get those sample passwords out, please!

(post from an e-mail to the ALL-SIS listserve)