Re: FYI: "Test the Web Forward"

We still have to schedule when you will be around the IG F2F in Shenzhen:-) But yes, we should probably discuss this.

Ivan

On Oct 31, 2013, at 12:03 , Robin Berjon <robin@w3.org> wrote:

> Hi Ivan,
> 
> On 31/10/2013 11:33 , Ivan Herman wrote:
>> one minor (or not?) technical issue that came up at a discussion in
>> Frankfurt is the 'driving' of the tests. I have not looked at the
>> test suite right now, but I presume the tests are aimed at being run
>> from a browser directly. That is of course fine for magazine and
>> journal publishers, for example. But for book publishers (ie,
>> readers) it may be a bit more complicated insofar as the natural
>> 'environment' is an ebook reader rather than a browser, ie, what they
>> consume is a book and not an html page. I am not sure what the
>> smoothest way of interfacing the test suite would be then...
> 
> That is absolutely not minor and very much spot on. The test suite is at:
> 
>    https://github.com/w3c/web-platform-tests/
> 
> It will of course always be a work in progress, but as things stand, even though it does have a lot of very good content, it still has quite a few rough edges.
> 
> One of those rough edges is how to run tests. Even in the "simple" browser case, that's not yet anywhere near as easy as we'd like it to be. We're working on making the harness much easier to use. (James Graham has some automation to show there.)
> 
> Hopefully, there is enough metadata in the tests (or rather, extracted from the conventions in use) that we could also generate epubs from it. There will certainly be some issues (I don't even know what the document.location is supposed to be inside a book, tests that check aspects related to origin and location might be either problematic or currently meaningless) but those can be worked through. That could be a topic for a Publishing TestTWF (and a further incentive for convergence :).
> 
> I'd be curious to see the results of the HTML parsing tests. I strongly suspect that most readers (at least the major ones) already parse (everything) as HTML instead of XHTML.
> 
> Another aspect that is also nascent is the use of WebDriver in order to automate tests, notably manual tests (since not everything can be directly automated). I am not aware of any reader currently able to be automated using WebDriver, but it would certainly be very useful for QA.
> 
> -- 
> Robin Berjon - http://berjon.com/ - @robinberjon
> 


----
Ivan Herman, W3C 
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
FOAF: http://www.ivan-herman.net/foaf.rdf