Briefs

eWeek (05/31/04) Vol. 21, No. 22, P. 56; Coffee, Peter

Peter Coffee reports that innovation in database technology is proceeding apace, and highlights many of the breakthroughs and potential uses--not to mention the more controversial facets--of database advancements. He cites a recent report in Nature that details life sciences-oriented features of the Oracle Database 10g that could be applied to the enterprise, such as pattern recognition, embedded machine-learning algorithms, support for vector machines, unstructured text mining, and regular-expression searching. Coffee recalls a conversation with Microrim's Wayne Erickson that focused on the difficulty of searching for data that cannot be encapsulated numerically or textually, and mentions that MIT's iDeixis project is trying to meet that challenge by comparing cell phone camera images to the characteristics of other images located by Web-crawling servers. Potential useful applications of iDeixis lie in the Web's massive repository of imagery for matching against what the camera views. Coffee notes that expanding the usability of databases remains a formidable challenge, and refers to Erickson's observation of a substantial distinction between meaning and appearance. The chasm between the two terms can be crossed through data annotation features, and Coffee writes that database vendors are under pressure from life scientists, music collectors, and digital photographers to roll out richer and simpler features. The author points out that improved databases have their troubling aspects, particularly in how they could give people uncomfortable access to other people's personal information.

Here is the rest of the story.