Apparently I am not alone in the short term turnaround of technology solutions. The Federated Search blog has mentioned a couple of other instances and asked me to answer some questions. I’ll do the best I can and may add an update later in the comments.
1.) What solution can you deploy in two week?
Oh the discussions and problems that have come up! To build or modify a federated search when your staff has many responsibilities is kind of pushing it for a 9 business day turnaround. I have very competent staff, the type that, given the time, probably could build one from scratch, but problems always occur. The short answer is that we decided to temporarily go with using EBSCOhost (10 main databases) as our default search while the web team sets up and tests dbWIZ.
2.) What resources will it take to make this happen and do you have these resources?
Having the type of staff who is competent enough to troubleshoot is the number one resources for such a project. We also had to have access to a test server in order to make sure the dbWIZ tool works before it goes live. While testing the tool we found that it wouldn’t return search results. This problem has not been remedied and is why we had to set up the temoprary search with EBSCO (will be up my monday).
To keep in mind with the timeline of implementing a federated search tool, you also have to allow for time to education the library. They need to understand how it functions. So it is not as simple as finding a tool and implementing it but also testing and teaching.
3.) What are you finding in your exploration of open source solutions?
After exploring the tools discussed in the previous blog, along with LibraryFinder, noted by a contributor on the comments, the web team had to take the fruggle way to a decistion. This “two weeks to a federated search” was complicated due to the fact the library has no more money to spend (budget crisis). So pretty much all the ones at cost were immediatly not possible. That left dbWIZ and LibraryFinder. The web team preferred the dbWIZ look and search capabilities over LibraryFinder. But an immediate concern was voiced by one member. He said that freeware may require you to maintain the subscriptions to the databases within the tool as well as within you main database pages. This is the case for dbWIZ. Double duty work is not a big deal if you are only using 5-15 databases as the base of the federated tool. But if you want all the databases available for custom searches you better think about how much time it takes to maintain databases, double that number, and see how much it costs to pay your employee for their time. When you can purchase some federated search tools for 7-8 grand (which probably will manage the subscriptions for you) then you might find it worth the cost.
Therefore, I then had to speak with the head of reference to find out how they really want to use this tool. Is it just to give students who have not had the opprotunity to learn how to use the library properly the ability to access their required number of articles? Is it to develop custom searches for different dispines? Lucky for us, the head of reference said that it is meant for students who just need something, not for graduate and faculty type searches. Knowing this, I can justify the extra work that it will take my staff to implement and maintain the dbWIZ.
4.) How many sources will you have?
At the moment we plan 10 (same as 360 Search) but until we are able to test how long it takes to perform the search, we will not have a final count. If the search with 10 databases is fairly quick, then 10 it is! If it takes a long time, we will probably weed some out.
5.) Do the connectors already exist for these sources? If not, what is your plan for building them?
I do not know for sure…I am guessing that they are not since the tool was implemented just fine but the search results were not showing up. I’m not a super-techie and can’t really explain how to make the searches work. What I can tell you is that this is where things slowed down. So if you have to connect each yourself, again consider how many databases you are connecting compared to how much time it takes your employee(s). It might be worth purchasing a tool.
slederman said,
March 6, 2009 at 9:29 am
Thank you for your responses. I’d be very interested in updates as your “first assignment” progresses.