I have heard librarians say to students – more than once – that they should use library databases because we don’t really know how Google ranks things. But it occurred to me the other day that I know a lot more about Google Page Rank than I do about, say, Academic Search Premier’s relevancy algorithm. I’ve often wondered why, for example, a catalog search might bring up literary criticism of Tom Sawyer before it actually brings up the text of Tom Sawyer. I would like it if we held our vendors more accountable for things like this.
Another thought that occurred to me today while discussing Summon with some coworkers: should availability influence relevancy rankings? Should items that a library has in full text be weighted more heavily than those we only have citations for? On one hand, I can see the utility to the user – they’re more likely to get things they need more quickly. On the other hand, will they miss good research – seminal works in a particular discipline – this way? I think there’s utility in both approaches. Availability could be a bigger factor in say, Summon, while actual relevance would be more important in a subject-specific database.
That said, I can’t think of any library products or vendors that rank search results based on availability. It’d certainly be tricky…