W3C

ERT WG

27 Sep 2006

Agenda

See also: IRC log

Attendees

Present
JohannesK, Shadi, Daniela, Chris, CarlosI, CarlosV
Regrets
Charles
Chair
Shadi
Scribe
Johannes

Contents


EARL 1.0 Schema published

<shadi> http://www.w3.org/TR/EARL10-Schema/

SAZ: new schema draft out now
... notifications go to WAI IG and RDF CL etc
... thanks for hard work

min/max cardinality on assertion properties

Daniela: hi everybody

<shadi> http://lists.w3.org/Archives/Public/public-wai-ert/2006Sep/0037

JK: proposal 2 is a shorthand for having multiple assertions; then the single properties (e.g. result and locations) would be all the same for each subjects/tests
... you could do this compression with results without instance locations

SAZ: we don't want to propose omitting instance locations

CV: I like proposal 1, don't like grouping

CI: prefer 2, can live with 1

CR: probably prefers 1

JK: prefers 1, 2 could be dangerous

SAZ: how about not restricting in the schema (proposal 2), but in prose texts use porposal 1?

CV: too much confusion

RESOLUTION: proposal 1 (exactly one subject and test properties) agreed, have some more discussion on the mailing list

<scribe> ACTION: shadi to investigate whether using Bags or Seqs would be valid in EARL [recorded in http://www.w3.org/2006/09/27-er-minutes.html#action01]

<danbri> (de-lurking.... to note that rdf:Bags are very unfashionable in RDFland... and Seqs somewhat too... feel free to mail me on this... danbrickley@gmail.com)

SAZ: seems like people want to drop confidence

CR: drop it

CV: +1

SAZ: no volunteer for submitting a proposal about confidence?
... the EIAO European project uses confidence in a very specific way

RESOLUTION: drop confidence

SAZ: cardinality for validity and instance?

JK: range of instance is a PointerCollection; is it necessary to have multiple instance properties?

SAZ: pointers in PointerCollection are all about the same location

<shadi> http://www.w3.org/TR/EARL10-Schema/#instancelocation

SAZ: proposal: exactly one validity
... no objections

RESOLUTION: exactly one validity property for TestResult

SAZ: Software does not always act like a foaf:Agent

RESOLUTION: Software is standalone class (no subclass of foaf:Agent)

SAZ: how to identify objects that don't have a URI?

JK: in RDF you can create nodes for everything

SAZ: what about files on a hard disk?

CV: use WebContent

SAZ: isn't WebContent only for something to be found on the WWW?

JK: rdf:about="file://localhost/tmp/foo.bar"
... that's a URI

<shadi> <quote url="http://www.w3.org/TR/EARL10-Schema/#webcontent">Information on the World Wide Web.</quote>

JK: the main issue is not having a property to store the content for non-HTTP stuff

SAZ: snippet stuff? no

JK: use http:body (storing content Base64 encoded)

CV: move the body property into another namespace for use in EARL and HTTP-in-RDF
... rename WebContent to Content
... or clarify the meaning to include not only WWW content

<scribe> ACTION: shadi to send to list for discussing the testing of local stuff on list and next week [recorded in http://www.w3.org/2006/09/27-er-minutes.html#action02]

Summary of Action Items

[NEW] ACTION: shadi to investigate whether using Bags or Seqs would be valid in EARL [recorded in http://www.w3.org/2006/09/27-er-minutes.html#action01]
[NEW] ACTION: shadi to send to list for discussing the testing of local stuff on list and next week [recorded in http://www.w3.org/2006/09/27-er-minutes.html#action02]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.127 (CVS log)
$Date: 2006/09/27 16:10:43 $