- From: Olaf Hartig <ohartig@uwaterloo.ca>
- Date: Tue, 21 May 2013 13:48:46 -0400
- To: <public-lod@w3.org>
- CC: Michael Brunnbauer <brunni@netestate.de>
Hi Michael, On Tuesday 21 May 2013 19:12:17 Michael Brunnbauer wrote: > Hello Olaf, > > very interesting! SQUIN may be useful for us in the future. Cool! > A "maximum query execution time" feature would be nice - so partial > results (according to your query semantics) can be returned to impatient > humans. Do you plan such a feature ? Yes, I'm planing to add such a feature. The reason why I have not done this so far is that SQUIN currently implements the LTBQE paradigm based on a fixed pipeline of iterators, which is a very inflexible implementation approach. I'm currently working on a much more flexible alternative. Once this is done, I can start thinking about support for different optimization objectives (e.g., maximizing the number of solutions computed in a given amount of time). Notice, what SQUIN already does is "streaming output", that is, reporting solutions as soon as they have been discovered (however, this feature is currently available only for simple BGP queries without solution modifiers such as ORDER BY or DISTINCT). > Is it possible to supply additional seed URIs for a query (e.G. via the > API)? It is: If you use the command line interface you can simply add additional seed URIs via the parameter --lookup If you use the API it is a bit more complicated. The easiest approach would be to simply copy the code that implements the aforementioned --lookup parameter. You find that code in class 'org.squin.command.modules.ModLDCache'; see in particular methods 'getLDCache' and 'ensureAvailabilityFinished' If you have any issues or further questions, let me know. Best, Olaf
Received on Tuesday, 21 May 2013 17:49:21 UTC