After spending a couple years of my professional life trying to improve the state of metasearch (federated search, if you prefer), I took a little break from being so passionate about the subject. But, like e-books, it’s one of those library technologies that I just can’t seem to shake an interest in. And it seems like librarians are loosening up a bit and letting patron demands and satisfaction with “good enough” search results replace an entrenched skepticism about the adequacy of metasearch. It’s still amusing to note that it took metasearch for librarians to beatify content providers’ native interfaces, which before their advent were also often described as inadequate.
In the meantime, there have been a few more interesting developments in the metasearch space. Serials Solutions has announced that it is about to launch Results Clustering by Vivisimo for its Central Search product (there’s an informational webinar on the new feature on Tuesday, Oct. 10). The automated classification tool organizes search results by author, title, subject, and journal title as the results are returned. Ex Libris has also signed an OEM agreement with Vivisimo for MetaLib. And (ironically or fittingly, I am not sure which), Frank Bilotto, formerly head of publishing at Vivisimo, has joined MuseGlobal in the new position of vice-president, publishing and digital media.
The development comes at an interesting juncture, as faceted classification in online catalogs begins to take hold in libraries as well (faceted browsing and clustering are really two very different technologies–perhaps fodder for a future post or technical face-off).
The other day, I had the good fortune to stumble upon a federated search vendor that has been around since 1999. Deep Web Technologies only recently began exploring the library space, but might be familiar to librarians who have searched Science.gov. I had a great conversation about the past and future of metasearch with the company’s founder, Abe Lederman. Though primarily in the government and open web space, the company has also recently worked with a Fortune 50 company to enhance its enterprise search capabilities. (Side question: Is “enterprise search” replacing “metasearch?”)
It’s obvious (as in my self-interest) that I am intrigued by looking a little outside the traditional library automation space for new ideas (cough…Endeca…cough). Obviously vendors like Ex Libris and Serials Solutions are doing the same thing through partnerships with a company like Vivisimo, so why shouldn’t libraries themselves do the same?
[This post originally appeared as part of American Libraries’ Hectic Pace Blog and is archived here.]
Actually, clustering subsumes (i.e., includes as a special case) guided navigation. That is, it’s easy under a clustering regime to set up a fixed meta-data field as the ONLY text to use for clustering. Or, to use a fixed meta-data field to form the clusters at the top level of the folders, and then use the spontaneous clustering at lower
folder levels.
You can see one variation on this idea at http://www.asme.org (mechanical engineering) which shows meta-data-based folders by default (“channels”) and then shows spontaneous, free-text clustering by clicking on a tab (“topics”).
Actually, clustering subsumes (i.e., includes as a special case) guided navigation. That is, it’s easy under a clustering regime to set up a fixed meta-data field as the ONLY text to use for clustering. Or, to use a fixed meta-data field to form the clusters at the top level of the folders, and then use the spontaneous clustering at lower
folder levels.
Raul Valdes-Perez
Wanted! Library Partner
Deep Web Technologies is actively looking for a library partner that we can work closely with to deploy our state-of-the-art federated search engine. We can provide the right library with our Explorit 4.0 software and assistance in customizing a solution or we can provide the library with a service running at our data center all for FREE for the first year and discounted beyond that.
What we ask in exchange is feedback on our product and for the library to serve as a reference installation for us.
If you are interested in partnering with us, or would like to get more information please contact me ASAP at: abe@deepwebtech.com