W3C home > Mailing lists > Public > public-wai-ert@w3.org > December 2006

Re: [ACTION] Warnings in EARL

From: David Rooks <drooks@segala.com>
Date: Wed, 06 Dec 2006 10:19:58 +0000
Message-ID: <4576994E.9040605@segala.com>
To: "Paul Walsh, Segala" <paulwalsh@segala.com>
Cc: 'Shadi Abou-Zahra' <shadi@w3.org>, public-wai-ert@w3.org

I agree with you both that the mobileOK document defines a WARN as being 
a result. However, we discussed this very thing in the MWBP WG a couple 
of weeks ago. As far as i know, a definitive resolution was not taken 
but the group does appear to be leaning towards the idea that a WARN is 
not a result but additional information attached to a result. Hopefully 
this will be cleared up in the next revision of the mobileOK document.

http://lists.w3.org/Archives/Member/member-bpwg/2006Nov/0128.html
http://lists.w3.org/Archives/Member/member-bpwg/2006Nov/0129.html
http://lists.w3.org/Archives/Member/member-bpwg/2006Nov/0132.html
http://lists.w3.org/Archives/Member/member-bpwg/2006Nov/0138.html
http://lists.w3.org/Archives/Member/member-bpwg/2006Nov/0139.html

And, for the record, i like Carlos' solution. It is simple, effective 
and provides (in IMO) a much needed WARNING class.  I am opposed to us 
(i.e. ERT WG) going down the subclassing route as i do not believe there 
is a fully definable set of subclasses that need to be implemented,  and 
with Carlos' solution i do not see a need for subclassing of results.

I can not envision a scenario where a warning (or at least, a useful 
warning) is anything but plain text. What is the point of issuing a 
warning without an explanation?

Regards,
David.

Paul Walsh, Segala wrote:
> -----Original Message-----
> From: public-wai-ert-request@w3.org [mailto:public-wai-ert-request@w3.org]
> On Behalf Of Shadi Abou-Zahra
>
> And as to mobileOK, the WARN seems to be a result just like a PASS or 
> FAIL not an additional flag as proposed by CarlosI (however, it seems to 
> be a kind of PASS too). They also define exactly when the WARN result 
> should be issued by using pseudo code for each test. In WCAG we don't 
> have a clear definition of when warnings should be issued so this may 
> lead to tool developers misusing warning results to satisfy the users 
> who don't like to see errors.
>
> [PW] According to the last conversation re mobileOK (or at least from what I
> can remember, David Rooks?), a WARN would be a result just like PASS or
> FAIL, not an additional flag.
>
>   
>>> It seems we still need a compelling example of a real warning in the 
>>> context of EARL. We often talked about warning to describe situations 
>>> such as "nearly-passed" or "could-do-better".
>>>       
>> What is "nearly-passed"? A cannotTell?
>>     
>
> That is exactly the problem! ;) ..."nearly-passed" could just as well be 
> a fail. For example, "if you just close that one tag your document would 
> validate but right now it is invalid" result.
>
> [PW] Huh? Nearly pass = fail, i.e. it didn't pass. Sometimes you need to be
> black and white and move on.
>
> Kind regards,
> Paul 
>
>
>
>
>
>   
Received on Wednesday, 6 December 2006 10:20:15 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:18:27 GMT