[racket-dev] search (was Roogle?)

From: Matthias Felleisen (matthias at ccs.neu.edu)
Date: Fri Aug 5 15:33:05 EDT 2011

So where does this leave us: 

1. with very little data about real searches, which happen locally, via DrRacket  (would it matter if we could do a Guillaume-style data collection for a few dozen students?)

2. with an understood deficit on our search; I haven't seen anyone deny this

3. a few options on improving search when we're on-line via JS calls to search engines (this clearly requires some experimentation before we commit) 

4a. John's wonderful idea of search the examples in our docs. Well, I imagine he's searching for all one-argument functions, runs them on the example, and produces the list of functions that compute acceptable results. Now, if we hashed these searches, we get a crowd-sourcing problem -- perhaps. The more users/students search for pure functions, the better we get. Perhaps. 

4b. and we know that once text search is incorporated we may wish to look for some form of semantic search. 

Did I overlook something? -- Matthias

Posted on the dev mailing list.