I've not seen any web UI frameworks that in my estimation compare to what we use at the moment.
I'm not sure I understood what you mean. The web is miles ahead of the desktop in terms of styling and widgets. There's no desktop equivalent for CSS, and although there are many desktop UI widgets, they're hard to integrate if they come from different sources, and none of them are as extensible as their javascript-based counterparts.
The only reason we don't have a browser-based photoshop or zBrush is performance, not features, and the C2 IDE is mostly static, so such concerns should be no issue (especially since javascript text-based IDEs are feasible, as proven by brackets).
Whatever we did, a radical technology shift or reworking of the editor (at least to support Mac + Linux) would take a considerable amount of time, pushing things back a long way while we catch up with what the editor already has. (We're too deeply tied in to Windows and MFC at the moment for it to be practical to port what we have already.)
There are already tradeoffs with continuing the C++ editor, and those will only get worse with time. We are already hitting a ceiling on what the editor can do in terms of modularity and extensibility - I'm sure you agree an IDE SDK is just as much a pipe-dream than a HTML5 IDE.
Linux+Mac support are two other things that would take a ridiculous amount of time.
The longer we wait with the current editor, the worse the cost becomes. Besides, we've been sitting on the bleeding edge for a long time, it will be nice to let the wrappers/browsers/devices catch up a bit and get better organization features, which are currently preventing big projects from becoming a thing with construct.
Skinning was a major feature that ended up limited to recoloring/resizing only (imagine the power of CSS based reskinning).Translation practically flopped. There are many widely-used translation APIs for javascript, and they're mostly plug-and-play.
To what extent are those wanting editor enhancements willing to wait for this to happen?
I'd rather see modularity first, before a huge push for IDE enhancements, but after that I don't think there are many major runtime features left (the biggest one was multiplayer and that's already in).
We could keep extending the C++ editor, which is relatively painful, and further invests ourselves in a Windows-only English-only editor.
I don't think you're too fond of that idea yourself. The mac+linux versions might be incentive enough for you to port, and those would probably represent an extra 15% of the total income (though I imagine Tom would have a better idea).
A translated editor would also go a long way towards acquiring new customers, so I thing it's a good idea from a business sense.
Or we could think about reworking the editor sooner rather than later, which in the long term could bring a lot of wishlist features but would probably be on the order of at least a year or more, during which updates to the current editor and runtime would probably be sparse. And we'd probably call that "Construct 3" given the vast scale and cost of the reworking.
We've done it before with the move from Construct Classic to Construct 2, and that took time but I think it paid off amazingly well.
IMHO those make sense both from a business and engineering perspectives. For me at least it's clear the engine has outgrown the IDE.
This time around we could keep the same runtime which makes it less work. Thoughts?
You'd also keep the exporter, the XML project structure, the runtime debugger and the runtime SDK, as well as most of the eventing scheme (one could call it the "construct standard library"). Extensions would be easy to convert if you keep the runtime SDK. You could also make a tool to convert c2 projects to c3 projects.
As for construct 2, I've been extremely satisfied with it, and I think I got more than my money's worth out of it. I'm not sure anyone in the community can disagree, especially considering the free (and frequent) updates we've been getting all this time.
I'm in favor of a Construct 3. Why don't you post a poll discussing it?