Re: EARL for a mixed-input report

On Thu, 31 Oct 2002, Charles McCathieNevile wrote:

> This seems to make sense, except that I don't think you need to have two
> levels.

Since my original post, I've put up some real examples of this on the
web.  If you haven't already done so, can I suggest you look at the
HTML version
(e.g. <URL:http://valet.webthing.com/access/example-full.html>;
there's also a corresponding .rdf)
I thinks this makes the motivation of the report very clear.
Level 1 is the automatic analysis Valet has done in the past,
while Level 2 is new, but has to be based on a Level 1 report
(or it becomes just another checklist-with-a-GUI tool).  The EARL I
posted was for Level 2, but I didn't want to lose the one-line
summary from the Level 1 report completely.


>    The assertor is either the  tool or the person - the fact that a tool
> chooses to trust a person over itself means that it should suppress its own
> result, or that it should provide both results allowing someone else to
> decide the tool is smarter than the person.

But the tool is now doing two different things, one automatic and one
heuristic, quite apart from recording the person's assertions.

> The significance of reification is that it lets us say "joe says that" and
> "fred says that" without running into a problem because the things they say
> are contradictory (which is what would normally happen if we merged the two
> statements as bare RDF).

What bothers me is when I find use of rdf:[subject|object|predicate]
seems to be an only alternative to jumping through hoops.  In this case
I was drafting something with pure-ish EARL terminology, but found it
much simpler to express when I went back to rdf-land.

> On the other hand the new schema uses some nicer property names to make the
> reification less visible and make it more obvious what is going on - the
> tested, result, etc properties are less general than
> subject,predicate,object, but I think that is an advantage.

Yes, that's how I think it should be, and that's why it's bothering
me that it didn't work like that in this case.

-- 
Nick Kew

Received on Thursday, 31 October 2002 19:37:05 UTC