I want to get this out to people. Certainly want to bring @rgchris's Ren-C scripts up to date with the modern EXE, and get you back hacking on Query again.
As for other people: I'm still aiming for the starting target to be gaming--code golf puzzle challenge stuff. I think it can be very fun, and the Magical Statics point to the kinds of "a-ha" moments that I think the game could get people to.
But even if it's "just a game", it would have been bad to have put it out before having figured out even the things I've done this past week. I'm not sure how many more "foundational reversals" there will be. (I'm swapping the GROUP! and BLOCK! antiforms as we speak, and it's one of those "I'd feel remiss if that had been wrong" things.) It doesn't have to be a finished, I just have to have a certain baseline pinned down.
"A designer knows he has achieved perfection not when there is nothing left to add, but when there is nothing left to take away." -- Antoine de Saint-Exupery.
For PARSE to be usable on non-trivial data sizes, it has to be nativized. I've explained my strategy for that (one combinator per file, keep the usermode version in the same C file as comments so it can be maintained alongside the native, scrape the Rebol version out to make a module and then have ways of swapping or sporadically running the usermode versions in debug). It's been eye-opening to try honing the usermode code first though... so I'll keep doing that.
The "big ticket" items that have to be decided are:
-
Is FUSED! going to happen. I believe it needs to, but it will open a can of worms.
-
Switching @foo to Iterator Dereference. I think we need iterators, and this seems like the right notation for it.
-
Modifying While Iterating In Stackless. The concept of taking locks you never release while in a generator is tough. This stuff is foundational, I need to revisit it and decide what the best answer is.
For whatever reason, those are the things that are on my mind right now.