Achieving Integration: Explanation, Steps, Drawbacks

The big key is probably building the Web2.0 tools into the library's web site to get the tagging and metadata that would be read by the search engines. Another step is generating key lists of authoritative content (books, articles, films, music, etc.) that can be provided via RSS feeds to blogs and also get indexed. Federated searching is good but we have to get the user to the search page. One thought I have is that if we can entice users to the library's site, we can then provide them with Google-like searching using federated searching, relevance ranking, etc. that would be more intuitive to the user. The key is getting them there (see my comments below in the drawbacks area.)

Using Zip Codes and IP ranges

I like the idea of somehow getting the user to put in their zip code or something so they can hone in on their local libraries when doing a Google search. Airlines and weather.com let you set up a profile so the results will always hone in on your locale and you can change it for a specific search if you want.

• Open World Cat - there is supposed to be a feature that locates libraries by zip code though I have not seen it.
• Google Scholar, a library can submit their holdings, via SFX or other means, and tie those holdings to a particular IP range. When a patron in that IP range (usually associated with the campus or library, does a search, it will direct them to the library holdings.

One thing I think we have to keep in mind is that different users use different search engines so whatever we develop would need to work across the spectrum.

Registering with Search Engines

• Our project has to underscore the importance of libraries registering with services like Google Scholar though that is search engine specific.
• However, the user still needs to know to go to Google Scholar and not the main search page.

Improved Indexing and Ambient Findability

• Improved Indexing - If indexing was improved and then libraries were able to put their holdings, along with the better indexing (metadata) into search engines, people would be able to find things easier via the web at large.
• Using the idea of Ambient Findability and the ways people search for information.

Increasing Rankings with Linkage

• Relevance is created through increased linkage
• Libraries need to create content that is linkable and stay away from things like dynamic URLs and other systems which discourage linking
• Web 2.0 tools integrated in a meaningful way that allow increased linkage
• Allowing users to link to as many resources as possible within the library's site. * Making things easier to find AND linkable will increase users ability, need, desire to link to the library's resources.

Using Web 2.0 Tools

We may want to also have guidelines for libraries to use Web2.0 tools for their own sites and databases so that once there a user can have a type of search that is global to that library's collections and services without overwhelming the person with complex searching.

• Federated Searching - One search gets you everything the library has and does.
• Mashups - A library could mash up something using their own site search engine against the Web that also matches up results to the guidelines for determining authoritatative resources. Many of those could be programmed for a computer to determine. This would be a value added service.

(MB) I think that we can address this under the idea of increased linkage.

What are the drawbacks to integration?

There are several drawbacks. The first is that if I don't know or think to search a library site for content then I am going to assume whatever my search engine comes up with is what there is. Library web sites don't get to the granularity of the actual content in their collections (physical or electronic) until you go to search the catalog or the databases. All of that is intimidating to the average user which contributes to why they may not consider searching the library site even if they are aware of what a library could offer them. So we are back to how to automatically feed greater detail to the search engines via metadata so that the library shows up on the hit list and the user is guided to a more in depth search once there.

A second drawback is that if every library did that and you couldn't narrow down your search to your locale it would become unwieldy. That has been the beauty of databases like WorldCat--1 (theoretically) entry with a listing libraries that hold it.

The current separation in libraries of searching proprietary electronic databases which may be restricted to the user community for that library and the library catalog are a huge stumbling block to the uninitiated. If we never see many of our library users because they are not standing in front of us, we have to somehow guide them to these rich resources smoothly, seamlessly, and quickly. Many libraries are also adding rich local collections they have digitized which also may not be searchable without entering a separate database with its own search engine. We simply don't have one-stop shopping that is publicly available even if we do have federated searching which not every library has.

We need to work with Search Engine companies to encourage them to add features like zip code limiting, linking to library resources, etc. and creating tools that let the user get library resources higher on the relevancy rankings.

How can this search engine issue be addressed to meet ALA's needs?

I think the solutions are probably the same though the need for profiling would not be there if you are looking for info from ALA or its divisions. (Just in case we want to address that for our sponsor!)

Comments from a UMKC colleague, Laura Gayle Green:

One thought that is indirectly relevant to your topic is the ability to add a Mozilla add-on for those users who use the Firefox browser.

As you know, OCLC has managed to integrate its data into the main search engines (see http://www.worldcatlibraries.org/wcpa/top3mset/c51cddbd5ff69999.html for the “public” display from a Google link), but it’s not weighted very well, IMHO. I did a Google search for sharp cecil and “southern Appalachian” – the and was discarded of course (I knew that, but still like to have good Boolean habits), but the OCLC record did not come up in the first 2 pages (or was it 3) of results (with 50 displayed per page). I modified the search to be oclc sharp cecil and “southern Appalachian” and only then did I pull up this hit. The first search DID pull up hits from Project Muse and Cambridge Online Journals on the first two pages, though. I imagine that was because these words would have occurred several times within the article, raising the relevancy ranking, and therefore retrieving that hit “higher” in the Google listing.

Metadata is the most obvious way to tag pages with appropriate terminology. Is there a standard thesaurus (much like the LCSH, but more detailed) that can be used or developed to apply to library web pages and resources? I’m also imagining the difficulty of a controlled thesaurus for the web and some of the inherent problems. Joe and Jane Public may include tags on their Flickr site, but not quite realize that something similar exists in their library catalog. With the web, it might be necessary to include broader terms in the metadata as well as the appropriate narrower terms to help Joe and Jane Public in their searches – but hopefully without overwhelming the search results.

Looking at the stated goals on the ALA blog, I think some of the issues around this notion boil down to standards. Standards for metadata and record structure. Standards for terminology. And these standards have to include and accommodate much more than textual resources. How do we describe sound? How do we describe images? How do we describe realia? While the existing structures and standards for cataloging work and are constantly under revision, how can we apply this to our databases? Consider what kind of results a federated search engine should retrieve with a search for Wagner’s Der Ring Des Niebelungen or Degas’s sculptures? Are our resources adequately tagged (including Grove Music, Grove Art, Naxos Music Library, RILM, BHA, etc. etc.) to be able to retrieve appropriate results?

What about having an “@YourLibrary” link integrated into the Google Toolbar (http://www.google.com/tools/firefox/toolbar/index.html and I’m sure there’s one for IE, just this is the page I’m pointed to with FF)??? That could be a cool partnership.