Web Devout tidings


Web Devout search engine

Web Devout now has a native search engine to help users explore the site.

The search engine uses a link weighting algorithm similar to Google’s PageRank, as well as over 10 other factors to help get the most relevant results at the top. I made the engine from scratch for this site, and it currently only accepts simple queries (most common operators aren’t yet supported), but I will continue working to improve it in the future.

Update: Standard search operators such as AND, OR, |, and & are now supported, as well as double quotes for term grouping and parentheses for query grouping.

4 Responses to “Web Devout search engine”

  1. Jeff Schiller Says:

    Interesting. But why not just use a Google search query with site:webdevout.net ?

    Posted using Opera 9.02 on Windows.

  2. David Hammond Says:

    The Google search results wouldn’t be as up-to-date or integrated as the native search engine can be. Plus, there are a number of features I’m currently adding to the search engine that aren’t supported by Google.

    One example is “favor terms”. By prefixing a term with a question mark (?), you can increase the ranking of results that contain that term without getting rid of results that don’t. This can be useful if you think a page might have that term in it but you aren’t sure.

    I also plan to use the search index for other purposes elsewhere in the site, such as in the visitor statistics page and perhaps 404 pages.

    Posted using Mozilla Firefox 2.0 on Windows.

  3. Asher Gabara Says:

    The search engine is good, but what is query grouping?

    Posted using Opera 9.10 on Windows.

  4. David Hammond Says:

    Query grouping is the organization of the search query into subpatterns. For example, a (b|c) (d|(e f)), where a through f are various search terms. Basically, it allows you to get really picky or flexible with what you’re looking for.

    Posted using Mozilla Firefox 2.0 on Linux.