XML Work Group, May 18 meeting

Last Tuesday, I went to the monthly meeting of the Federal XML Work Group. This is a small part of the effort to establish a single federal enterprise IT architecture standard. From the website:

Extensible Markup Language (XML) embodies the potential to alleviate many of the interoperability problems associated with the sharing of documents and data. Realizing the potential requires cooperation not only within but also across organizations. Our purpose is to facilitate the efficient and effective use of XML through cooperative efforts among government agencies, including partnerships with commercial and industrial organizations. Contributions are welcome and encouraged!

The standard that evolves out of this process will enable not only the assorted agencies of the federal government to cooperate, but also enable cooperation between state and municipal governments as well. Given the collective buying power of government agencies, it will become the de facto industry standard. Technoflak has been deeply impressed by the level of commitment of all the participants to developing a standard that will have lasting and practical value for everyone.

The May 18 meeting was a combined meeting of the IRS XML Stakeholders group and the XML Working Group, hosted by the IRS. These groups have many of the same objectives and are working collaboratively. The XML working Group is co-chaired by Owen Ambur (Interior); the IRS XML Stakeholders is chaired by Sol Safran(IRS).

Arriving late, Technoflak missed most of Judy Newton’s presentation on developing a metadata registry with a focus on ISO11179. One thing I did catch was her concern that the software tools that automatically generate XML schema, say from SQL, do not document the source of the data. One way to ameliorate this problem is with a link to a 11179-based metadata registry. Newton encouraged participants to visit the Joint Technical Committee’s website.

Steve Wasko, of IRS public portals, spoke about their public portal strategy for metadata and taxonomy. The IRS and Mitre are working to develop tax administration terms and a companion thesaurus. The goal is to deliver better search results for taxpayers using the IRS website. Mitre consultants have spoken with fifteen IRS business units and sixty IRS staff members to develop a list of 800 terms. Their Office of Chief Counsel had used a uniform issue list of 16,000 terms, but agreed to use the new list of 800 terms. Technoflak could only marvel at the triumph of bureaucratic cooperation.

Wasko said there were problems integrating the metadata and new taxonomy with the existing content management system.

Claude Matthews, of IRS web services, spoke about the use of Metasoft to organize eLearning. He described the convergence between the IRS employee support systems, the learning & education systems and the media & publications systems.

He described the work of the Enterprise Data Management Office (EDMO), which involves procedures (not correspondence). It is important to keep in mind the different needs of the taxpayer, tax examiner and security. Matthews encouraged participants to write technical procedures in XML, making it possible to display information in different ways to different users.

Metasoft is based on SCORM, which is an extension of LOM and Dublin Core.

He characterized Metasoft as a developer’s tool and explained that Infopath is used to develop templates, because it facilitates XML tagging. Metasoft is used by InfoPath as a trusted source.

Matthews said that the Internal Revenue Manual should drive all IRS content. The current topic map is in simple English; the internal system is more technical. Successful system integration may require adoption of a new XML structure (OASIS? LegalXML?). Namespaces will consolidate many “stove-piped” areas.

Currently, content is driven by Business Operating Divisions and process owners, leading to duplication of effort and an updating nightmare. In the future, content will be driven by the content data model, MITS via BSMO/EDMO, an easily updatable trusted source. Matthews said a collision is imminent. Technoflak agrees.

Manny Tayas of Systinet spoke about UDDI best practices. From his first slide:

“The Universal Description, Discovery and Integration (UDDI) protocol is one of the major building blocks required for successful Service Oriented Architectures.”

“UDDI creates a standard, interoperable platform that enables organizations and applications to quickly, easily, and dynamically find and use shared services over standard internet protocols such as HTTP."

“UDDI is a cross-industry effort driven by major platform and software providers, as well as marketplace operators and e-business leaders within the OASIS standards consortium.”

Tayas compared UDDI’s function in web services to Windows directory in Windows. It is a key component of Service Oriented Architecture. UDDI enables interoperability between applications, register & discover functions, and coordinate & compose functions.

UDDI is standards based, flexible, with a wide variety of publish/discover tools available. Tayas stressed that UDDI is not a repository, but a description of services: basic contact information and identifiers about a company or service provider, categorization of web services using taxonomies, and technical information describing a web service. He encouraged participants to visit the UDDI website to get a clear understanding of its capabilities.

Modeling UDDI entities should start with thinking about what constitutes a businessEntity and a businessService within the organization. Once these are established, this information should be documented for others to consume and governance policies should be put into place to enforce. Tayas suggested the organization chart as an example of defining businessEntities. A lively discussion ensued on whether project level or security aspects might serve as a better basis for definition of businessEntities.

UDDI was developed as a general-purpose service registry and was not specifically designed for web services, though web services are the most logical type of service to publish in the registry and a good place to start.

Tayas explained the process of publishing from Web Services Description Language (WSDL) to UDDI. Using UDDI ensures interoperability with application vendors for discovering WSDL-based services. It will enable better query capabilities. He said using taxonomies are the key to promoting reuse and to establish categorization schemes before deploying.

Tayas concluded with a discussion of process and procedural considerations. It is critical to standardize publications and establish publication guidelines.

Presentations

Minutes of past meetings