When explaining the power of microblogging (e.g. Twitter) to others, it helps to share my own occasional "Aha!" moments.

On Sunday, Speaker Nancy Pelosi announced that an economic stabilization bill had been written and was available at financialservices.house.gov. I tried going there for about 10 minutes, but the servers were swamped.

So I went to Twitter search, and merely searched for "bill". The immediate results included several comments about the bill, and at least two links to download it from non-swamped servers. This tells me a couple things: I trusted in the critical mass of contributions to Twitter that I would find what I wanted - and sure enough, it was there. It also shows the power of searching the immediate NOW.

A Google search for "bill" includes a 5-hour old news link, imdb's of movies like "Bill" and "Kill Bill" and wikipedia entries on Bill Gates and Bill Clinton. But for the moment I was searching for "bill" on Twitter, there was only one "bill" that mattered most.

This is how Twitter brings together the massive amounts of information being fed it NOW with what I am searching for NOW. This is just one way Twitter works, if you know how to use it.

I'm working a presentation that includes the concept of search as social media, both in terms of ranking algorithms that are based on socializing links, as well as how search results can expose emergent properties of social consensus. An example of the latter arose as I was writing about the gorilla among wikis, and deciding how much that gorilla should weigh.  Let's check Google:

Weight of Gorilla (lbs.) Number of Google Hits 100 1,020 200 1,930 300 13,100 400 12,100 500 33,900 600 14,400 700 1,180 800 242,000 900 23,500 1000 915 1100 9 1200 353 1300 3 1400 8 1500 194 1600 214 1700 0 1800 57 1900 3 2000 1,380

Obviously, "800-pound gorilla" is the 800-pound gorilla of gorilla weights, even though adult males average about 400 pounds, are about 500 pounds at the extreme in the wild, and obese gorillas tip the scales at about 600 pounds in captivity.

Why 800? Maybe someone out there knows. What do you see in the other values? A 100-pound gorilla? Really? The values are higher around realistic gorilla weights, then drop off at 700 before spiking at 800. At 900 they are relatively high but then drop at again 1000. Why not 1000? Is it because people want to say an "X hundred pound gorilla" and not a "thousand pound gorilla" - so "ten hundred" doesn't work? Then, past 1000, there are some little spikes. Not 1100, but 1200. Not 1300 or 1400, but 1500 and 1600. I guess 1600 is two 800-pound gorillas. Then not a single 1700 pound gorilla to be found, but when you get to a ton, you're up there again.

This is a silly example, of course, but with a little thought, I'm sure you could construct some searches that show you social consensus points for other kinds of data.

The folks over at Smashing Magazine have written a great, highly-researched article called Google PageRank: What Do We Know About It?. It explains what Google PageRank is, how it's calculated, and why we should care.

Jason Calacanis, formerly of Weblogs, Inc and AOL, launched a new search engine a few hours ago called Mahalo. Although the site will tell you that it means thank you in Hawaiian. I already knew this because I've heard the Laurie Berkner song about a hundred times already. (It's my older son's favorite disc right now). I bet Steve Nelson knew it too, but he's frequently in a state of aloha.

Mahalo, the Web site, has a simple premise. Provide a page for each of the top 10,000 search terms. The information on the page is written an edited by a responsible person employed by the site. Robert Scoble suggested it was kind of like Yahoo! when Yahoo! was just a directory.

They are calling themselves a search engine, but I think it works more like Wikipedia, except you might have a bit more trust that the content is credible. I might go to Mahalo first if I were trying to look something up as if looking it up in an encyclopedia. I wouldn't use it to figure out how to use javascript to dynamically update the cascading styles on elements of an unordered list.

The top 10,000 search terms idea is interesting because it runs counter to the idea of a long tail. Google is very much oriented towards the long tail. It indexes lots of obscure stuff that only a few people care about. Jason's site only indexes the head and part of the neck. Or maybe it's just the forehead. I don't have a good idea of how many keyword phrases fit into the beast that is all things people search for on the Web.

I wonder where the top 10,000 search terms come from. It would be interesting if they paid attention to the terms people typed into their search box. It would be really interesting if people could do the equivalent of a google bomb by linking terms to pages that have yet to be written. For example, why isn't there a Leon Atkinson page?