Following some discussion on the PEG, Bruce Gruenbaum and I have been working on an updated version of the collection classes I originally published in 2006. The interaction has lead to more of a ground up re-design than a rewrite with a strong initial focus on the kinds of collection classes specifically needed for modeling one to many relationships between objects. I have now published the pre-implementation documentation for this work prior to doing the final coding so that there is an opportunity for community input before I have every thing already done. Please check out http://www.oehive.org/CollectionClassesForOORelationships and leave any comments or contact me directly if more interaction is needed.
While I'm quite interested in this community initiative, I'm still not convinced about a collection library that is:
1) custom made (for performance and integration reason);
2) without generics;
3) without advanced query mechanism
Has anything been revealed regarding these points in a near future version ?
As to performance, please read the child page about what we will be testing and add any requirements of your own. Note, of course, the strong focus on collections for implementing object to object relationships. There are clearly other needs, such as those in the Java hierarchy, which will get addressed later. Yes, I'm concerned about performance and it would be nice to get some help from the language, but preliminary testing suggests we may be OK.
No, as far as I know, real generic support is still not on a firm time-table. I really missed it in the 2006 implementation ... it even drove me to use include files and preprocessors! That's desparate! But, this time around I don't know. Given the focus on relations, really we are only looking at collections of objects so, in the iSet family (see UML child page and Properties and Operations child page) there is really no need for any type except Progress.Lang.Object. So, the only place where generics would really apply is the key in the iAttrSet family. This time around, I think I can solve that without generics and with the nonsense I did before. We'll see if I'm right, but if I am, then the generics may not be so important here.
I know you are fond of the advanced query thing, but I wouldn't hold your breath. In any case, it is really outside the scope of the present work because it doesn't apply to the kind of basic object to object relationships which this is intended to support.
If anyone is interested, there are some interesting comments being made on the OE Hive pages for this topic.