- From: Rui Lopes <rlopes@di.fc.ul.pt>
- Date: Mon, 4 Apr 2011 17:43:17 +0100
- To: ERT WG <public-wai-ert@w3.org>
Dear all, My take on this, below: On Mon, Mar 28, 2011 at 11:41, Shadi Abou-Zahra <shadi@w3.org> wrote: > # Issue > > In one way, EARL provides a vocabulary of *semantic* terms that can (and > should) be used in many different ways. Like we use terms from FOAF, Dublin > Core, and potential DOAP vocabularies, we also want to promote reuse of EARL > terms in different contexts. For those people, the conformance section does > not make sense as currently written. > > On the other hand, we want to continue to define a *syntax* for tool > developers to exchange evaluation data. That is, a particular way of using > the EARL terms so that an EARL producer and EARL consumer can exchange a > well defined EARL report. We have definitions for these already, but they > are somewhat mixed with the vocabulary definition. Last week I met with Shadi at W4A, where we discussed a bit this point. Summarizing my take on this: 1) Indeed, separating both worlds is good to further disseminate EARL as a reusable vocabulary. RDF has been created from the ground up such that a Web of triples is not dependent on strict schemas; 2) Strict conformance is something more akin to "traditional" vocabularies at W3C (i.e., XML-centric, not RDF-centric). This has leveraged incredible ecosystems of tools around XML languages such as XSLT, XQuery, XML Schema, XProc. The only issue I see in these two perspectives is the non-recommendation status of the EARL Guide: will we still have an ecosystem of EARL producers and consumers? I don't have a definite answer for this, although my feeling is that *yes*, there will still be an ecosystem. Going onwards with the proposal is an acknowledgement of this issue, independently of the outcome. Cheers, Rui > # Proposal > > Move the conformance sections from their current location to the EARL 1.0 > Guide document, but keep a cross-reference that is clearly focused to tool > developers. This way the current specs become pure vocabulary definitions, > and the EARL 1.0 Guide becomes a more relevant source for tool developers. > This involves minimal changes yet several benefits. > > > Regards, > Shadi > > -- > Shadi Abou-Zahra - http://www.w3.org/People/shadi/ | > WAI International Program Office Activity Lead | > W3C Evaluation & Repair Tools Working Group Chair | > >
Received on Monday, 4 April 2011 16:44:10 UTC