At least for me (and perhaps others if the goal is to lure new people) a collection of reasonably well documented example projects (preferably concentrating on one thing only) or source to existing (non-engine-utilizing) game project is worth more than extensive engine or framework (even if it includes examples). I'm usually after one thing at a time and I prefer to make it myself at the lowest level so I know what is going on. In most areas I'm not so fluent I would automatically know how to do things, I first need to see what other people did. Understanding what and why and how, are things one needs to tackle first before even thinking about things like performance and portability. In this context (learning) a library or engine project with 100's of functions or features may be exhausting to crawl through. When I go look at one method, it leads to several new ones, header files to bunch of dlls, dependency injections and interfaces. A project doing one thing or limited set of things would be more useful.

In my books the E in PGDCE should stand for effort and not for engine. Whatever you do, comment it with extra effort and share it. It will strengthen the community and bear fruit in the long term.