Evaluation and Repair Tools Working Group Teleconference

Review of updated EARL 1.0 Schema

MS: let's jump directly to the conformance section first
... thought it was helpful

http://www.w3.org/WAI/ER/EARL10/WD-EARL10-Schema-20090223#conformance

scribe: separation between producers and consumers is important
... seems to be little difference between the two

CV: does the term validation make sense?

<cvelasco> http://www.w3.org/TR/CCPP-struct-vocab/#Conformance

CV: CCPP does it differently

SAZ: MikeS had an action item, and confirmed that different specs define conformance differently

CV: they established first document conformance, then producers, then processors
... more fine grained

SAZ: validity term not an issue?

CV: CCPP also use it, should be ok

SAZ: an issue of organization, want more highlighting of the sections?

CV: yes, also address document conformance

SAZ: EARL is not necessarily a single document, rather data
... also Johannes raised an issue about unclarity of what producers/consumers should do

<JohannesK> "Model conformance" instead of "Document confomance"?

SAZ: maybe also relates to formatting and organization

[suggestion: Data Conformance]

MS: separation between consumer and producer

SAZ: found that there is large overlap, although producers could be more minimal because they do not need to output all data
... but consumers should be able to process all data defined by this document

MS: maybe need to be spelled out more clearly
... or do the opposite, mention it briefly and see what comments come in

http://www.w3.org/TR/CCPP-struct-vocab/#Conformance

CI: agree that current wording is a little vague
... not sure how to improve it

SAZ: going in the right direction?

CI: think so

<JohannesK> What about the properties with no restrictions? Is a conforming EARL producer required to be able to support these properties?

SAZ: like to continue with this approach, seem to be on a right track
... could work on improving the organization and formatting
... think that we will need to look at the restrictions closely
... especially during Candidate Recommendation stage we may identify conflicts or unexpected situations

<scribe> ACTION: shadi to continue refining the conformance section by spelling out report, producer, and consumer conformance more clearly [recorded in http://www.w3.org/2009/02/25-er-minutes.html#action01]

<trackbot> Created ACTION-78 - Continue refining the conformance section by spelling out report, producer, and consumer conformance more clearly [on Shadi Abou-Zahra - due 2009-03-04].

SAZ: EARL producer does not have to generate optional properties
... would not be possible to check if they can actually support it

CV: would need a test file with an optional property, and see if the output also contains that property or if it has been dropped

SAZ: that is a new type of tool, like an aggregator
... it is a consumer and a producer at the same time

<JohannesK> A person using that producer tool could check

SAZ: we could require these not to drop information from the input

MS: doesn't address the actual question

SAZ: could say producers MUST be able to generate all required properties, and SHOULD be able to generate the optional ones

CV: don't think this is real conformance, should be able to generate all

CI: why are the properties optional then?

<JohannesK> because in some cases they don't make sense

SAZ: seems like we may need something like partial conformance
... for example to encourage tools that only output certain type of information

MS: how do we define partial?

SAZ: tools that support any parts of EARL

MS: assertion without an assertor?

SAZ: could require that if a class is generated, all required properties must be supported

<scribe> ACTION: shadi to add information about aggregator tools and about partial conformance [recorded in http://www.w3.org/2009/02/25-er-minutes.html#action02]

<trackbot> Created ACTION-79 - Add information about aggregator tools and about partial conformance [on Shadi Abou-Zahra - due 2009-03-04].

<JohannesK> Shall we add "Schema-aware processing is not required" to the consumer section like in CC/PP?

SAZ: was confused about redefining Domain and Range
... Ivan confirmed that it is not a redefinition
... it is ok to do

MS: so it gets multiple domain and ranges?

SAZ: yes, Johannes was right

status of EARL requirements and EARL Guide documents

MS: who is editor for the requirements?

SAZ: I was, never got round to it
... we have change requests since the f2f

MS: I could look at these
... what about the Guide?

CV: Johannes and I
... will work on it

MS: maybe provide something in 2 weeks from now?

CV: yes

Next and future meetings

next meetings: 4 March, 11 March