Re: correct RDF Re: Locating In EARL Example

>> <rdf:li rdf:parseType="Resource">
>> <earl:line>9</earl:line>
>> <earl:href>spending.html</earl:href>
>> <earl:xpath>/html/body/p/a</earl:xpath>
>> <earl:name>anchor</earl:name>
>
> earl:name won't work here:
>
What I'm trying to say, using the earl:name element, is that the items 
within the rdf:li element are associated and can be identified by the 
earl:name element.

>   <rdf:Property 
> rdf:about="http://www.w3.org/WAI/ER/EARL/nmg-strawman#name"> 
> <rdfs:label xml:lang="en">name</rdfs:label>
>     <rdfs:range 
> rdf:resource="http://www.w3.org/2000/01/rdf-schema#Literal"/>
>     <rdfs:domain 
> rdf:resource="http://www.w3.org/WAI/ER/EARL/nmg-strawman#Assertor"/>
>   </rdf:Property>
>
> So you would be asserting that the location is an Assertor...
>
Oh oh. That's not what I'm trying to say.

> In trying to come up with addressing schemes, I think it would be helpful 
> to write some prose about the things you want to record - I find it hard 
> to guess often what a property is meant to do unless I made it up. (I 
> assume that other people don't always find it obvious what my properties 
> mean either... :-)
>
Good point. Here's some very rough prose - I think that we can put an 
unordered list of marked text within the rdf:li element that describes the 
thing causing the accessibility error. We don't know all the stuff that's 
going to go in there yet but it's likely to be a large variety of things. I 
think it will be just a list of items with a name that identifies the whole 
bunch. Several bunches of data may be required for describing some things 
that cause accessibility problems (like example 
http://checker.atrc.utoronto.ca/test197-earl.html).

> If you want to model the atributes that things have, and the values they 
> do (or ought to) have, there might be a better way than making a new earl 
> property for each attribute or element we come across. I think it should 
> be possible to create a very limited number of properties like element, 
> attribute and value, and specify them properly, but I should think about 
> it a bit more...
>
Sounds good to me. I like your suggestion of the basic 3 things - element, 
attribute, and value.

> This example also makes me wonder about confidence. I cannot understand 
> why the confidence would be low in this case. Nor what you would do with 
> the information.
>
The confidence rating is used to state how confident you believe the test 
result to be. In the example 
(http://checker.atrc.utoronto.ca/test197-earl.html) our accessibility 
checker flags this error for every anchor element. Our checker isn't smart 
enough to interpret the link text and interpret the destination document and 
determine if the two match. So when it flags this error it gives a low 
confidence rating - it's not sure there really is an accessibility problem. 
If a human was to perform the same test, they could state the error with a 
high confidence rating.

This information is useful to anyone reading the accessibility report. If an 
error has a high confidence rating (example: image is missing an alt 
attribute) then that error must be fixed. If an error has a low confidence 
rating then the error may not need to be fixed.

This is not a perfect system and could be improved but I believe it's a 
useful thing to have for now.

Cheers,
Chris

Received on Monday, 11 April 2005 09:22:35 UTC