What will be nice to have, if you ever consider making this more than a nice show-case sample application, is to break it down in multiple components starting with a 'core' software development kit to build on and lots of small and specific components.
If you can get some community involvement in defining the interfaces then there is a great chance that part of it are going to be used easily either directly or being extended (or even totally different implementation), once the interfaces are defined everybody can go on and implement differently if they see fit but still having the option to easily interchange components that adhere to the same interface.
There are a number of 'core' functionality that you can split in a 'software development kit': exceptions, collections, data streams, serialization... then a number of general purpose services like logging, localization, data access...
Note that if one is going to evolve AE to demonstrate how all of the Progress products work together ... shouldn't the PSDN SDK also include all of those products?
Well, if you're referring to the evaluation version then the AE can be packed with all necessary products in the same way... as evaluation, could be probably used in pre-sales engagements as well for show-case of all those products and how it can be integrated.
I has more referring to the low-end 'core' (framework) components that can be used in day to day development by everyone and the fact that maybe PSC could step in and define some OERA components at interface level although a full implementation wouldn't hurt... but having the model and the interface defined one can start implement something along the line and there could be more implementation that could actually be interoperable, as opposed to many totally different implementations that can be used only as a whole... it's a simple re-usability issue, break that in small packages and define the interfaces with external systems and then we can actually use one implementation or another. And this is going to be easier to manage if there ever will be some sort of community involvement, one can't be involved in all areas from core functionality, data access, user interface, messaging, bpm... better try to break things, define the 'interface' between them and find 'owners' for each component/module.
No, I was referring to the PSDN SDK which some of us pay good money for. The OE licenses it provides are fine except that the .NET components are an eval license limited to something like 90 days. The Sonic license is a bit kludgy ... all the functionality is there, but it is an eval license and thus is missing a number of options which one would use in setting it up for production, so no opportunity to experiement and practice with a real setup. And, so far, that's it. None of the rest of the Progress products.
As you can tell from the OERA OSI effort, I am very much in favor of developing components, but I also very much don't think it should be a case of PSC doing something and then giving it to us, even if there is then some mechanism for feedback. I think we need the feedback from the people who actually need to use this stuff in production systems from the beginning.