Genuine decoupling is a challenge. Some of the bones are there already, but we're likely to only meaningfully reach this goal if we have parallel implementations of two separate editors.

Getting a polished UX is going to require a level of integration that might not be possible to do in an entirely abstracted way.

There's a reasonable risk that if someone drops in another WYSIWYG editor, elements like link and image insertion won't work as well.

This is an example of how we might use Loomio of feature capture. However, noted this doesn't have voting or way to prioritise as we have been talking about over on core list. Still features couple be proposed, talked over and put to a consensus as to whether they should/shouldn't get worked on.

It seems like it's a good vehicle for arguing over architecture, which happens a lot on the silverstripe-dev list.

@simon15 the current proposition refers to simply focusing effort on TinyMCE 4 rather than the decoupling. (Loomio allows a discussion and then a community proposal for action based on discussion, which in this case after @samminnee1 and I discussed the original decouple idea, the proposal was shifted toward getting agreement to work on 4.x upgrade instead). Still experimenting with best use practices but so far Loomio is nice for getting an idea of people position of these often opinionated matters.

@camfindlay1 This just further exemplifies my point in this thread. It's way too easy to get confused about what you're voting on and what you're discussing.