« Periodic Table on the Web | Main | New York, New York - so good they ... wrote a website about it »
February 24, 2003
Trends in technology : parallel universes
Lincoln Stein and the Bioinformatics Nation
Not many people know this, but I used to work for some banks. Shhh. I also used to work for a number of companies which made software for banks, mainly information and trading systems. In the good old bad old days (ca. 1985) if you wanted to get, say, foreign exchange rates from a Reuters Money Monitor screen, you needed to know exactly what bits of information were at which location on the screen (actually, in the screen record) and then "scrape" them out. Such tools were called "screen scrapers" and were horribly sensitive to changes in layout, of course. Later, companies like Reuters, Bridge (now part of Reuters), Telerate (once part of Dow Jones, then Bridge, now merged with Moneyline), Pont (now dead) and Bloomberg started offering record based feeds which made the whole thing a lot easier, but financial content aggregation standards were not there (this was before XML, remember?) so the "glue" was still written by the information system vendor companies I worked for. Today, there are very few players in the financial information system market, and financial information systems are like pocket calculators - you pretty much get a free one with every box of "Frosties". Need FX rates? Look on the web.
I don't work for banks any more. I work in the pharmaceutical IT business. I want to work in biotechnology, which means keeping an eye on bioinformatics developments. People like Lincoln Stein, who's been active in the Perl community for some time, and is now an old hat at this bioinformatics stuff, are coming at the problem of handling the bioinformatics information deluge from about the same place that we were at in 1985. Actually, it's a bit better than that : the Web and XML is here ; lessons have been learned about data visualisation techniques ; and disk space/processor speed/memory are not the limiting factors any more. The bioinformatics community is also more open than financial information vendors ever were, so I doubt that the financial information "feed wars" of the late 80's won't occur in bioinformatics. Watch this space ...
Jim Kent is another bioinformatics convert, and developed a lot of the tools needed for the AceDB project - and then GigAssembler for Celera to do the genome assembly at UC Santa Cruz ... and the human genome browser ... let's face it, the HGP owes Jim a bit of a debt, don't you think?
Posted by daen at February 24, 2003 10:28 PM