Here is the first. You have to listen to a few minutes to get the context and background, but it's well worth it–try around minute 42 in the recording:
http://www.dotnetrocks.com/default.aspx?showNum=367

The other one is here. He talks about how eBay and Amazon can acheive high levels of scalability, among other things: http://elegantcode.com/2009/06/05/recording-of-udi-dahan-on-soa-e-van-01-june-2009/

Ah I don't think I've seen that from Udi Dahan, is it a presentation/blog?

I have watched the interview Eric did with Greg from QCon 2007 but I didn't realise the idea of domain events had actually come from there. That's cool!

Greg was meant to do his talk at QCon London earlier in the year but I think he injured his knee the week before so he wasn't able to come – would have liked to see that presentation live.

I saw Greg's recent post about CQS [Young] where he described CQS at the architecture level being about separating commands/queries at the message level rather than method level which I think is quite a nice description.

Thanks for clearing some of that up & for your post where you've gathered all the resources. I've been going through those, still not finished.

Cheers,Mark

It's all about managing user expectations. As long as they can see that something is happening, waiting 3 seconds is okay. Actually, Udi Dahan talked about using the browser, Javascript, and server-produced JSON files to offload the loading the display of new content to the browser rather than doing all of the work server side.

Perhaps I can answer a few of your questions:

Regarding the expectation for global consistency, typically the approach that Liz offered is the best one–simply show the change as "Pending". You can adjust your SLAs so as to ensure near real-time consistency of the most critical elements of the system.

CQS [Young] can easily work for small systems. Just remember that the justification for doing DDD (and DDDD) versus other approaches, such as transaction scripts or anemic domain models is embodied in the complexity of the domain itself. The ability to build massively scalable systems is some respects is almost a byproduct of following the principles outlined in CQS [Young].

Martin Fowler has also written about domain events. When I watched Eric Evans presentation on what he had learned since the book, he was referring directly to Greg’s implementation. Eric interviewed Greg nearly 2 years ago at QCon in 2007 regarding domain events and state transitions and Eric was very intrigued. I think Greg proved to Eric that it does work.

One thing to be aware of is there is some discussion among those that are using and investigating “Command Query Separation” at the architectural level about the name of the pattern because of the confusion with Bertrand Meyer’s version of CQS for individual objects. Greg is currently exploring other names. I’ve got my vote in for ”Command Query Distribution”; we’ll see how that turns out. You may want to refer to it as CQS [Young] for the time being.

Keep the great blog posts coming.

Just wanted to say that I've really enjoyed your book club posts. You've provided just enough detail to peak my interest and get me thinking, and provided the links to the resources to let me research these items further.

Thanks for the great series of posts and keep them up!

-derick.