Re: test against Gmail

mobileOK is largely orthogonal to questions of "one web" and content
adaptation, and by design. As Jo says, mobileOK says nothing about
desktop content and doesn't care how you decided to cough up your
mobile content.

However, the mobileOK test does contain a subtle assumption that it is
being applied to a mobile-specific URL -- e.g. m.gmail.com and not
gmail.com -- and this is why it says you should default to delivering
mobile content. This in turn has some implication for how you do
content negotiation / adaptation between desktop and mobile content.

>From my time galloping across the mobile web, I believe the large
majority of sites do use a different set of URLs for mobile content
(and possibly the desktop property redirects to the mobile property),
and, it's these sorts of sites and URLs that mobileOK is targeting.
This model, by design, fits neatly with mobileOK.

It's also entirely legitimate to serve desktop and mobile content from
the same URLs. I can understand that the right default for such sites
is often desktop content, which seems to be a problem in theory
vis-a-vis mobileOK.

In practice, there is no mobileOK test that tries to determine what
your default behavior is. The tests always use an Accept header that
asks for mobile content, and a fixed User-Agent. Regardless of content
negotiation / adaptation strategy, it should be possible to determine
that the test requests want mobile content. So, the note about
defaulting to mobile content may not present a problem in practice.

As a result, I'd also support removing this sentence, as it seems
controversial and does not map to a real test anyway.

Sean


On 12/4/06, Rotan Hanrahan <rotan.hanrahan@mobileaware.com> wrote:
>
> It's a tough call for a site administrator. You may have an excellent
> adapting Web site, capable of delivering excellent presentations to a
> wide variety of devices, especially the diverse Web-enabled mobile
> devices. However, you may also know (perhaps from the nature of the
> content/service you provide) that the majority of users will not be
> using mobile devices and therefore you choose to default to non-mobile
> content whenever you receive a request from an unknown device.
>
> This is legitimate, but despite your ability to provide excellent
> content for known mobile devices (perhaps in all cases adhering to the
> underlying aims of mobileOK) your default action falls foul of the
> mobileOK necessity to provide mobile content to unknown devices.
>
> There may come a time when the media people start a tvOK initiative for
> Web-enabled TV. And we may even see a gameOK initiative for Web-enabled
> game consoles. The adapting solutions will happily support all of these
> and more.
>
> But when it comes to the default response, all these initiates will be
> in conflict. It will not be possible to be xyzOK in default
> circumstances.
>
> I know that the adaptation technologies my company provides, and those
> of our partners, associates and (of course) competitors, can deliver
> mobileOK *if* the operator of the service chooses to default in that
> way, but experience shows that most providers of "one-URL" strategies
> still prefer to default to old fashioned PC browser content. That may
> change as usage patterns shift.
>
> Perhaps one day there will be no need for the default behaviour, because
> the request protocols will demand some form of disambiguity (perhaps
> even to the point of tidying up the Accept header once and for all) but
> I suspect my hair will be quite grey by then...
>
> Meanwhile, should we be considering a defaultOK initiative? :)
>
> ---Rotan.

Received on Monday, 4 December 2006 17:40:44 UTC