Sphinn - SEO For Semantic Search Engines

50
Sphinn It!

SEO For Semantic Search Engines

Went Hot: July 14, 2008 - 2:34 pm
Posted By: UtahSEOpro 89 days ago
Topic Type: News Story (Jump to http://searchengineland.com)
Category: SEO

A new generation of search engines are starting to become publicly available and so it's time to start thinking about how they will affect SEO efforts.

6 Comments

Who Desphunn This (0)

Comments

from Cognition 89 days ago #

Votes: 0 | Vote:

+ -

A couple things to add regarding Cognition and semantic search engines:

First, Cognition also searches on top of the Medline database to allow for a semantic searching of Medline. Plus, Cognition has a case law database for semantically searching federal and Supreme Court decisions since 1950.

Second, On the second query (Who makes diet coke?), the writer doesn't mention Cognition's result. Interestingly, if he had capitalized Coke, he would have immediately found the answer! This is another aspect of semantic search engines that people need to remember: they are case sensitive due to capital letters changing the meanings of some words.

But overall, a great article. Thank you.

from rizaberkan 89 days ago #

Votes: 0 | Vote:

+ -

Good approach overall, however examples focusing on "who" question only is a sad choice to make a comparison between these search engines. The most common question type is "what" and you could ask something like "what drug treats headache" to see what is happening in the results. I have explained in hakia blog that one needs at least 200 different queries to make any kind of assesment about a search engine. Here is the link http://blog.hakia.com/?p=151.

from eKstreme 89 days ago #

Votes: 0 | Vote:

+ -

Thanks for your comments. It may not have come across well in the article but I'm a big fan of all these SEs and use them regularly.

Re Cognition searching other databases: yes I know but ony the Wikipedia search makes for accurate comparisons with the other engines. An interesting comparison would be Medline/Pubmed searches on Cognition, Hakia, Go Pubmed, and others. However, that would delve into example queries that are not mainstream. BTW, from my experience, the Cognition Medline search is the best out there with eerily accurate results.

The capitalization issue: it's a problem. My feeling is that if a search engine requires case-sensitive queries, people will not use it as often. Searchers are now set in the habit of typing in lower case. Yes a small fraction of queries is properly capitalized, but the vast majority are not. Honestly, I hate that bit about Cognition!

Re testing a search engine properly: absolutely agree. I said this in my post that the objective is to get people to start thinking about these new search engines. No one can do justice to any SE unless they use it every day for weeks. I only pointed out some examples just to show the differences and similarities.

Pierre

from hugoguzman 89 days ago #

Votes: 0 | Vote:

+ -

I don't think that these engines will become mainstream enough to warrant major emphasis in the foreseable future.

It's fun to look into the technology and all, but these outfits simply won't have the resources to handle a high volume of queries. They're also not refined enough to outgun Google at this point.

I think that semantic search will have to wait until voice recognition technology catches up, so that instead of typing in semantic queries, users will simply utter them (or maybe even just think of them).

from onreact 88 days ago #

Votes: 0 | Vote:

+ -

You forgot one of the best semantic search engines, Lexxe:
http://lexxe.com/main.cfm?sstring=Who+built+the+empire+state+building

from linkmoses 88 days ago #

Votes: 0 | Vote:

+ -

Some food for thought RE: the future of links in a semantic search world.

Below is abstracted from my LinkWeek column Semantic Link Building coming next week. Call me a tease...

You wrote...

Not being able to accurately identify authority could be deadly for the success of these semantic search engines.

This notion of authority is also the reason why links will continue to be crucial, HOWEVER, it wont be the links most folks think it will be and are still hopelessly chasing today. Just as Google devalued millions of links across hundreds of non-descript directory wannabees, the engines will apply far more scrutiny as to the links they feel send useful signals, ignoring the ever-growing mass of links that are meaningless. The key will be understanding which links matter and why, having the content that will inspire, engender, and earn those links, and and knowing how to move that process along properly. The web is made up of two things: Pieces of content and links between those pieces of content. That's really it. The billions of links are a quality control mechanism too useful to ignore. The semantic SEs that succeed will be those that seletively ignore the links that should be ignored.