Re: Location Class

On Monday 19 September 2005 18:42, Shadi Abou-Zahra wrote:
> Hi,
>
> During the previous teleconference we decided to use the current "assertor
> approach" for the location: we will not pre-define the location instance to
> *always* be a bag but that it can be either single type or compound type
> (where more than one trigger constitute one occurence of a result).
>
> Please find a quick draft below for discusion on tomorrow's call:
>
>
>   <rdfs:Class
> rdf:about="http://www.w3.org/WAI/ER/EARL/nmg-strawman#Location">

[chop]


>   <rdfs:Class
> rdf:about="http://www.w3.org/WAI/ER/EARL/nmg-strawman#singleLocator">

Hmmm.

Zero or more of [ title, description, line, column, xpath, fuzzy]

No mention of charoffset or byteoffset?  CSS-selector?   xpointer?

And no mention of relationships: for example, a column without a line
is nonsense, and line/columns must presumably come in pairs if there
are more than one.  Surely line/column should be treated as a pair,
with the column being optional?

And the concept of a *single* locator comprising (say) multiple titles,
descriptions and xpaths is, erm, interesting.


>   <rdfs:Class
> rdf:about="http://www.w3.org/WAI/ER/EARL/nmg-strawman#compoundLocator">

[ title*, description*, Locator+ ]

Again, should we be *that* flexible about numbers of titles/descriptions?


-- 
Nick Kew

Received on Tuesday, 20 September 2005 15:51:11 UTC