- Product Trial
WorldWideScience.org just received yet another pat on the back from the blog Inside Science Resources. Khue Duong, Science Librarian at California State University, Long Beach, posted a review titled, “WorldWideScience, a Cut above the World Wide Web.” Yes, the title summarizes it well.
Users are often surprised to learn that the engine behind WorldWideScience.org is a federated search engine, bursting the myth that federated search is an antiquated, dusty bit of technology that doesn’t aggregate source results and is dreadfully slow. Here’s the first sentence of the review:
Are you still looking for a reliable federated search tool that goes beyond the run-of-the-mill results?
Enter Deep Web Technologies’ next-generation federated search, Explorit Everywhere!, the powerhouse behind WorldWideScience.org. In the review, Khue Duong explored the search functionality (including Advanced Search) on WorldWideScience.org, and performed four searches to test the application results. Three of the tests performed well, returning relevant results:
Without displaying the additional results, searching “Isle Royale National Park” as Full Record yields 66 papers and 49 data sources. Regarding papers (document type: articles, reports, etc.), the first 40 displayed results seem to be to the point; the rest probably has some mentioning of the search term in its document. The first twenty data results has some mentioning of Isle Royale National Park in the title. The rest, such as the DNA sequences from the DNA Data Bank of Japan, shows that the sample originates from Isle Royale.
One of the four searches didn’t yield as many relevant results.
Note that adding an additional layer of parentheses in the search of (“isle royale national park” AND (wolf OR wolves)) also yields many misleading results in all three categories: papers, multimedia and data.
This is hardly surprising. On public search portals like WorldWideScience.org, many sources do not support title search well. Trying the above search as a full-record search brought back many good results, including the top-ranked results containing all of the search terms in the title.
We would also like to point out that the additional search results should be displayed when the search completes as these additional results may contain some of the most relevant results available. Also displaying the additional results should significantly reduce the variability in the results returned by running the same search multiple times.Khue Duong, however, still recommends WorldWideScience.org as a science resource for researchers:
Overall WorldWideScience.org is another one-stop-shopping platform that one should consider when searching for publications, data, and media files from governmental and international scientific domains.
Over the last year, WorldWideScience.org has garnered lots of attention as not only an excellent search portal for scientific information, but a multilingual search portal as well. In January 2015, Microsoft published a Case Study about WorldWideScience.org. A few months later, Deep Web Technologies published an article in Multilingual.com magazine. And WorldWideScience.org isn’t the only portal using our federated search. Explorit Everywhere! is used for several other public search portals, such as Science.gov and Askia.Uneca.org.
We’re proud of WorldWideScience.org and proud to support the DOE Office of Scientific and Technical Information (OSTI) in bringing this free, publicly available, multilingual search portal to science and technology researchers around the world.
Published on Mobile Groove
Author: Charles Knight
Today we are kicking off with a new look and a new series of alternative search engine iPhone and iPad app reviews and road tests with Biznar, a mobile search app available for both the iPhone and the iPad that takes a unique approach to searching for business related information.
Why “alternative?” Because that’s where the excitement and the innovation is! Sure, our smartphone devices present us with three whole choices — Safari / Google, with Bing and Yahoo! (GBY) available in your ‘Settings’. However, our requirement (even demand) for more personal and precise search services and results on our personal mobile devices opens the door for dozens of alternative search engines to make their mark — and gain market share.
That takes us to Biznar, a serious and comprehensive business search engine developed by Deep Web Technologies that uses a technique known as Federated Search to deliver results that GBY do not. (Biznar also has a sister app Mednar, which specializes in medical search.) More about the company and its mobile strategy in the companion post.
Deep Web Technologies has built a reputation as the “researcher’s choice” for its advanced information discovery tools, and I concur. As I stated in the press release when their App was released in November of 2011: “For business and medical research, Deep Web Technologies has selected just the right approach for their Biznar and Mednar Apps: advanced Federated Search.
Why is this approach significant? Because Biznar takes a different tack than crawling and indexing the Web. That’s a good idea since the Web includes good and bad source material, old and new articles, legitimate and not so legitimate scholarship and a whole lot of marketing — all of it warehoused for some period of time by the time your search query comes along.
On the (mobile) mark
Biznar has preselected a corpus of around 70 high-quality business sources and searches them all in real-time the moment that you launch your query. These sources include leading information destinations such as ResourceShelf and MarketResearch.com and Fortune, as well as Deep Web Technologies partners APA PsycARTICLES and APA PsycBOOKS.
Barbara Qfederated search article. “A good federated system imposes a tremendous burden on the builders so the users can feel the search process as effortless.” Indeed, at Deep Web Technologies, that is exactly what we feel we are doing. We’re creating search systems for our clients that require very little effort on their end, but do exactly what they need. Her assessment of what produces quality searches (or the lack of them) is spot on as well:
More important, however, are the problems of truly making the systems perform effectively for end-users. Basically, a lot of human intelligence and expertise, not to mention sweat and persistent effort, has to go into these systems to make them “simple” and effective for users. For example, most of the databases have field structures where key metadata resides. A good federated system has to know just how each field in each database is structured and how to transform a search query to extract the needed data. Author or name searching alone involves layers of questions. Do the names appear firstname-lastname or last name-comma-firstname? Are there middle names or middle initials? What separates the components of the names – periods, periods and spaces, just spaces? The list goes on and on – and that’s just for one component.
The article also mentions our company President, Abe Lederman as well as several public-facing portals powered by Deep Web Technologies. Thank you, Barbara!