I told both Jason and CK awhile back that I'd pass along my thoughts. I wanted to sit back for awhile and watch how things all started to develop before offering up anything.

There are two users to a search engine. You have the person searching for information, and you have the person providing the information. In the case of Mahalo, the first use-case I get. Curated, edited, focused information that I'm looking for. In fact, I had my first Mahalo moment while searching for information on Ted Leo and the Pharmacists, and Mahalo was a big help. Awesome.

But when Dave Winer mentioned that Jason didn't bring a Win-Win to Gnomedex, it struck me as the words that I was looking for to describe the second use-case, although he probably meant his words in a little bit different context.

The second use-case falls flat on its ass. There is currently zero incentive for a blog author/content producer to support Mahalo. On Google, the value proposition is clear -- the more people a producer suggests to Google, the more people who are using it. The more people who are using it, the more traffic the producer's site ultimately gets. Yes, some people cheat and get more traffic than they should, but ultimately everyone has a chance to receive "votes" by way of organic links. It's a classic win win.

When Mahalo first launched with curated results, the existence of this proposition was at least debatable. Google backfill kept the basics of a search engine in play. Now that Mahalo is producing original content instead of SeRPs, however, it isn't really a search engine, but a knowledge base. From a content producer's point of view, why on earth would they want to contribute pagerank/trust rank/support to something that will ultimately overtake their keywords and traffic on Google? The slightest change in the rankings can dramatically affect a sites ad sales/conversions/whatever. Supplemental information makes sense I suppose, it's just like linking to a blog post. But again, that's seems like an awkward conflict of interest in the role of a search engine.

Remember when Google started giving their properties premium placement in the SeRPs? The damn near riot that almost ensued? People were grabbing pitch forks and lighting torches chanting "do no evil". That's nothing compared to the outrage people would have if Google started indexing their own open-ended random content on all subjects ahead of their second group of users, the content producers.

Mahalo will probably succeed with whatever Jason does. Answers.com, about.com, wikipedia, etc are all really popular high traffic sites. They just all happen to also be incredibly boring. If Mahalo wants to be exciting and redefine search (which seems to be J's stated goal), then it needs to create a win-win for the people who are building out the rest of the web. Figure that out, and Mahalo stands a chance to really help a lot of people.

UPDATE: In fairness, I should note that while sitting here thinking about the win-win comment, I saw the mahalo "how to write a resume" link scroll by on delicious. Talk about putting it all into context ;)
UPDATE #2: Both Jason and CK respond in the comments.

Vander Walhas an interesting post that's in reponse to another interesting post byMatt McAlister. Both seem to suggest popularity scales may be overrated.Many websites rank items via some sort of popularity. Digg.com uses voting mechanisms, everyone seems tohave some sort of "Tag Cloud" going, some measure clicks or email forwards orpage views. The thought is, the more popular an item on the web, the higherquality of that item.

But what about niche topics? Is coverage on these sites being "normalized", asMcAlister writes? On the face of it, popularity is relative to everything else.If I'm really into pet rocks, a really popular pet rock page isn't going tobe popular in relation to, say, Chinese kids singing like a boyband. If I'm searching out specific information within a somewhat obscuretopic, ranking things by popularity would only seem to make it harder tofind.

Vander Wal writes that popularity gets in the way of information seeking.Maybe, but it would seem that keywords intermixed with popularity would helpmake results more relevant naturally. After all, Google seems to handle mysearch for "petrocks" just fine, and they rank largely on popularity (in fact, thats kindatheir whole gig ;). On the other hand, digg.com might not provide the bestmeans for searching out a topic, but is it supposed to? Seems to me digg is asite that allows you to see the most popular information floating around rightthen and there. Sure, it's collecting meta-data, but, I mean, does digg evenhave categories or tags? It's category is "Technology", and it's tag is"popular." The rest of the meta-data is decent, but suffers from occasional lowquality (as opposed to tags, which would balance themselves out if properlyimplemented).

Still, there's plenty of information that I'm sure is thrown to the waysidebecause it's not showing up on the popularity radar. Is there a way to judgeit's content in a way other than popularity? Sure, subjectively by an editorialteam, but that's simply infeasible on a large scale as it would require expertsin a ton of different catagories. Wikipedia meets dmoz maybe? I donno.

Anyway, I think proper search mechanisms in tandem with popularity measurementis a solid approach. Either without the other doesn't quite cut it in terms ofinformation (re)location.