On Aug 12, 2013, at 7:15 PM, Mark Finkle <[email protected]> wrote:
> > > Defining such an interface and managing the multiple client and server > versions is not awesome, but of course it can be done. And loading a > privileged page is equivalent: the interesting part is populating the > interface (or privileged page) from the Android Account object. > I still reserve judgement on the Android part. Using a GeckoView (doesn't > exist yet) or a WebView (leery of support across various Android versions) > has risks. There is also the "look and feel" issues. We'll need to talk to UX > about how this would "feel" in the Android system. Remember, Android already > has a native way of doing these things, desktop Firefox doesn't really need > to worry about that. Also, making this look like the right "Android" UI could > be challenging too. > > Obviously, getting something to work is first and foremost, but moaking it > look and feel good is following closely. Let's not back ourselves into a > corner. This work should not result in any changes to the Authentication API which would back us into that corner. The ability to go native is preserved. And if with android you guys want to *start* native first, then I don't have a problem with that. > Lastly, I might be too noob to have caught the reasoning, but why are we > intent on using web content? Does it make key generation easier since we are > pushing that to the server? Is it a "we only need to make the UI once" > thought? Cause that's not true: We'd need mobile and tablet(s) versions. I'm intent on starting with web content, because it's something our team needs to do anyway, and I think it's the most efficient way to get us patches landed in elm that let us take a next step in terms of refining UX (seeing and feeling it in product), as well as gives us a implementation in firefox capable of producing key material, which is a pre-requisite to working on storage. To your point, I expect we'll need a myriad of different skins for authentication that occurs against firefox accounts. Part of the motivation for breaking out the web bridging into a different server is so that we can handle all of those, localization, etc in a separate context. Does that make sense? lloyd
_______________________________________________ Sync-dev mailing list [email protected] https://mail.mozilla.org/listinfo/sync-dev
