David Ascher reports on a filesystem-based tagging technique developed by Stephen Hahn, who works (and blogs) at Sun. On his personal blog, Stephen describes a Perl-based prototype that represents tags using directories and symbolic links. The scheme can chew up a lot of these resources, though filesystems with file attributes -- he mentions Sun's UFS and Linux's ext3fs, I'd add Microsoft's NTFS -- would reduce the resource cost by a lot, at the expense of portability.
Given the popularity of tagging on the Web, Stephen observes:
It's therefore a little strange that no one has written a simple utility to give you similar category construction capabilities on your typical Unix-like file system.I think the reason is that the tagging craze is mainly fueled by social effects. It's true that I tag my own content online in order to help me keep track of it, but I'd be unlikely to do so if I were the only person engaged in that process. Categorizing my stuff to help other people find it is part of the motivation, but it runs deeper than that. Observing who else bookmarks my items, and what terms those folks use to describe my items, helps me measure the effect of my work. Finding other items those folks have bookmarked and tagged leads me to new resources relevant to my work.
That said, there's a huge opportunity to bring this potent hybrid of personal and collaborative information management to the corporate intranet server. Recently, when I saw a demo of the Xythos document manager, I was mentally superimposing del.icio.us onto its search and navigation screens. I suspect that enterprise content management will move strongly in this direction. If it does, the underutilized metadata features of various file systems will finally get a workout.
Former URL: http://weblog.infoworld.com/udell/2005/07/14.html#a1267