Hi Lars, Karen,
On 6/9/15 10:22 PM, Karen Coyle wrote:
> On 6/9/15 2:18 AM, Svensson, Lars wrote:
>>> R-210 "Define valid properties"
>>> >http://lelystad.informatik.uni-mannheim.de/rdf-validation/?q=node/423
>>> >The title is slightly confusing: a statement with an 'allowed' property may still
>>> >be 'invalid' (because its value doesn't have the right form, for example). We
>>> >suggest to replace it by "Define allowed properties"
>> I'd suggest "Define allowed properties for a context"
>
>
> This gets very complex, and it reminds me that we haven't had the conversation about terminology that we might need -- context, graph, node??
>
> The other complexity is whether "valid" is a closed or an open state. That is, if I say:
>
> valid properties are: A, B, C
>
> and my data also has property D, is that a violation of my rule? In the W3C group this is being talked about as "open" vs. "closed" graphs. There are times when you might want to say "only A, B, or C" [1], and other times when you want to say "A, B, or C, and I'll ignore anything else." If with "ignore anything else" there is the question of: will I keep it, or drop it on the floor because I don't know what to do with it?
>
> So you can see that this potentially becomes a number of more specific sub-requirements. Do we want to take those on? It seems that unless we do get more specific, what we have said cannot be coded.
>
> kc
> [1] Then there's the "OR" problem - exclusive or not. And that compounds the sub-questions above.
>
Not putting 'context' too much in spotlight was one of the reasons we've suggested the simple "Define allowed properties"
I think I can live with Lars' suggestion, though. "context" is vague enough, I expect it will not conflict with whatever technical choice is made later.
"valid" is a tricky point, indeed, and that's why we're suggested "allowed". If a rule says that only properties A, B, C are allowed then if there's a description with property D, it's a violation of the rule. End of story!
Cheers,
Antoine
|