Sanjiva Weerawarana wrote:
This is not a matter of test cases and fixing things. What we're talking about is a radical rewrite. What you're suggesting is like saying Axis2 should've started with test cases showing where Axis1 is broken.
No, what he's saying is that you start off with requirements and a framework for understanding when you've won. Then you compare what you've already got (Sandesha) to that, and see what the gap looks like. And it is ONLY at that point when you make any decision that looks even remotely like "rebuild from the ground up". (Did we do that for Axis1/2? Well, not really - but we *did* essentially have that process at a human level with all the mail and f2f meetings and architecture sessions where the entire group was invited to participate.)
Now it may well be that you can have a *fairly* good idea that that will be needed even at the outset - but there is a big difference, from a social if not technical level, between making that decision as a result of solid technical analysis where everyone's working together and making that decision by fiat.
If you're really looking for buy-in from the Sandesha community, the way to get is is to FIRST demonstrate exactly what's wrong/missing in Sandesha with good tests and JIRAs. Then dims' point was that *if* incremental changes are possible to start off, let's find them - I don't think anyone's closed to the idea of a ground-up reimplementation, but I think we all need to start on the same page if that's going to happen.
My $0.02. Thanks, --Glen --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
