|
From: Jody G. <jga...@re...> - 2006-10-31 21:47:18
|
Hi Marc, Over the last year you have suggested I look at the oracle sdoapi jars. I have started up my oracle experiment again (unsupported/oracle) and am attempting to set things up in the manner you suggest. Paul and I have been in contact with your friend Xavier (although I missed him at FOSS4G) and he was going to bother the licensing guys on our behalf. Since "oracle/unsupported" does not have to be in the build I am going to depend on the jars directly, and trust the user to download as per oracle click through madness. If we end up liking the experiment we can either stub the jars again or bother oracle in earnest. I have been in contact with some other open source projects and options very from the strict stance we are taking, to committing the drivers into CVS and trusting oracle not to notice, ... oh and of course not supporting oracle. So now for the question about your sdoapi.zip recomendation... Docs: - http://www.oracle.com/technology/software/products/spatial/htdocs/sdoapi_javadoc/overview-summary.html Download: - http://www.oracle.com/technology/software/products/spatial/htdocs/xplatformsoft.html I recall the approach you recommended was to use to: - use the sdoapi to create the requests (on the grounds they may be efficient) - convert the resulting objects to WKT, and parse them into JTS for local spatial operations (or deSTRUCT them straight into JTS again) One thing I am considering is letting GeoTools to use the oracle geometry constructs directly when the new feature model is available (they have some nice OGC curves for example). It would also be constructive to compare this work with the ISO Geometry module that is currently being formed, perhaps we could arrange for a wrapper around the oracle constructs. Cheers, Jody |