Digging into BOSS
by Mike Gunderloy - Jul. 10, 2008Comments (2)- Related Blog Posts
- Guest Post: Yahoo's Cloud Team Open Sources Traffic Server
- Cloudera Announces Hadoop World, and Hadoop Marches On
- OStatic Buffer Overflow...
- OStatic Buffer Overflow...
- At Hadoop Summit, Yahoo! Announces its Tested Distribution
As we mentioned earlier today, Yahoo! has rolled out Yahoo Search BOSS: the Build your Own Search Service APIs that make much of Yahoo's core content widely available. Though the various offerings involved are not open source, this "open content" is potentially interesting to open source developers working on the web. Here are some details about what's inside.
BOSS, in its current form, consists of two components: the BOSS API and the Boss Mashup Framework. The API itself is a fairly straightforward RESTian web service. It lets you retrieve web, image, and news search results, and spelling suggestions, directly from Yahoo!. By supplying appropriate parameters in the REST URL, you can do things like filter out porn results, limit results to specific types of documents, and (of course) specify search terms and target sites. Results are returned as XML.
The BOSS Mashup Framework is a Python library with a vaguely SQL-like syntax. Using it, you can combine BOSS results with other bits of XML, JSON, or RSS/RDF. In addition to merging results, it can handle sorting, grouping, removing duplicates, and so on. Armed with this Framework (and of course Python skills) you can easily combine Yahoo's search results with just about any other data you can get your hands on.
Although the idea of open search results is exciting - and the Mashup Framework looks potentially interesting on its own - open source advocates will want to approach their use with a bit of caution. Everything involved is licensed under its own particular terms of use, and those have the potential to impose some severe limits on what you can do with this software. For one thing, you're only licensed to use these bits in conjunction with a Yahoo-based search offering, which would seem to forestall any thoughts of using the Framework as a general-purpose mashup tool.
Potential users should also take note of the monetization provisions spelled out in the FAQ. Basically, although the service is free while in beta, those who deploy it will ultimately be required to display Yahoo! text ads. This may prove an aesthetic bar, if not a legal one, that some potential consumers of this service do not wish to pass.
