XML Community of Practice
Meeting Notes
July 20, 2005
The Department of Education’s Office of Federal Student Aid (FSA) hosted this meeting.
Following introductions, Owen Ambur displayed the most recent submissions to the ET.gov site at http://et.gov/component_search.aspx and encouraged attendees to consider whether the xmlCoP should try to help build CoPs around any of the components registered on the site. He called attention in particular to the three components registered by NIST and KC Morris spoke briefly about them.
Holly Hyland provided a brief update on the status and plans for FSA’s XML registry. She indicated they were contacted by five agencies following her previous presentation to the xmlCoP in March 2005. In May they won a best practices award for the registry from Postsecondary Education Standards Council (PESC). [Editor’s note: An announcement about the award is available at http://www.pesc.org/publications/press-releases/Best%20Practices%20and%20Services%20Awards%205-2005.pdf] FSA currently has three schemas under development. They have hired a data architect and are using Embarcadero for data modeling. The models will be stored in Popkin System Architect, and they are aiming toward the establishment of a mandatory data model for FSA. Frank Napoli asked if they are considering external data sharing needs too.
Holly’s notes concerning the registry are available at http://xml.gov/documents/completed/fsa/edrr.htm
Participating via teleconference, Tim Mathews mentioned the freebXML registry and, with respect to XML naming and design guidance, suggested that a core set of rules must be applied on a mandatory basis in order to achieve interoperability. Todd Vincent suggested the declaration of global types is one such rule but that rationales must be provided for any rules that are deemed to be mandatory. Joe Carmel suggested that prohibiting the use of abbreviations and lower camel case would preclude the use of such standards as MathML. Considerable discussion occurred with respect to the need and wisdom of trying to impose mandatory rules versus strongly suggested guidance, specifically whether the word “should” should be used instead of “must”. Owen asked for a showing of hands as to which word should be used with respect to the use of upper camel case, and there were no votes in favor of trying to impose such a requirement with no possibility of justifying exceptions. Owen asked Tim to change the wording of that provision in the draft from MUST to SHOULD and Tim indicated he would do so.
Following the break, Ken Sall briefed the group concerning XML naming and design issues raised by the Intelligence Community’s (IC) Metadata Working Group staff. He noted the IC is using the draft of the XML Developers Guide posted on the xml.gov site in April 2002. [Editor’s note: The Guide is available at http://xml.gov/documents/in_progress/developersguide.pdf.] The IC Metadata Working Group is comprised of 15 member agencies but has touch points with 35 organizations. Attendance at their meetings generally range from 50 - 100 people. The current draft of the proposed XML naming and design rules is unacceptable to the working group but they want to make it usable.
The IC also had concerns with the first draft of the XML specification for the Federal Enterprise
Architecture (FEA) Data Reference Model (DRM). For example, security markings have now
been incorporated into the DRM; otherwise the IC would have difficulty exchanging DRM
instances that might contain classified information. Likewise, prohibiting the use of acronyms
may preclude use of the XML naming and design guidelines because some acronyms used by the
IC may be unclassified whereas the expansions of the acronyms to spell out the full words may
be classified. Some issues with the current draft of the XML naming and design guidelines are
show stoppers for the IC and must be corrected.
Joe Carmel noted that the Library of Congress wants to use xsd:any for legislation, to avoid breaking the schema in the future. Ken cited xsd:union as another requirement and, from Ken’s description, Owen noted that it sounds like xsd:union may also be applicable to the data context/taxonomy elements of the DRM. Another example Ken cited is empty elements. Joe noted they are needed in documents and argued the prohibition on their usage should be data-centric. Todd Vincent agreed that the guidelines should distinguish between documents and data.
With respect to the distinction between guidelines versus rules, Ken asked how rules would be enforced. For example, would OMB cut off funding if an agency does not believe a rule makes sense in its business processes? Todd agreed that enforcement is a very important point and he cited the example of the Department of Justice conditioning grant funding on conformance to the Global Justice XML Data Model (GJXDM). [Editor’s note: DOJ is scheduled to host the August 17 meeting of the xmlCoP for the purpose of briefing us on their XML naming and design rules. See http://xml.gov/agenda/20050817.htm.]
Ken also questioned the proposed prohibition on the use of xsd:appinfo. Todd noted that security-related concerns have been cited but that he did not understand them. In addition, Ken questioned the need for two normative versions of each schema, one with full documentation and another without. Owen noted that KC Morris had pointed out the schemas without documentation can automatically be derived from fully documented schemas, so it is necessary only to have the latter. Richard Ordowich mentioned performance issues associated with the size of the GJXDM.
Finally, with respect to the proposed establishment of the Oxford dictionary as the authoritative source for definitions of terms used in XML schemas, Ken noted there are several different editions and none of them is available without cost on the Web. He also noted that UBL uses the English rather than the American spellings for terms, whereas the proposed U.S. federal XML naming and design rules specify the use of American spellings. He suggested using either the Concise Oxford English Dictionary (2004 edition) or the New Oxford American Dictionary (2005 edition).
Changes requested by the IC Metadata Working Group are documented at http://xml.gov/documents/completed/icmwg/NDR-Position-Paper.pdf. Ken displayed and referenced that document at the meeting. It points out 22 problems and identifies recommended changes for each, including three major points:
• the document should describe best practice, not mandatory rules,
• examples and justifications are essential and should be added, and
• mandatory rules (if any are needed) must be the minimum absolutely
required to encourage interoperability and data sharing.
At the close of the meeting, Roy Morgan announced that usage of the proof-of-concept XML registry had been low in recent months and that it would be retired shortly. [Editor’s note: See the Registries page at http://xml.gov/registries.asp.]
Among those in physical attendance were:
Owen Ambur, xmlCoP
Tim Bornholtz, ED
Azad Faruque, DOS
Puja Goyal, NIST
Renee Higgs, Lockheed Martin/SSA
Holly Hyland, FSA
Tom Merkle, NIJ
Roy Morgan, NIST
KC Morris, NIST
Frank Napoli, LMI
Quyen Nguyen, NARA
Richard Ordowich, HHS
Ken Sall, SAIC
Chris Traver, DOJ/OJP
Among those who identified themselves as participating via teleconference were:
Todd Vincent, Legal XML
Tim Mathews, LMI
Joe Carmel, retired from U.S. House of Representatives
Please convey any additions or corrections to Owen_Ambur@ios.doi.gov