Joshua Smith <mailto:[email protected]>
2016 January 29 at 06:12
Our customers want “an application”. It’s largely a perception thing,
tbqh. It’s okay that it’s actually a browser under the covers, as long
as you can’t tell there is a browser under the covers.
The must-have features for us are:
1. Ability to install something on the desktop and/or start menu
(windows) and to the dock (OS X). (It could be a shortcut if it had a
customizable icon.)
2. No browser chrome, resizable window (with no file/edit menu on
Windows!), full-screen capable (at the O/S layer, not the HTML layer)
These are features that we've talked about adding to Firefox for
Progressive Web Apps with a Web Manifest, although I don't know of any
plans to do so at the moment.
3. Support for mailto: links, and for opening the user’s default
browser (OWA can’t do the latter)
FWIW, OWAs should actually be able to do that, per
https://bugzilla.mozilla.org/show_bug.cgi?id=847518.
4. Persistent local storage
Nice-to-have features:
1. Exemption from any service worker cache eviction policies or size
limits (this is more of a must-have, tbqh)
I suspect (more) persistence of local storage and exemption from service
worker cache eviction are two sides of the same coin. And the exemption
from size limits is something that I'd expect you'd want for local
storage too. They also seem doable for Progressive Web Apps, although
again I don't know of any plans.
2. Ability to install without having to put it in a “store” (chrome
lacks this)
3. Ability in the browser to tell if it’s already installed (chrome
lacks this)
4. Ability to launch the app from the browser
Distributed "installation" is inherent to Progressive Web Apps, which
can be accessed from any site, although they aren't installed the same
way as Chrome Desktop Apps and OWAs, since their installation/update
model is more web-like (automatic, incremental, and silent).
Nevertheless, Service Workers is intended to make it possible to
implement more package app-like models, and these are all features that
it would make sense to provide for such apps.
Also, as Windows and OS X are adding support for custom URL schemes as
a way to launch apps and pass them a relative URL, I’d love for that
to be supported. We use this on iOS and Android, and it is incredibly
powerful.
I'm not familiar with this work, but it sounds intriguing!
FWIW, Web Manifest makes it possible for us to implement a subset of
these features in Firefox, albeit with shallower native integration
(f.e. no native uninstall, and perhaps the app would run in the same
process as Firefox, although it would run in a dedicated window).
There aren't yet firm plans to implement such features, but
understanding your use case better might help us along that process!
I’ll be perfectly honest with you: This sudden decision to pull
support for OWA desktop apps because “nobody wants to work on it”
makes me very hesitant to deploy anything using Firefox at this point.
That’s not how a business should be run. You have millions of
installed apps that are suddenly going to stop working in a future
release, many of which I suspect people /paid for/ in your web
storefront. If using Firefox implies that I have to dedicate a FTE to
making sure the features we need in Firefox are maintained, that makes
Firefox unaffordable. If Mozilla wants to be taken seriously in the
enterprise, they need to fix this approach.
I hear you, and I do understand how this change impacts Mozilla's
credibility (even though it's the most responsible thing to do, given
the organization's decision not to prioritize the runtimes).
But note that the decision to disable the runtimes is my own (as the
module owner for the Desktop Runtime module, and the relevant peer for
Fennec), not Mozilla's. Although I'm a Mozilla employee, in this case
I'm acting in my capacity as a member of the Mozilla community, not an
employee of the organization.
-myk
_______________________________________________
dev-webapps mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-webapps
