A client might search blogs, and request the first major heading, first large bloc k of text and first photo, if any, of the blog. My claim is that the client would likely get a good ontology of the blog post. My further claim is that bloggers, blog developers and clients would come to informal agreements about what long and short form search of a blog gets one. Today, the web gets more intelligent about moving pages and part of pages closer to active retrieval points, it does this blindly. The graph system needs to support this kind of ad hoc system, the ability to flow data into requests with some flexibility in differences between the schema requested and schema available.
There is no reason sqlite cannot keep hundreds, thousands, tens of thousands of container formats. We need a way of informalizing the choices. More later.
No comments:
Post a Comment