On Mar 25, 2009, at 11:20 AM, Seaborne, Andy wrote: > I also feel nervous that the parameters feature is being driven by > optimization concerns and an appeal to the analogy with SQL. I'd > feel more comfortable if it were driven by some application-writer > added-value; that is, an external advantage, not an implementers and > tuning advantage. So far, such a use case has not come to light. I added parameter support to my implementation (API, *not* protocol) simply because I found it made writing applications easier. That being said, I think the biggest value here is at the client API level, and wouldn't be troubled if it wasn't standardized in the protocol. Optimization seems compelling, but I would tend to agree with Andy's reasoning here. gregReceived on Wednesday, 25 March 2009 17:06:39 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 1 October 2009 14:42:11 GMT