I found Ted Neward's post on WS and Corba quite interesting. I tend to agree with most of what's being said by Ted, Don and Steve, except for a few particular points.

First of all, let's start with Steve's comment "Something else that Don doesn't mention about what CORBA (and DCOM and others) got wrong is something that drove Don to help create SOAP in the first place: the need for the same infrastructure at each end of the wire. Not only the same infrastructure, but in some cases, the same heavyweight infrastructure". Now, in principle, I'd agree. However, in all honesty, the entire WS-* stack is becoming so large and changing so rapidly and requiring ever mode infrastructure on each end, that I'd say this seems to be the case less and less now :( Yep, indigo is just that: infrastructure. And let's be honest.... supporting WS-ReliableMessaging, WS-Security, WS-Addressing and all the rest of the boys does require some non-trivially implementable infraestructure.

Now, I don't complain about it (all that much), but it does seem to me (incorrectly, perhaps) that with all the WS-* specs coming out, the whole WS stack is more and more complex, and some of the simplicity of the model seems to be getting lost alongside the way. I agree with Ted, though, that the composability of the overall stack is great, but, doesn't that add its own share of complexity to the mix?

Sigh. Personally, I just gave up trying to keep up with the WS-* specs for now. Too much hassle, no real return on [time] investment on doing so right now. Shortsighted? Perhaps a little bit, but the whole WS-* mess just doesn't seem to be worth it right now, except for a few key specs (and even those seem to be still churning... how long do you guys think you can keep fidgeting with WS-addressing, for example? Come'on, we need something that works NOW, not something perfect that never stays put!).

As for WS-ReliableMessaging? Well... can't say I fully see the point now. If SOAP is just meant to be an application-level protocol... and one that is supposed to be transport-independent, why try to replicate at its level all the good functionality existing transports provide? Isn't one of the biggest points of making it transport-independent being able to take advantage of the benefits each transport provides?

OK, rant over :) I don't even know if this makes any sense...


Tomas Restrepo

Software developer located in Colombia. Sr. PFE at Microsoft.