First day on the job and I have been told that the library will forgo 360 Search do to budget constraints. The web development team and I need to come up with a new system to be put in place by March 1st…no pressure! i realize that what we put in place does not have to be the final word on the matter, but it will need to be a reasonable solution. I found a list called “What determines quality of search results?” which has been very useful. I have also started to look around the web and so far here are the possible directions to go. I have not played with them yet, but if you have any opinions or experience with the products, let me know.
MasterKey Metasearch by LibLime is free and open source, if your library can host it. It searches:
- digital collections
- websites
- subscription databases
- electronic publications
- one or several library catalogs
- Project Gutenberg and Open Content Alliance: over 235,000 Ebooks available
- Wikipedia: almost 1.5 million encyclopedia articles
- OAIster: over 10 million repository resources from academic institutions around the world
- Open Directory Project: almost 5 million human-cataloged Internet resources
dbWIZ is another open source federated search tool. This tool is already able to search a long laundry list of commonly subscribed to databases.
Google Scholar is of course an obvious default choice to use only during a transition. I certainly use the tool myself when first looking up information, but do not rely on it for more advanced searches. I think, however, all federated searches should include searching Google Scholar because so many people are comfortable using it. That way they can see the difference in results.
One suggestion was to use the EBSCOhost search due to the popularity of its databases. Again, a short term “top 10″ type of search.
There are a number commercial items worth looking into in the future, as we can’t afford them now.
WebFeat is the “original federated search.” Although it looks like a good tool and seems to be used among libraries.
Fast Unity may or may not search subscriptions. I’m a little fuzzy on their website as they boast that TVGuide.com uses them and that you can have people advertise…maybe it is not the best tool for a library.
Similarly, Deep Web Technologies boasts that libraries will find it easy to use and implement…which makes me a little weary. Perhaps further investigation will help!
SIFT by Raritan Technologies uses a classification system to organized the search results, using a file based system. It seems to be very customizable and they say their are reasonable priced, although the price is not listed.
Of course a great looking system does not mean it can intergrate easily into the local organization, design, and proxy service. We will also have to revisit how such a tool should be used (as a top 10, by subjects, or an “all search”). 9 work days is not a lot of time…so chances are we might not love what we decide upon. Yet again, maybe we will get lucky!
Jeremy Frumkin said,
February 17, 2009 at 4:28 pm
Take a look at LibraryFind – it’s open source, built on ruby-on-rails, is very UI focused, and can harvest metadata for local indexing in addition to federating out a search.
slederman said,
February 17, 2009 at 5:23 pm
Hello,
I wrote about your post at the Federated Search Blog:
http://federatedsearchblog.com/2009/02/17/two-weeks-to-a-new-federated-search-tool/
I hope you’re willing to respond to the questions I posed for you.
Jacquelyn said,
February 17, 2009 at 7:19 pm
I definitely will do my best as soon as I have some answers for you.
Sebastian Hammer said,
February 19, 2009 at 12:15 am
Hey there, and good luck on the new job. If you have some skills with JavaScript-programming and HTML/CSS, you might be able to throw together a decent proof of concept in 9 days using Pazpar2, the Open Source engine behind most of our metasearch solutions — especially if you don’t mind working late a couple of evenings. The package installs easily on most Linux boxes and includes a very basic interface (charmingly green) that still shows merging, relevance ranking, and facets. The appearance and functionality can be tweaked along the way to fit perfectly into the local organization. It’ll hit anything SRU or Z39.50 compliant out of the box, which includes several major full-text aggregators, and later on, if you get the go-ahead to spend some money, LibLime can help set you up with connectors to just about any service on the web that has a search interface using OpenTranslators. You might be able to get folks on the YAZ list to give you a hand. The best part is, you can invest some sweat and save money on third-party software if that’s what the situation requires. Good luck with an interesting but challenging first assignment.
Larry Donahue said,
February 20, 2009 at 12:47 pm
Good article, and I wish you success in your effort to find what you’re looking for. I’d love to know why you think that our boasting that our technology is easy to use and implement worries you a little (yes, I’m from Deep Web Technologies).
On the ease of use, check out some of our public websites: http://www.biznar.com (for business research) and http://www.mednar.com (for medical research) and judge for yourself. Our technology is also used at http://www.scitopia.org and http://www.science.gov, among other places.
On the ease of implementation, we say that because we do most of the heavy lifting. We’re also working on turnkey implementations, but not quite there yet. Hope to have that in place within the end of Q2, perhaps Q3. Would love input and feedback.
Larry.
Jacquelyn said,
February 20, 2009 at 1:10 pm
I think the boast worried me a bit because there wasn’t documentation to support it that was easily accessible (so is it true or not!). I’ve worked with good systems people to put together mediawiki which is supposed to be easy to implement. To download and put into production right out of the box, yes it wasn’t too hard. But the customization was a huge pain, especially when the instructions didn’t get the result you wanted and you didn’t know why. Also, there were no examples of libraries listed that had used it, just mainly businesses, which made me think there must be a reason libraries (seemed) to not be using it. Looking at the business sites I was able to try out the front end, but it was the back end I was more concerned with.
So thanks for letting me know about the “turnkey”implementation possibility and that you do work hard with the libraries to implement the tool. I guess that didn’t come across to me in your web presentation.
Larry Donahue said,
February 23, 2009 at 3:37 pm
Thanks for the feedback, Jacquelyn. We’ll work on the web presentation.
Larry.