W3C home > Mailing lists > Public > w3c-wai-er-ig@w3.org > October 2001

EARL 0.95 Crinkles

From: Sean B. Palmer <sean@mysterylights.com>
Date: Sat, 20 Oct 2001 03:03:25 +0100
Message-ID: <058301c1590b$85eb7ec0$7ce793c3@y0r1d9>
To: <w3c-wai-er-ig@w3.org>
> 22:15:55 <sbp> what should I do next for ERT?
> 22:16:09 <wendy> well, hmmm.
> [...]
> 22:17:05 <wendy> secondly, i'm still real interested in what
> bugs you about 0.95 and how you would move to 1.0.

Here's a brief outline of some "crinkles":-

@prefix : <#> .
:a = """Why is the Operator class not explicitly connected to the Person
class?""" .
:b = """Can contactInfo be enumerated any further (good plan for
extensibility)?""" .
:c = """Model the master/slave relationship?""" .
:d = """Can we have some properties attached to the context information for
digitally signing the document?""" .
:e = """Why isn't earl:email unambiguous?""" .
:f = """Add stuff like: earl:fax, earl:streetAddress, earl:phone,
earl:organization, earl:gender, etc.""" .
:g = """Is a CC/PP setting a platform? I don't think so.""" .
:h = """What on earth is a CC/PP setting doing in EARL anyway? Do we have
use cases for this? Can anyone provide a scenario, and how this property
may be used in EARL?""" .
:i = """TestSubjects should be unique if they have either of an earl:date
or an earl:version, rather than just earl:date as we have now.""" .
:j = """We should make the sub classes of Test Subject fully pairwise
disjoint, and get some definitions from the glossary and UAWG.""" .
:k = """Let's put in some more properties to let people point to bags of
sites... or do we need to? We can let people attach multiple testSubject
range things to one TestSubject instance.""" .
:l = """Datatypes. Need I say more? earl:Date needs careful
investigation.""" .
:m = """Do we really need date ranges? What are they all about? Will
anything process them?""" .
:n = """Put a range on testSubject. It is highly unorthodox to name a
property after its domain rather than its range, although it does seem to
be understood correctly in this context.""" .
:o = """sameSubjectAs? daml:equivalentTo? Not mathematically so, as we have
already discussed, but we need to pin down the semantics more carefully."""
.
:p = """lastModified: an interesting property. Will EARL implementations
recognize this as being a subPropertyOf earl:date?""" .
:q = """Same goes with earl:released.""" .
:r = """Standard result properties, please. No more customization of
properties.""" .
:s = """earl:cannotTell - must be very careful in the naming of the
properties, and of their semantics.""" .
:t = """Confidences should be scrapped, as discussed in the F2F. High or
nothing. Having said that, there is a point where doing comparisons of
human results may be useful. Notwithstanding that, if people want this
property, it should hang off of the statement as a whole. Can one set
default property values?""" .
:u = """Test Cases... TPDL... aaaargh!""" .
:v = """Point to the test cases themselves? That's the ideal solution.""" .
:w = """Alternatively, EARL could act as the metadata framework for
TPDL.""" .
:x = """testMode should hang off of the whole Assertion""" .
:y = """The whole schema should be documented more.""" .
:z = """Use RDDL for the namespace.""" .

Cheers,

--
Kindest Regards,
Sean B. Palmer
@prefix : <http://webns.net/roughterms/> .
:Sean :hasHomepage <http://purl.org/net/sbp/> .
Received on Friday, 19 October 2001 22:04:30 GMT

This archive was generated by hypermail 2.2.0 + w3c-0.30 : Thursday, 9 June 2005 12:10:39 GMT