http://www.imagingmagazine.com/db_area/
archs/2002/10/tfm0210f1.shtml?contentmanagement
“There are two schools of thought as to when metadata should be applied to content (in a process known as metatagging). The first school advocates applying tags at creation, a theory that, not surprisingly, many vendors in the content management and taxonomy software vendor community support. The second school calls for categorization of content at the search end using various algorithms that analyze content for meaning. These algorithms aren’t dependent upon metadata applied to the content along the creation path. This method, again not surprisingly, is championed by categorization and search vendors.”
In some recent discussions that I have been fortunate to sit in on, there have been some (very real and valid, imho) concerns raised over the incredible overhead and other implementation challeges of cataloguing or meta-tagging learning objects. I know this wouldn’t be news for the folks thinking about learning objects and LOR implementation in Canada and abroad in the post-secondary context, but it would be nice to hear more consideration about currently existing ways of automatically generating metadata and classifying objects. This article points to a number of existing vendors in both taxonomy and automated classification software. None of these are cheap (or even work that well, from my limited exposure to them) so some of the motivation to develop new tools to do this within post-secondary may be about cost and that this field is still pretty open. Still, I’m a little leary of ‘build’ decisions that happen before the buy ones have been fully evaluated. And again, imho, it’s not the case that one needs to stop evaluating this – for me this is the reason behind iterative development processes that include ongoing business and requirements analysis throughout the entire lifecycle of the project, not just at the start. <end of rant>
Anyways, this article is worth a read. Whenever I used to teach ‘searching the internet’ classes I used to try to help the users (usually very new users) distinguish between pure search engines and directories, and to realize the scenarios in which using one or the other made sense. Maybe someday soon that explanation will no longer be necessary. – SWL