Interfacing mobileOK checker/markup validator

Hi Olivier,

As you may know, mobileOK [1] requires validity and well-formedness for
the markup (both to the page declared doctype and to the XHTML Basic 1.1
DTD).

As a result, the mobileOK checker [2] proceeds to some validation
internally, and the validation results are given in the output of the
tool.

Of course, as you well know, turning the output given by the validating
library into a somewhat usable output for the user is quite a bit
challenging.

Given that you've already gone through the pain of doing so with the
markup validator, it would seem like a waste of effort to implement a
something similar in the mobileok checker.

One fairly obvious solution would be for us to just notice that the page
is invalid, and then put a link to the markup validator results where
the user could then hopefully get a more readable set of results and
some practical guidance on how to fix the identified problems.

Given that the validator has an option to force the doctype with which
the page is validated, it wouldn't be a problem to do the two types of
validations I alluded to.

The main obstacle is that the mobileOK checker has a set of well-defined
HTTP headers (User-Agent, Accept and Accept-Charset [3]), and many
mobile sites rely on these parameters to send an appropriate response
back.

So, just linking to the validator as is wouldn't work in this context;
we would really need an option where the validator would do its
validation as a mobileOK agent.

I don't think this type of profiling exists at this point in the
validator, and I'm not sure if it would be useful for other cases. But
is this something you could consider including in the markup validator?
I would be willing to contribute a patch to that effect if this is
something you think could be added to the validator.

Thanks for any feedback on this,

Dom

1. http://www.w3.org/TR/mobileOK-basic10-tests/
2. http://validator.w3.org/mobile/
3. http://www.w3.org/TR/mobileOK-basic10-tests/#http_request

Received on Thursday, 7 February 2008 16:09:49 UTC