When I started using QL back in 1975 – a statement which dates me I realize – I recall being struck by the crudity of Boolean searching and its scant contact with the concept and context driven research behaviour of experienced searchers. Search questions had to be bent out of shape to fit into Boolean: paradoxically legal jargon and Latin was easier to search. The entire process put a premium on knowing the vocabulary and logical structure of the law.
Then we learned how to game Boolean. With nested searches and lengthy synonyms and then proximity locators. We overcame the limitations of George Boole. Easier to do on Lexis or Westlaw. Those of us who used Dialog knew all about combining saved sets in order to locate our needed information within the overlaps on the Venn diagram.
In the early 1990s both West and Lexis (as they were then called) introduced Natural Language searching. It wasn’t really – more like preset synonyms and broad proximities – but it did improve recall and accuracy and overcame The Curse of Thamus. See http://www.yale.edu/lawweb/lawcrs/arc9798/lasdab1.htm
Late in that decade it looked as if smart engines would employ feedback loops, asking users to tick preferred initial results, getting us “more like this” – a cumbersome but interesting development in Dow Jones. That feature seems to be present in the current Lexis search engine, although the results I’ve looked at recently don’t seem to advance recall or accuracy. Dolphinsearch offered us yet another way of getting closer to concept searching, but deployed on legal materials, the results weren’t up to the hype. Lexis seemed to be toying with the ideas though – see http://www.infotoday.com/newsbreaks/nb011210-2.htm
Then web searching took more and more of our attention. Despite the crudeness of the Google engine, today’s students and younger lawyers have been raised to look there first. Don’t worry that it only indexes the first 101 KB of a page or that it will miss much more than the first ten pages of a PDF, the sheer quantity of material makes Google an unparallelled resource. See http://searchenginewatch.com/reports/article.php/2156481 It’s interesting to read now the original concept paper for Google , entitled The Anatomy of a Large-Scale Hypertextual Web Search Engine, at http://www7.scu.edu.au/1921/com1921.htm
I was momentarily optimistic when I saw that PC World had a preview of the variety of tools from this article in the June issue of PC World – New Search Stars: What’s the trend in Web search? http://www.pcworld.com/news/article/0,aid,120962,00.asp That is until I actually read the piece and discovered that the piece profiles a variety of content portals, still deploying the same traditional tools.
I see that folks are now revisiting some of these ideas
My question to Canadian researchers is what sorts of strategies are we in fact deploying to get round the crudity of the search tools. I tend to use Google Advanced Search, and then limit results to domains (like .gov or .gc.ca) or by type, like .pdf files if I think I’m going to get deluged. I’ll run supplementary searches by right mouse clicking within Mozilla to drill down on something that looks promising – another reason to like Mozilla – but it’s a crude sort of feedback.
What sort of innovations should we be asking the legal database vendors or the Googles of this world to develop?