Librarian Central

Editorial Value Meets Algorithmic Search

By R.V. Guha, Software Engineer

Search results on Google.com have always built upon the wisdom of crowds. When Google first launched, we distinguished ourselves from existing engines that were either purely editor-based (hand-built directories such as ODP and Yahoo!) or purely algorithmic (such as Inktomi and AltaVista) by using PageRank to combine the scalability of an algorithmic approach with the value derived from editorial judgment.

Google has also been at the forefront of enabling webmasters to provide customized search for their users. In the past, some visual aspects of search pages and results could be customized to match the hosting website's look and feel, but there was no way for enthusiasts and experts to use their knowledge to actually improve search result quality.

Custom Search Engine combines the benefits of syndication and the algorithmic application of expert knowledge to enable anyone who's passionate about a topic to improve the search experience.

—R.V. Guha, Software Engineer

Custom Search Engines (CSE) combine these two benefits -- syndication and the algorithmic application of expert knowledge -- to enable anyone who's passionate about a topic to use his or her knowledge to improve the search experience, without having to run web crawlers, search servers and so on -- and to reap the twin rewards of fame and fortune for creating this improved search engine. And best of all, it's free.

What can you do with Google Custom Search Engine? Well, at the most basic level, you can specify a collection of items to search; in others words, you can restrict your searches to only the sites and documents you want to include in your results. Or, you can search over the entire web but prioritize certain specific sites and documents in your results. You can also offer search refinements that make it easier for your users to find the information they're looking for.

Despite all this functionality, the power and scope of a Google Custom Search Engine would be somewhat limited by the time and energy you'd be willing and able to spend refining and adding to it. Fortunately, Custom Search Engines are collaborative; by sharing your engine with certain colleagues or other experts, or by opening them up publicly, you can work with others to build and tune your project. Finally, there are two options for housing your search engine; you can either use the Google-hosted template design and access it on a unique URL that we provide, or you can seamlessly integrate your CSE into your website and show results within your own site, in a format and colors that blend in with your design.

You can do quite a bit already with Custom Search Engine, but we're always looking to improve our products, and these days we're focusing on some enhanced CSE features. Soon our users will enjoy even more control over their search results' ranking and display. For example, we're exploring the possibility of more automated Custom Search Engine creation that would allow search over a pre-existing list of preferred sites -- be it a blogroll, top 10 resources, or any other link list. Going forward, we also hope to address other product limitations -- for instance, the fact that today you can't add to your Custom Search Engine sites or databases that aren't crawled by Google (in other words, you can't create a CSE to act as a metasearch engine for your libraries' subscription databases). So, stay tuned. 

So, there's Custom Search Engine in theory. What about in practice? To give you an idea of what kinds of things you can do with CSE, we've showcased some featured examples. You might be particularly interested in the educators section, with examples like Colorado State University's custom search of its own web pages (see the searchbox on the top right) and the Expanding Your Horizons search engine, which aims to expose "resources for girls, parents and teachers" on the subject of women in the science, technology, engineering and math fields. Here's the search engine that Cornell Law School is using to expose legal research guides. Also, in the past, we blogged on Librarian Central about how Tamas Simon's Digital Reading Search and The Librarian's E-Library Custom Search Engines are great examples from the librarian community. For a much larger selection, check out the Baby Boomer Librarian's recent list of library-related Google Custom Search Engines (including the LISZEN search engine, which searches over almost 700 librarian blogs).

As you can see, the possibilities are relevant, current, powerful and nearly endless. Ready to create your own Custom Search Engine? Let's get started.