[whatwg] Proposal: "Offline-Capable" Meta Tag and API Indicates Application's Ability to Function Without Network Connection

[Brian Blakely]

Hi Ian,

I agree to the extent that nobody knows what works best at this point
(though I could point to some examples of good implementations). UA
implementation would certainly evolve, just as Fullscreen implementations
have been.

My major concern is that, as web developers learn, leverage, and master
various offline technologies, widespread adoption alone could take years,
and that is before developers begin to finesse their UIs.

The main purpose of the proposal is to accelerate the uptake of the offline
Web.  A UA hook for users helps to break us away from "the web is
online-only, forever" and a simple API for a dev will encourage
implementation.

Cheers,
-Brian


On Wed, Jun 6, 2012 at 7:49 PM, Ian Hickson <ian at hixie.ch> wrote:

> On Fri, 27 Jan 2012, Brian Blakely wrote:
> > On Fri, Jan 27, 2012 at 4:50 PM, Ian Hickson <ian at hixie.ch> wrote:
> > > On Fri, 27 Jan 2012, Brian Blakely wrote:
> > > >
> > > > "What kind of app are you considering that needs 700MB at once?"
> > > >
> > > > I'm considering videogames that the user would like to play offline
> > > > (plane flight, subway, etc), as well as massive software packages
> like
> > > > Adobe CS. A good application designer would allow the user to choose
> > > > portions of the app that they would like to cache long-term, but
> suppose
> > > > the user needs the entire thing?  In that case, 700MB could likely
> > > > lowballing by quite a bit.
> > >
> > > I think appcache handles this particular case reasonably well (modulo
> it's
> > > other known limitations, anyway). The caching progress can be easily
> > > reported to the user (either by the UA or the page), so the user can
> know
> > > that they should leave the tab open while it does the update, and yet
> the
> > > page is usable in the meantime.
> >
> > I completely agree Ian, app cache would be the means by which a
> > developer sends their assets to the user's local storage device.
> >
> > This proposal deals chiefly with standardizing the messaging around
> > that. The developer sets up the application to be ready for offline use
> > (via App Cache, localStorage, IndexedDB, cookies, etc), and informs the
> > UA when the user can go off the wire.  The UA then informs the user in a
> > predictable way that will become familiar to them as they continue to
> > use that particular client.
> >
> > Background downloading and other mechanics introduced in this thread
> > enable a native-like app download process that is, again, always the
> > same on the same UA, instead of varying from application to application.
>
> I think we should wait for sites to start showing their own UI for this
> kind of thing -- "ok, I'm now fully cached" -- before we add a mechanism
> for the script to ask the UA to show UI for this. Without the experience
> gained from authors doing it themselves, we don't really have enough
> information about how to design the feature.
>
> --
> Ian Hickson               U+1047E                )\._.,--....,'``.    fL
> http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
> Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
>