RE: feedback on data format for WCAG-EM reports

Hi Shadi, Wilco,

This sounds good to me: the "method" property would report the procedure
employed to select the sample. And, to minimize ontological commitment, it
could well be a property of the WebPageSample class. That is: is the
"method" something exclusive to the randomPageSample? I don't think so:
eventually, other types of samples could be the result of applying some
selection methods. Thus, it is not necessary to attach this property only to
"random" samples.

Regards,

Samuel.

-----Mensaje original-----
De: Shadi Abou-Zahra [mailto:shadi@w3.org] 
Enviado el: jueves, 03 de julio de 2014 11:01
Para: Yod Samuel Martín; 'Wilco Fiers'
CC: 'ERT WG'
Asunto: Re: feedback on data format for WCAG-EM reports

Hi Samuel,

I was wondering if there may be particular properties for structured and
random samples. At least the random sample could have something to describe
"the method used to select the sample". On the other hand, this could be
added to webPageSample directly and used where relevant. For example, to
refer to Step 3 of WCAG-EM for structured samples and to provide a
description or reference for the random samples.

So, I guess we could add a "method" property for the "method used to select
the sample" directly to webPageSample and use the approach that you
describe, which is indeed cleaner in terms of semantics.

Thoughts?

Best,
   Shadi


On 1.7.2014 12:51, Yod Samuel Martín wrote:
> Hi Wilco, Shadi,
>
> Just a comment on this:
>
>>> # structuredSample and randomSample
>>>    - maybe create an abstract "webPageSample" class for the two?
>> Wilco: Makes sense to me. I haven't really looked at subclasses or
> anything like that. I'll add this to the file.
>
> First, I don't think structuredSample and randomSample should be typed 
> with their own classes, as they would rather be mere properties of 
> evaluation (unless more properties are planned to be added to each of 
> them). Thus, structuredSample would not be subclassing webPageSample; 
> instead, webPageSample would be the range of the property 
> structuredSample. It is a similar concept (there would be no 
> difference indeed in the JSON example), but I think this better 
> matches the intended semantics (the same reasoning would apply to
randomSample).
>
> Second, if webPageSample has its own class, I think there should be a 
> wrapper around the list of webPages. According to JSON-LD, this:
> {"structuredSample": [
>      {"uri":   "http://www.example.com/whateverSection", "state": "Plain
> text"},
>      {"uri":   "http://www.example.com/anotherSection", "state": "Plain
> text"}
> ]}
> means that the evaluation has *two* structured samples (one which is 
> the webPage at .../whateverSection and another at .../anotherSection).
>
> I guess the intended meaning is better represented by something like this:
> {"structuredSample":
>      { "webPage": [
>          {"uri":   "http://www.example.com/whateverSection", "state":
"Plain
> text"},
>          {"uri":   "http://www.example.com/anotherSection", "state":
"Plain
> text"}
>      ]}
> }
> which means that the structuredSample has two web pages instead. This 
> allows, in addition, attaching other properties to the structuredSample.
>
> Regards,
>
> Samuel.
>
>

--
Shadi Abou-Zahra - http://www.w3.org/People/shadi/ Activity Lead, W3C/WAI
International Program Office Evaluation and Repair Tools Working Group (ERT
WG) Research and Development Working Group (RDWG)

Received on Thursday, 3 July 2014 16:26:59 UTC