The author of Tree Style Tab has commented (in Japanese) on the change to deprecate add-ons making use of XUL. [1]
Roughly what he said might be summarized as follows (sorry if I misunderstood his intention): Tree Style Tab is useful because the add-on changes the behaviors of the tab globally. That way, it can cooperate with other tab-related add-ons whose authors didn't intend to make the add-ons work with TST. Therefore providing the Sidebar API doesn't help because you can't expect add-on authors to write code just to make add-ons work better with TST.
At this moment in time, Firefox relies on XUL for both the Firefox UI and its extensions. Mozilla wants to move away from using XUL. What's the sensible approach for this?
In my opinion they're doing exactly what they should.
Step 1. Offer new extension APIs that are compatible with other browsers and decpreciate XUL.
Step 2. Encourage existing XUL-based extensions to be rewritten using the new APIs.
Step 3. After the majority of well used extensions use the new APIs, rework Firefox to drop XUL and use web technologies for the UI.
Step 4 (Not confirmed, but could happen). After the new Firefox UI is stable, offer lower level APIs to give further control over the UI.
I think a smarter approach would be introduce this new API and only deprecate the portions of XUL that haven't been replicated in the new API. Once you reach parity you do the rest as you suggest.
Depreciating functionality doesn't mean the functionality goes away immediately, it just means it's not recommended to use.
Mozilla isn't stopping people from using XUL by depreciating it, they're suggesting that developers should avoid it, as it'll eventually be taken away. The two extension systems will both be available until a sufficient portion of Firefox extensions have been migrated over to the new APIs.
Or to put it in another way, good news, they're doing what you said!
At this moment in time, Firefox relies on XUL for both the Firefox UI and its extensions. Mozilla wants to move away from using XUL. What's the sensible approach for this?
In my opinion they're doing exactly what they should.
Step 1. Offer new extension APIs that are compatible with other browsers and decpreciate XUL.
Step 2. Encourage existing XUL-based extensions to be rewritten using the new APIs.
Step 3. After the majority of well used extensions use the new APIs, rework Firefox to drop XUL and use web technologies for the UI.
Step 4 (Not confirmed, but could happen). After the new Firefox UI is stable, offer lower level APIs to give further control over the UI.
I can't see a better way of moving away from XUL.