In keeping with its promise to provide easy programmatic access to most of its data services, Yahoo released YQL, or the Yahoo Query Language. YQL comes with a REST-based interface to an online query processor that lets developers query data from a variety of sources, which are not limited to Yahoo's own properties.
In a recent blog post, Yahoo! Query Language Thoughts, Ryan Barrett, a member of Google's App Engine team, reviews the YQL API, comparing it to several other Web-based data APIs, including Google's own App Engine data store. Barrett notes that:
YQL is similar to a number of modern query languages, including XPath/XQuery, Microsoft's LINQ, Google's GQL, Facebook's FQL, and the query interfaces for Amazon SimpleDB and CouchDB...
It serves primarily as a unified interface to Yahoo's various web properties, such as Flickr, Yahoo! Maps, and Upcoming. The query engine itself is hosted by Yahoo; users make queries by sending HTTP requests to a REST endpoint.
Currently, YQL is not part of a cloud database. It's strictly a hosted service for query processing. On the plus side, this means that YQL isn't limited to a single data source. It's not even limited to Yahoo's own properties. YQL can operate on any third party data source, as long as it's in one of the usual suspect formats: RSS, ATOM, JSON, XML, etc.
Barrett also notes YQL's conceptual similarities to a relational data model:
At a high level, YQL uses the familiar relational database model of tables and rows. However, to map this onto hierarchical data, YQL needs to know which level of the hierarchy represents a row.
All other operations are done in memory, subject to a 30 second processing deadline. This limits those operations to minor post-processing and cosmetic tweaks, which rules out many interesting use cases on even modest size data sets.
What do you think of Yahoo's YQL as a general-purpose Web-based query language?
> <P>What do you think of Yahoo's YQL as a general-purpose > Web-based query language?</p>
The NIH syndrome is usually an indication of disconnection with the real world in particular when you are not the market leader who can press his own quirks on everyone else.
> The NIH syndrome is usually an indication of disconnection > with the real world in particular when you are not the > market leader who can press his own quirks on everyone > else.
It might be argued (now would be a good time), that all the <insert your favorite framework/persister here>QL's is the result of NIH: "No I won't learn how that damn relational model works, it's too mathy, so I'll build some half-assed verbose substitute. Just because."
I have not reviewed the different offerings. Generally vendor specific products like LINQ does not work well with other database technologies (hence why MS is dropping the technology).
Maybe Yahoo wants to stay out of the cloud arena, and just provide a different/better interfaces to the variety of clouds? Amazon, Google, even Intuit is in the cloud game with http://ipp.developer.intuit.com.
Maybe the question should be: Can Yahoo be an honest information broker?
Previous Topic
