On Sat, 15 Oct 2011, Jay Hargis wrote:
If you have a lot of dynamically generated queries, the application will grow beyond trivial and you can spare the extra cycles both figuratively and literally you'd be happy with SA in the end. It's an amazing tool. I personally have rarely...perhaps never, benefited from the added bonus of being semi database agnostic as a result of using a database abstraction. Switching databases almost always includes a large rewrite. So in my experience, that consideration hasn't provided much for me. But it is there.
Jay, Database abstraction is not an issue. This model will use postgres because when it is used by my client and the regulator, they'll want multiple users to be able to access it. Rather than using a different database, my concern is the ease of having user-selected criteria delivered to the postgres back end, and the results returned for use in external modeling (specifically, GRASS and R; how I'll programmatically exchange data with those, and display the results from them, are issues to be addressed a few months from now). You and Lawrence write convincingly of the potential advantages of my learning and using SA. I'll take your advice. Thanks, Rich _______________________________________________ Portland mailing list [email protected] http://mail.python.org/mailman/listinfo/portland
