I just finished reading Brian Lamb's article on wikis. Generally I was impressed, Lamb seemed to take a fairly critical view that I think was lacking in a number of the other articles. However, I do take issue with his brushing off of the structure issues that can arise when collaborating via a wiki. He seemed to suggest that user problems with wiki structure stemmed mainly from unfamiliarity rather then genuine issues with usability. Much of his argument in this regard rested on his faith in the inherent search functions of a wiki. This attitude seems counter to the library and information science perspective. I certainly don't want to bring up and go through the many arguments for proper cataloging and sort over search, but they do exist and are prevalent enough that they shouldn't be ignored.
Lamb also suggests that recent additions and update lists, as well as other assorted tools would somehow address the lack of structure issue and ultimatly improve findability. I couldn't disagree more. If anything tools such as the recent change list hinder findability by focusing user attention towards a small subset of the total information contained in the wiki. In terms of searching the recent change list is no different from a new book list in the library, while interesting and full of helpful information it is not a tool for finding things.
Contextual links also come with their own problems, particularly when combined with unsupervised editing. If the only structured means of finding a page is through a contextual link, all that is required to orphan the page is for the link to be edited out of the related article. Given the known problems with freetext searching a page that has lost its single contextual link may never be found.
None of this is to say that wikis are not a wonderful collaborative tool, but to ignore the issue of structure risks making a waste of a lot of positive effort. It would also seem that the inherent self organization of groups fades as the group grows increasingly diverse. As libraries of all sorts generally deal in diversity, the issue of structure, it would seem, should be even more relevant to them.