Comments on: Federated Search Finds Content that Google Can’t Reach Part I of III

By: shakuntala

shakuntala — Mon, 15 Feb 2010 16:07:23 +0000

Dear Sir, I am doing Master of Library and Information Science, I have choosen the topic Federated search engines for my seminar. Please guide me what subtopics i should cover under it?
which are the famous federated search engines available?

By: Deep Web Technologies seeks web designers for theme development project » Federated Search Blog

Fri, 11 Sep 2009 17:12:49 +0000

[...] to construct the CSS and graphic files for a theme for our upcoming Software-as-a-Service based federated search [...]

By: A Federated Search Primer - Part III of III | Deep Web Technologies Blog

A Federated Search Primer - Part III of III | Deep Web Technologies Blog — Tue, 24 Feb 2009 08:41:19 +0000

[...] is Part I and Part [...]

By: Dr. Michael R. Alvers

Dr. Michael R. Alvers — Wed, 14 Jan 2009 11:06:22 +0000

Let me put a more philosophical perspective on the issue. What would be an ideal “search engine”? I’d say the one that finds my answers. Either a simple phone number (at one end of the spectrum) or the meaning of life (on the other EXTREME end). Questions like “what are the investments in biotechnology in Europe in 2008″ fall between the extremes. My ideal search engine is an expert. An expert has the necessary knowledge and can communicate with me.

Of course nobody wants to fill forms. We want answers! Scientifically correct, comprehensible, complete, fast and with a probability of correctness (physicist call it error bars). I think it is just a matter of time when the deep web also will be “crawled”. Filling out forms by using APIs (application program interfaces) should not be so difficult. And what keywords to search for is also clear: the keywords given to e.g. Google. I don’t see this as a major problem. The major problem is the lack of semantics behind “keywords”.

Keyword search is not enough. Searching for “heart diseases” leads (in PubMed.org) to about 50,000 results. Considering “all known” concepts (to an expert – or GoPubMed.com using MeSH = Medical Subject Headings) one will find 850,000 results. Ok, no one can/will screen 50,000 or even 850,000 results. It again needs expert knowledge to drill down to let’s say “Heart Diseases” and “Early Growth Response Protein 1″ (known to experts). The remaining 23 articles (found with GoPubMed) are highly relevant, scientifically correct, complete (as the background knowledge is) and the results are delivered with 6 mouse clicks. Fast. Try to find these articles with one of your favorite search engines.

By: Greg Lindahl

Greg Lindahl — Tue, 13 Jan 2009 00:02:47 +0000

toxnet.nlm.nih.gov has a robots.txt which bans searching.

By: thegeniusfiles

thegeniusfiles — Sun, 11 Jan 2009 18:11:31 +0000

Very interesting article. I’m looking forward to the next article. As a university student in computer science and information systems, this subject is both endlessly fascinating and endlessly frustrating to me. I have used several of the search engines available the University of Wisconsin system. They can indeed find information which is difficult to obtain through other means; however, the user experience is uniformly terrible. This is a field which would greatly benefit from more attention to the user interface, and the presentation of results.

By: Hope Leman

Hope Leman — Sun, 11 Jan 2009 15:44:25 +0000

Very edifying. This essay should be required reading in all library science programs (like the one I am enrolled in at the University of Pittsburgh). This is an outstanding primer on a subject that should be better understood by anyone interested in search.