It's not an unreasonable starting point, if for no other reason than
there's been little effort to preserve theme compatibility across
releases, or to build the UI in ways that minimize risk of theme breakage.
> To gather some data regarding this issue, I visited the Most Popular
> Themes page on AMO [...]
A few problems and nitpicks:
* Nightly has only had 12 days of development since the last Aurora
uplift, so you're effectively missing most of the 6-week development
cycle. (i.e., Nightly and Aurora are effectively the same still).
* All but 3 (BlueFox, OfficeBlack, Smallring) are actively maintained
and explicitly compatible with Firefox 10 (which, as I just noted, is
basically the same as Firefox 11 right now). Testing such addons doesn't
say much about the risks of compatible-by-default.
* As we've found with many parts of the Mozilla ecosystem (including
extension compatibility), the long tail matters too. I'd have to check,
but I'd be surprised if we've shipped a release since Firefox 3 that
didn't have the Top-10 addons all compatible on release day. Recent
Firefox releases have shipped with 97+% of AMO addons compatible on
release day. So clearly that's not enough. :)
> I didn't test every
> single aspect of the themes, nor was this meant to be some sort of
> comprehensive or representative test, but hopefully someone who has
> the data or the means to obtain them will chime in.
Understandable, but that's an important part of determining if they're
really compatible. Primary UI for the main window is certainly
important, but so are functional video controls, door hangers, menus,
prompts, other windows (Places, Downloads, Dev Tools), etc etc etc.
> Compat-by-default addons will solve a major gripe that I have seen
> expressed by users on tech sites: Firefox updates break addons. This
> grouse is unlikely to go away, and will possibly increase in
> magnitude due to frustration from dashed expectations, if #693901
> does not apply to themes.
I don't agree. All the grumbling about addons that I've seen has been
about extensions, not themes. Which isn't surprising, given that
something like ~80% of users have 1 or more extensions, whereas < 2%
have a theme installed.
Justin