My Photo

Your email address:

Powered by FeedBlitz

April 2018

Sun Mon Tue Wed Thu Fri Sat
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30          
Blog powered by Typepad

Become a Fan

« Getting Big Things Started | Main | Thinking about Jet Lag at 32,000 feet »

May 05, 2006


Feed You can follow this conversation by subscribing to the comment feed for this post.


What does it mean for queries to find queries? I can think of 3 different notions:
1. The queries themselves are identical, for instance by string match.
2. The queries are logically equivalent. (Is this even easy to find for, say, SQL queries?)
3. The queries are different themselves but return the same data.

As for relatedness, consider a Google search query: "andrew johnson president". Is this related to "Andrew Johnson OR president"? The two have different results, but match on some of the terms.

Ray Garcia

A symmetrical dicussion could be made for all data being queries, including data about data. Even a single fact, "what is your birthday", captured in a single column record in a database, gets disassociated from the original query, including the means and context under which that query was presented by the inquirying person of system and the person object under review.

Might the corpus of queries have it's own ontology? When thought of this way the discussion may be one of knowledge representation and its inference instead of data and queries.

Dan Linstedt

Hi Jeff,

Intersting thoughts. I would tend to speculate that queries are in fact, unstructured data. Then, to follow: being unstructured data - they can be mined. But also, the information in the queries need to be stemmed, stop-worded, run through natural language ontologies, and so on.

Once persisting the data about the query, and data about what structures were hit - new correlations can be assessed. I'd even go so far as to speculate that the correlative analysis based on ontological study could in fact either "correct" the next query, OR better yet, adapt the model underneath to better meet the needs of the question being asked.

To your point - in other words, help the queries find queries, help the data find the queries. As usual, form and function MUST be glued together.

My two cents anyhow. I'm now blogging on "modeling architecture" and it's impacts at my new site:

Daniel Linstedt

The comments to this entry are closed.