W3C home > Mailing lists > Public > public-mobileok-checker@w3.org > February 2013

Re: Requesting page with 'mobile' user-agent

From: Jo Rabin <jo@linguafranca.org>
Date: Fri, 8 Feb 2013 13:50:07 +0000
Message-Id: <4C913CAD-ACBD-4284-856B-99797DE3222E@linguafranca.org>
To: public-mobileok-checker@w3.org
The issue that we faced with this was that we could not predict what user agents would be recognised as "mobile" at any particular time. Hence we made up a mobileOK checker user agent string which needs to be recognised by a Web server in order to do a redirect or display a mobile friendly page directly.

The process that a mobileOK checker follows, including the HTTP request headers and the algorithm for following redirects, is described under http://www.w3.org/TR/mobileOK-basic10-tests/#http_request


On 3 Feb 2013, at 19:39, Justin Freshwater <justinfreshwater@gmail.com> wrote:

> I noticed that the mobile checker gives a 0% for ebay.com, yet there is a page m.ebay.com which is eBay's simplest mobile version and gets a score of 57%, and then when I view eBay on my smart phone I get redirected to hp.mobileweb.ebay.com/home.
> But if I am typing in ebay.com into the mobile checker, which received 0% and appears to be the full desktop version, the mobile checker should be GETing the page with a user agent that contains 'mobile' to trigger the different content delivered and/or redirects.
> I acknowledge that this particular example could be just how eBay is handling the user agent string or mobile checking, but I am asking if the page is retrieved in the way I described with only 'mobile' in the user agent.  And if not, I am recommending that that is how it should be done.
> This recommendation is based on what I have read from the W3C's own Device Description Repository working group for using testing against only 'mobile' in the user agent as the most common thread between all mobile device user agent strings.  And secondly, the reason I think you should be able to simply type one address that can deliver a desktop version and check the same address on the checker for the mobile version, is that it goes along with the principles of "one web".
> So does it in fact work this way?
> I would appreciate any comments and hope I incite further discussion.
> - Justin F.
Received on Friday, 8 February 2013 13:50:37 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:21:22 UTC