Last week law.com picked up an interesting item from American Lawyer “Software Glitch May Have Erased E-Mail Text in Enron Suits”. The information in the article is inconclusive as to the extent (if any) of the problem or the resolution.
News items such as this are a good reminder of the number of human decisions required to create an “automated” collection of information (or knowledge for that matter). I’m guessing it is impossible today to create a large database and actually test it to make sure it is 100% complete and accurate. Whether it is internal system or third party supplied legal research content, we tend to assume (because we must) that we can rely on it (of course, our contracts with third party vendors probably provide that we have no recourse if there turns out to be an issue in any event).
I do not doubt that extensive testing is done by vendors before they market their products. However, other than finding an issue in the course of using the material (eg searching a legal research database and not finding a case you know should have been retrieved by the particular search), I don’t have any thoughts on how to practice “safe” computerized research.