Well, thank you for your input everyone. I have put together some
conclusions. Please let me know if you disagree or have something to add. I
also have some questions towards the end, for those brave enough to read
the whole thing and still be wanting more.
Developers of FxOS applications should develop with 1.1 in mind for now.
Mozilla is planning to make the Flame be the reference device, but this
will only be relevant in the mid/long term, once there are more 1.3 devices
available to the general public.
Interestingly, Mozilla is currently encouraging developers to use 1.2+
simulators, as well as practises (chiefly flexboxes) that don't work well
in earlier versions and whose support cannot always be detected at runtime.
This means that a number of newly arrived developers are going to develop
new apps without being aware of these downsides.
These developers can be divided into these categories:
A) Those who have 1.0.1 or 1.1 devices. They will try their apps on their
own devices and realise that something is amiss
B.1) Those who have 1.2+ devices: they may create apps that don't work on
old devices, and they may be unaware of this problem
B.2) Those who don't actually have a FxOS device and just use the
simulators. (Do these actually exist...?)
Those under (A) may be upset if they have been using a 1.2+ simulator and
suddenly find they can't use their own app on their own devices. At least
they'll realise the problem in time and fix it (hopefully).
Those under (B) will only become aware of the problem at the Marketplace
review stage. However, if the review isn't thorough enough, problems could
be missed, hitting end users.
OS upgrades are not the solution. At the moment they are not automatic,
require using adb in the best of cases, and even then they doesn't always
work. I'm a tech-savy individual and I couldn't upgrade my ZTE Open to 1.2
following the instructions (although now I think I should stick to 1.1
anyway). This means we definitely cannot expect end users to upgrade their
phones at all.
(Speaking of which: LG is evil. Its case serves as evidence that Mozilla
shouldn't be so confident about devices always being upgradeable/hackable:
hardware vendors will always ultimately do whatever they want, regardless
of their users' best interest).
But there's still hope. I haven't been through the Marketplace submission
process yet, so I can only speak from what I read. The Marketplace review
criteria [1] state that, apart from checking the manifest, a reviewer will
use the app for a few minutes. Also, Lisa tells us on this thread that 1.1
is being used for reviewing.
My perception is that the apps available on the marketplace right now are
not terribly complex. If this is true, reviewers should be able to catch
these problems. Also, the criteria say that reviewers may be able to point
developers in the right direction when problems are found. This could
potentially help a lot. I wonder if the Marketplace team have data
available on how this is working out so far.
I wonder how much of a problem the current situation is after all. For
example, how many owners of FxOS devices are not developers? FxOS phones
lack some very popular apps (eg: WhatsApp). Why would these people choose
theses devices in the first place? Do they actually care about apps? Maybe
they are just content with basic phone functionality and browsing.
If this is the case, then all this is still not a problem. It will become a
problem in the future, as new web technologies emerge that are not
supported by 1.3 or whatever comes later, but we'll still be in a better
place. Mozilla's strategy of starting FxOS by kindling the fire with a
couple sub-standar versions will have paid off.
[1]
https://developer.mozilla.org/en-US/Marketplace/Submission/Marketplace_review_criteria
Pablo Brasero Moreno
pa...@pablobm.com