Hydrocephalus?

4 comments

I don't know what Talltroll has been drinking, but he seems to have water on the brain, what with his new blog post entitled "Trust is a reservoir, when it should be a river"

In all seriousness it's superb commentary about algorithms "probably" in place right now and well worth a read!

hydrocephalus inspired algo busting

Comments

Really is a good post

It would change SEO a bit if the search engines got that smart - you would need to really integrate the trust building into the month-to-month business.

page rank?

Given Google's addressing "spam" during the time of the development of the observed trust issues, is a high "trustrank" domain actually one that has a low "spamminess score" ? Viewing it that way, one can easily rationalize permitting Amazon to maintain trust because of the shear number of it's non-spammy pages. The spammier the SERP, the higher they turn that "re-rank the pages by spam score of the domain" knob.

The higher you turn those kinds of dials, the more size matters.

Citation indices have always been flawed

because at the end of the day they're merely a primitive, more often than not pathetic compensation for crappy on-site analysis algos.

Google's PageRank was just that from its very inception - an academic citation index, overblown to cover the WWW.
However, what does it essentially prove if scholar X has been quoted 1000 times, whereas scholar Y has merely achieved to 100 citations? Maybe Mr X is simply better at self-promotion, more controversial, the laughing stock of the academic community, better connected with the media or published by one of the fat cat publisher? (I.e. the superior "link baiter" ...)
Whereas "unlucky" Ms Y may be specialized in an entirely unsexy field only a handful of people will understand anyway, counting herself lucky to be published at all by some picayune backwater institution, possibly even in some unpopular language. So what?

Plus - what makes the two comparable in the first place? If you're working on demonstrably refuting the second law of thermodynamics, chances are you'll get way more attention than if you're into research focused on the spectral analysis issues of fungicidal housepaint color chromatography. But does that make the one more "important" than the other? Nonsense!

Add to that the issue (never seriously addressed) that a commercialized Web is dominated by entirely different mechanics (competitors linking to competitors? Fat chance!), and it becomes pretty obvious why PR had to be demoted almost to oblivion years ago. (Not that too many members of the SEO community seem to have caught on ...)

As a searcher, I'm primarily interested in what I can expect to find at the end of a SERP link, because that's where I expect to spend my time (and, possibly, my money). What other people or sites may think of it is essentially immaterial unless I specifically choose to follow someone's express recommendation (e.g. directly from their site, via e-mail communication, offline promo, etc.). But that's quite a different approach than using a search engine for filtering and grading tons of purportedly similar options.

Maybe all this bias of confusing quantity with quality was systemic and unavoidable at the start due to the all-dominating "network" and "community" paradigm, blithely extended by today's focus on social bookmarking, tagging, blogging, overhyped wisdom-of-crowds expectations, etc.

Obviously, the one approach will be as vulnerable to manipulation as the other.
- When it was on-page body content, in came invisible text.
- When it was meta tags, keyword stuffing had its heyday.
- Directory referrals? Yahoo! are still making a mint of that.
- Keywords in your domain name? The birth of the 0800-free-prescription-drugs-viagra-viagra-viagra.com site.
- When it was links galore, link popularity programs was what people went for.
- Vintage domains? Snapped, auctioned, sold at a premium now.
- "Trust" sites? No problem for those with a padded wallet, either - just go and buy some.

And: What does "trusting" sites, people, links without actually analyzing (let alone spidering) what they're referring/linking to boil down to if not mere hearsay? Call it RumorRank or TabloidRank if you will, big deal, but don't try to dupe me by terming it "relevance" ...

Damn, fanto, that post's

Damn, fanto, that post's nearly as long as my original...

>> Obviously, the one approach will be as vulnerable to manipulation as the other.

I think there will eventually be an approach that's immune to manipulation - but we are a loooong way off that. Until you can teach a machine to actually understand a site much as a human would, you won't seit.

By then, I suspect that SEs as we know them won't exist. We'll be in the realms of intelligent agents, in effect everyone having their OWN SE, customised just for them. The current SE companies will become data brokers

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.