E-mail
Attending: Joel Farrel, chair; Valerie Smothers, Carl Singer
Leap 2 A validation
Valerie commented that she has had problems validating LEAP2A documents against Atom unofficial schemas. XML Spy trips up on the inclusion of RDF elements under the entry element. Simon Grant, LEAP2A's developer, had indicated that Relax NG or Schematron m ay be better approaches to validation.
Joel commented that the XML validator may not know if something is valid, but he didn't see why it would say it was invalid. No one currently has to use relax ng or schematron to validate MedBiquitous XML. It would be nice if we could stick with that.
Valerie asked Joel what IBM's take on schematron and Relax Ng was. Joel commented that IBM doesn't define anything that requires them. They do participate in consortia that define document types where you can do a schematron validation on top of xsd validation. They don't have schematron tool; it's nothing they make a big deal of.
Carl commented that Valerie's findings were consistent with their tendency not to reply on xsd for validation. They know what they are looking for and handle it programmatically. They have used rdf in atom documents. They tried using a Microsoft WCF-based web interface and had much trouble in getting an Atom document to pass the systems internal validation. He also had difficulty in getting the document accepted by .net library.
Joel asked is JISC had a Relax NG schema that undetstands their extensions. Valerie commented that they did not but that they seemed to be heading in that direction. Joel explained that current atom users rely on tools like Adeera to parse the XMl and tell if its valid. Adeera is from apache. It's a library that needs to be called from a Java application. It creates java based objects from the xml.
Carl commented that depending on how big the e-portfolio is, validating in a memory tree is probably ok.
Valerie commented that validation is important from conformance/standards perspective. Carl recommended building an open source tool to validate. Valerie replied that the working group members were not in a position to contribute to open source tools at the moment.
Joel offered to send an email to Sam Ruby, an author of Atom, asking if there is an xsd. See what he says. If there is one, maybe we have to add some additional element definitions. Maybe that's good enough for medbiq. That may not be good enough for jisc. They may need to develop schematron rules.
He'll agreed to forward whatever answer he gets from Sam. We can discuss.
Future calls
Valerie reviewed that there is a call on January 19 with ADL to discuss Web Services Testing Forum. The next call after that is Feb 10, which is a joint call with the Professional Profile Working Group. Valerie will distribute meeting and call information for both calls.
Joel reviewed that Doug Davis is confirmed for Jan 19 at 11 AM EST. Valerie agreed to confirm the logistics.