Is there a future for semantic search engines

[Roger Martin]

Nowadays, most semantic search engines DON'T WORK - better off with keywords, most of the time. Is there a future? Is someone working on this?
Also, won't XML metadata make the debate redundant?
I've checked them all: NLight, Lexiquest, Arisem, Subjex...don't seem to do the job.

I would appreciate any insight on those questions.

Thanks

[caine]

Hi Again,

Not quite sure what you mean by a semantic search engine. They don't understand real world concepts, so deeper language analysis, that humans are capable off are definetly out of the reach of computers, programming and electronics as a whole.

The search engines do work on a principle of 'key word analysis', they are just not very good at it when you take into account the amount of information they have to reference in order to meet your search criteria, it becomes quite obvious that by today's technological progress, that the searching is extremely limited, maybe to a few words, which allow a speedy answer

In regards of XML, i think you could be possible correct, but xml is not a speciality of mine.

[InternetEsquire]

I am working for a company that is developing a semantic engine that will be installed on top of a two-tier relational database. The two-tier relational database is a 64 bit application that organizes node data into hierarchical categories with multiple parent paths -- this in striking contrast to symlinks. The result is not unlike a neural network.

The beauty of a 64 bit two-tier relational database is that searches that would otherwise take several weeks to complete can be performed in a thousandth of a second. However, the system is interactive, and it will require the input and feedback of millions of users to create a semantic engine that can pass the Turing test.

[caine]

Internet Esquire

Could you give me the url of the company that is developing the two tier 64 bit engine.

I think you are right about it not being able to pass the turing test. The complexity of the original assumptions and the premises noted, suggest many decades before we are even close.

[InternetEsquire]

caine,

I'd post the URL, but I don't want to find myself in hot water for self-promotion. If you e-mail me, I'll be happy to give it to you.

As for the time to get this up and running to the point where it can pass the Turing test, we're probably looking at three years.