incubator-jena-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mattmann, Chris A (388J)" <>
Subject Re: Implementing GeoSPARQL
Date Wed, 04 Apr 2012 19:55:20 GMT
Hi Rob,

Thanks for your email! My replies inline below.

On Apr 4, 2012, at 12:23 PM, Rob Battle wrote:

> Chris,
> I am the primary developer of the GeoSPARQL index in Parliament.  The spatial part is
certainly pluggable. We have an interface for the index and have implemented it with a variety
of backends including a persistent RTree from the degree project [1], PostgreSQL w/PostGIS
[2], and an in-memory quad tree using JTS [3].  

Gotcha, OK. 

>  One part that is not pluggable is the query engine for for Parliament (which replaces
the default ARQ query engine).  We optimize the query such that the parts of the query that
can be answered more efficiently by the spatial index are executed first.  We would certainly
consider using SIS if it could be easily integrated and provides better performance than what
we are currently using.  

Cool. I'm not sure SIS will be more efficient than JTS, but SIS is ALv2 licensed, so it will
at least carry with it less
virality and won't cause your own software to be LGPL (or GPL), in a copyleft fashion. And,
there are a budding
number of people here and a positive community of many people who are interested in helping
you solve 
use cases that are interesting and challenging.

> TBH I hadn't heard of it before I saw this thread.

No worries. We're trying to grow a community here at Apache and it's been slow, but extremely
and a great overall experience.

> Unfortunately, I'm not sure if our implementation is ALv2.  Parliament is BSD licensed
and some of the libraries we link to are LGPL.

Depends on the version of LGPL, and specific instances, but what is known is that JTS brings
with it LGPL dependencies and virality,
with respect to licensing. There are different ways you can link against it, and legal "syntactic
sugar", but rather than worry about that, 
we're trying to be an ALv2 "clean" version of similar functionality.

And I (specifically, because we're all individuals here at Apache that have our own "itches"
to scratch) am interested in a real
use case for SIS beyond Quad Tree and our simple GMaps demo...GeoSPARQL sounds like a great
use case!


Chris Mattmann, Ph.D.
Senior Computer Scientist
NASA Jet Propulsion Laboratory Pasadena, CA 91109 USA
Office: 171-266B, Mailstop: 171-246
Adjunct Assistant Professor, Computer Science Department
University of Southern California, Los Angeles, CA 90089 USA

View raw message