Why Google's Online Enyclopedia Will Never Be As Good As Wikipedia
slate.com
slate.com
I swear sometimes Google is run by geeks who read Ayn Rand one day and thought "gee the market economy can solve any problem!"
So, this is the old problem of how to rank... Maybe social news style voting? But different from reddit/dig/HN, because it wouldn't be "news", so newness shouldn't be a key factor.
Searching on Google probably is just the regular searching, so it works the same as any other site.
Searching on Knol seems to allow duplicates, so there is a rating system.
Wow, that thought experiment actually worked better than I thought it would!
But how is pagerank determined? In the usual way, of the aggregate pagerank of incoming links (and other factors that google uses).
I was suggesting that there might be a better way to rank knol articles, than only using Pagerank. This might not turn out to be the case; but that's what I was suggesting.
Also, I really think the fixation on Pagerank is quite a devious thing Google has done, because it completely obscures all the other complexity of a search engine. Be it stemming, synonyms, inferring search intent, etc, most of that stuff doesn't really come out of Pagerank. There are "thousands of signals" or whatever it is that the Keepers of the Search have, and they all play varying roles in the actual ranking process. Could knol articles be more effectively winnowed than by web search? Yes. Would it be appropriate do treat knol articles specially in web search? Probably not (imagine user backlash).
Within Knol, they also have the advantage of ratings and other stuff though. What really bothered me about your initial comment was the "maybe they use some social news site style voting"... you could have actually looked and said "oh! they have ratings", instead of just blindly pontificating.
I've heard that point of view of Pagerank before, as a marketing device. I personally do like the idea that an algorithm can make a difference, and I absolutely believe it can. Whether it really did in this case, however, is a question of fact.
What you say about search engine complexity is true; although many of those techniques were known before pagerank (though I'm sure they're been developed further and new ones added since then). The Pagerank patent itself goes to some pains to mention other techniques being integrated. But I think the significant thing about Pagerank is that it really was something new.
Google's overall success had many factors: fast results; an uncluttered search-only webpage; no paid placements in search; and doing a good job on all those factors you mention (though I think the specific ones you state are to do with search, and not with ranking - at the time, there was no problem finding the right result; it was ranking it near the top that seemed to be the issue... at least, according to Google PR...).
To be honest, for me personally, I think the leanness of Google was most appealing (fast, simple, honest), together with a sense of coolness, of it being new, and by a couple of brilliant guys. It was also (virally) introduced to me by someone cool in tech. The search results themselves were "good enough". I remember doing some comparisons, and finding Google better - but it really didn't matter to me... it just confirmed the attitude I already had.
Now, regarding my "blind pontification" in your opinion: I did look for ratings/ranking - on the front page, and I tried a search, but I didn't see any ratings (That's why I said "I don't see any explicit ranking." It turns out they don't appear on the front page, or for search results; you only see them after you have selected an article. EDIT actually, they do appear in search - but that take so long to load (for me) that they didn't appear (and there's no placeholder), even when I was double-checking.
The "pontification" part bemuses me... I was suggesting/asking/wondering what a good way would be to do it. That's the opposite of pontificating (in my opinion, anyway). I like to wonder about things - it's a source of new insights, and it works for me.
A request: if you think I'm saying something silly, could you straightforwardly point it out to me? People (and I really think especially I) do make mistakes; and I (at least) appreciate being set straight.
I must say, I've got used to straightforward discussions on HN, but it seems to have shifted recently. I feel sad about that.
So in summary: I thought you were being silly but it turns out this was just a big understanding. I hope this doesn't make you hate me or HN.
I'm in Australia. The text loads instantly; but the ratings take 7 seconds (I'm sure it was longer last time), but with no indication that something is missing...
In this type of situation, when you think I'm being silly, my request is for you to say: "Knol does X". Or "Knol does X and you're silly" is fine too. As it was, I had no idea where you were coming from.
And I could spend more than 7 seconds on a site before commenting. =)