What's easier to understand, google pageranks or the clotting cascade? Both seem based on well-defined, well-studied mechanisms but are, in fact, frequently unpredictable. Try dosing someone's coumadin after a course of azithromycin and a diet heavy on leafy vegetables. Their INR could be 2 or 8 and I'd believe it.

It's the same with Google's pagerank. A mediblog from a certain rising second year had a page rank on 7 last week -- the highest of any medical blogger, anywhere, and equal to the PR of some of the most popular blogs on the web. Today, his site is down to 3. (At least, that's what's reported on my google toolbar. The real PR used by google could been different. But the proof is in the traffic -- I wonder if he noticed a boost, and now a decrease, in visitors arriving from searches).

And that's just pagerank. As for actual search enging results position (SERP), I'm just as confused. For instance, for a few weeks after this post, my blog was the #1 site for the phrase "dammit, jim" -- Then it dropped to #14, and now it's #5 or so. The recent movement up is likely due to blogborygmi's new pagerank, transient as that might be. But there appears to be a time-sensitivity to search terms, as well. Terms in recent posts get a leg up on other returned google results. I've searched, but haven't found this noted elsewhere.

Also, as a point of pride, blogborygmi had been the sole returned site for the google search of "grapeshit." Then, for a few weeks, all references to this fabulous word vanished. Then, google returned six results from all over. Now, two. How am I supposed to monitor progress as this term sweeps the nation?

Based on these examples, it's hard to believe I'm trying to make a point. Bear with me! I just met a toxicology student who, when confronted with a novel patient ingestion, has more luck with google than pubmed. And he's not the only scholar using Google results for urgent patient care.

Google is a private company, of course, and they provide a free service that works very well. But if the internet is becoming the central repository of human discourse, and Google is the indespensible guide to the internet, there should be more transparency about how it works. Why do some terms rise and fall, while others disappear and come back? The answers might be of trivial importance now, but someday it could be lifesaving.

Yes, disclosure of the pagerank and SERP algorithm allow for exploitation and duplication of Google's functionality. Even without the disclosure, people have tried to manipulate the system, only to see their sites seemingly punished. Is Google really this petty and arbitrary? Is this what's necessary to make and maintain such a good search engine?

The bottom line is, will scientists and scholars remain content if their reference librarian pulls books and journals out of a black box and says, "trust me, this is what you want" ? Maybe... But only if they're always, always right.