Although Google manages to serve up pretty good results most of the time, sometimes it makes sense to give the search engine a hand by limiting the search to only provide results from a particular set of pages, or domains. So in this post I’ll describe a couple of “emergent” or “discovered” custom search engines that are available in tools you might already use.
(Custom search engines provide one way of achieving this, of course – set the limits over which you want results returned from, et voila… But creating custom search engines, as such, is not necessarily something that would occur to most people.)
Let’s start with delicious, the social bookmarking service, in which users bookmark links to delicious, with one or more tags.
Did you know that there are now a range of tools within delicious that let you search over the titles and descriptions of different sets of bookmarks?
If you pick a particular user, the default Search these bookmarks search will just search over the title and description fields of the bookmarks saved by that user. If you further limit the view of the bookmarks to those tagged in a particular way by a particular user, then the Search these bookmarks search will be limited to just those bookmarks. In other words, Search these bookmarks is context sensitive to the user, tag or user’n’tag combination that is currently selected.
(Remember that the full text of the bookmarked pages is not being searched – only the bookmark title and description fields – which is one good reason why it makes sense to fill in a bit of description about every bookmark you make: it makes (re)discovery of links at a future time easier…)
So where else do people create there own resource collections, or resource feeds Google Reader, maybe?
And as it happens, another emergent, “auto-created” custom search engine can be found just there:
The Google Reader search provides a blogsearch facility that lets you limit your search to the content of the RSS feeds you subscribe to in a variety of ways: the content of all your feeds, the content of the items you’ve read, the content of feeds bundled in various folders, and so on.
So for example, you could bundle a set of RSS feeds together in a single folder, and then, as if by magic, you have a custom search engine that searches over just the contents of those feeds.
With Google’s “official” blogsearch tool no longer functioning as such (rather than just indexing feed content – that is, just actual blog posts – it appears to be indexing blog web pages, so you get contaminated results that may only be a “hit” because your query was matched by sidebar content or other blog website fluff), the Google Reader search tool goes back to basics…
…the only problem is, that so far as I can tell, there is no way to subscribe to the results of any of these searches, and there is no published (or community documented) API for the Google Reader search facility… (so if someone can watch the AJAX calls and produce one, I’d be really grateful :-)
(By the by, can you define filters on folders in Google Reader, a bit like iTunes Smart Playlists?)
See also: Search Hubs and Custom Search at ILI2007.
PS if you are looking for an effective blogsearchengine, Icerocket has been grabbing the buzz lately…