Given available resources and release timeline, there will be time for @zmike to focus on only one of these tasks. Here is some background on the potential options:
Settings: The settings infrastructure for Enlightenment is, in a word, bad. There are a nearly infinite amount of settings dialogs full of externally-documented options; options which, in many cases, make almost no sense in the context of the dialog they reside in. This is a huge hassle both for experienced users (eg. me) as well as newcomers, and it is one of the most complained about aspects of the user experience. Rewriting this will be a substantial task, but the result will provide a more sensible user experience for the hundreds of available configuration options.
Gadgets: The gadget subsystem(s) in Enlightenment was organically developed over a period of 10+ years. The API is completely nonsensical, very hostile and unfriendly to new developers, and extremely difficult to use. Gadget-related bugs are the second largest category of reported issues, surpassed only by the jenga-like filemanager subsystem. Rewriting and improving this will allow for redesigning of existing gadgets as well as providing an API which facilitates more rapid integration of new gadgets, potentially even in non-C languages.
The less-voted-for option will become the focus for E22 development provided that a more urgent issue has not arisen.
This poll will terminate at some point after the E20 final release.
I think gadgets should take top priority for two reasons. First, for a DE E still has relatively few native applications and gadgets. Second, and more important, I keep seeing complaints from those who mourn the loss of of this or that from e17. Yes, it can be a pain finding settings at times, but KDE can present similar challenges at times and people manage to live with it.
I vote for gadgets. As former user of KDE, I can live with complicated settings system, but having sane task manager (with multi monitor support), as written in Base Gadgets would be great ;)
Though settings are a pain to get, once set, most of us leave them alone. Gadgets on the other hand have a lot of utility. I miss the old alarm from 19, and adding in an API that would make it so that you could tie it in with the calendar/timeclock to actually get meetings setup and alert would be a boon for those that don't have outlook. The edgar module has made it so that gadgets can be written easier, but there are things I would like to see built-in like the afore mentioned alarm, notes, etc, which would all lend themselves well to being gadgetized.
I was thinking Settings first, since Gadgets are usually "Nice to have" vs Settings being "Need to have".
But if we have an easy to find document explaining the settings that are there so far - the linked document is a great start - I can live with it being initially quirky.
I use very few Gadgets myself, but maybe I would use more if they were easier to create. This gets my vote, even if it only means the API gets easier to use. Maybe it'll even make more devs and users see the potential in E.
I think we should improve the Gadget Settings. Sorry, what I meant to say is there should be a Gadget for Settings. So, in conclusion, if we could add something to the Settings for Gadgets, that'd be great. ... That's there already? Awesome job! Fast work!
I noticed the poll is at a 50% deadlock. Think I might mosey on up and break it.
EDIT: Curse you, the person who simultaneously voted the opposite of me!
Settings or Gadgets, Settings or Gadgets... Yes!
Sorry, I couldn't resist. :-) While the Settings system dearly needs to be cleaned up, I find myself agreeing with Srart and TwoD. Once I've got the settings where I want them, they don't get changed, at least not much or often. On the other paw, cleaning up the Gadgets system would make it easier for devs to write interesting new goodies. (Alas, I'm not much of a C programmer, so I can't help.)