W3C home > Mailing lists > Public > public-wai-ert@w3.org > September 2007

RE: replacing earl:inferred with aggregation logic (Issue #7a)

From: Carlos Iglesias <carlos.iglesias@fundacionctic.org>
Date: Wed, 5 Sep 2007 13:25:22 +0200
Message-ID: <09700B613C4DD84FA9F2FEA5218828190E7E10@ayalga.fundacionctic.org>
To: "Shadi Abou-Zahra" <shadi@w3.org>
Cc: <public-wai-ert@w3.org>
Hi Shadi,
 
>> [...]
>> IF all sub-tests were carried out in the same mode
>>     the test mode of the combined test is the same.
>
>Good approach.
>
>> ELSE
>>    IF there is at least a mixed (or undisclosed) mode
>>         the test mode of the combined test is mixed
>
>Good point, my proposal did not address "undisclosed". I'm not sure how
>much precedence this value should have in the logic.
 
For me it makes sense. If any of the part is no disclosed you don't have 
the necessary information to say if the whole will be manual, automatic 
or whatever.
 
>>    ELSE
>>        IF there is at least a manual mode
>>             the test mode is manual
>
>This means that "automatic" + "manual" = "manual" ...bug or feature? ;)
 
Yes, I think this is the tricky one

>>        ELSE
>>            IF there is at least a semiautomatic mode
>>                the test mode is semiautomatic
>>            ELSE
>>                IF there is at least a automatic mode
>>                    the test mode is automatic
>
>I think this nesting of IF...ELSE is rather confusing, I hope we can
>reduce the rules to a small set that is easy to overview.
 
Yes, sure. It was intended just for "internal use" and to help to see the
logic.
 
> Here is
>another attempt of the rules:
>
>#1. if all sub-tests were carried out in the same test mode, then the
>test mode of the combined test is same.
 
OK

>#2. if the test mode of any of the sub-tests is "undisclosed", then the
>test mode of the combined test is also "undisclosed".
 
OK

>#3. for all other test mode combinations of sub-tests, the test mode of
>the combined test is "semi-automatic".
 
I think this doesn't work with the current definition of semi-automatic, 
specially with the "...a software tool was primarily responsible..." 
 
If I have an automatic one and a manual one, why the result of the 
combination must be mainly guided by a software tool?

>Does this cover all the combinations? Is the outcome what we want
>(especially the strong effect of "undisclosed" and "semi-automatic")?
 
I agree with the undisclosed part but not with the semi-automatic one, 
need to think more about it.
 
Regards, 
 CI.
Received on Wednesday, 5 September 2007 11:25:16 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:18:28 GMT