XML Working Group

Meeting Notes

July 21, 2004

The meeting was hosted by the Merit Systems Protection Board at its office in Washington, D.C.

Owen Ambur announced that the next meeting, scheduled for August 18, will focus on XML registry services and be hosted by Booz Allen Hamilton at their office near Metro Center. (The draft agenda is available at http://xml.gov/agenda/20040818.htm) He also noted that the second in what may become an annual series of XML authoring/editing tool forums is being planned for this fall. The focus of this year’s event will be on legislative and regulatory documents and a second track will address architectural approaches to XML content management.

Since time did not permit him to do so at the June 16 meeting, Steve Hamby demonstrated the Navy Interactive Electronic Technical Manuals (IETMs) application that will be deployed on all Navy vessels and possibly Coast Guard cutters. http://xml.gov/minutes/20040616.htm The manufacturers provide information in XML format, using their own schemas, and the Navy incorporates it into the manuals. The application uses XQuery for structured as well as full-text searches, and it uses XSL for sorting of results. Roy Morgan noted that the presentation of the data is very important because, for example, a warning note must always appear on the page to which it applies. Steve indicated the application uses Corel’s SVG viewer, which is bundled with the application, but that Adobe’s SVG viewer may not deliver exactly the same presentation.

Steve then turned to the primary purpose of his appearance at this meeting, which was to represent the Integration Consortium in briefing the group on the components of enterprise application integration (EAI) in relation to eGov, lines of business (LOBs), enterprise architecture, and XML registry services. He characterized the common scenario with information technology systems as modern cities connected by dirt roads. Owen noted that the slide <http://xml.gov/presentations/ic/sld003.htm> Steve used to illustrate that point pretty well characterizes the eGov projects and line of business initiatives, each of which has been pursued as a stovepipe application.

Steve pointed out that EAI includes, among other components, adapters and information integration tools. Owen observed that the use of such tools is an important part of a transition strategy for migrating from the “as is” to the desired “to be” state in IT architectural planning and investment control. In that regard, Steve noted that the cost of EAI generally cannot be justified for the first investment but only in terms of building the necessary infrastructure to support the enterprise as a whole. (emphasis added)

Steve describe the Global Integration Framework (GIF) as a universal lexicon for integration terminology. http://xml.gov/presentations/ic/sld013.htm GIF provides a central registry of GIF-compliant interfaces, e.g., for linking Web Services to a business model, WSDL to OWL, etc. It is a consolidation of best-of-breed practices from the W3C, OMG, and The Open Group (TOG). It is based upon the TOG’s Architectural Framework (TOGAF), which addresses business, data, application, and technology architectures similar to the Federal Enterprise Architecture (FEA). However, unlike the FEA, TOGAF does not encompass a Performance Reference Model (PRM), and Steve suggested Software AG may be able to facilitate the inclusion of a performance reference architectural model as well.

GIF uses an N-tier model and each tier should be separate. Benefits include: reuse of integration objects, end-user/subject matter expert (SME) driven changes, and lower cost of integration. Steve reiterated that it requires spending more on the first application but that it will be justified as an investment in infrastructure, rather than in terms of a benefit/cost analysis on the first application. In that regard, Owen suggested that a single set of EAI components could service all of the eGov projects and OMB’s LOB initiatives. Steve agreed and noted that doing so would enable agencies to leverage legacy stovepipe applications in an open, citizen-centered manner.

With respect to XML registry services, Steve noted that WSDL and J2EE documents can be stored and made available in XML format. He also referenced Semantic Web Rule Language (SWRL) and said the final architectural component to be addressed is security. See http://www.daml.org/2003/11/swrl/ & http://xml.gov/presentations/ic/sld026.htm With reference to the security layer, Owen observed that it encompasses a lot of heavy duty processing and asked whether the Integration Consortium had engaged the XML hardware vendors. http://xml.gov/presentations/ic/sld026.htm Steve indicated that might be a good idea and Ian Foster observed that convergence is occurring among the XML security hardware vendors.

With reference to citizen-centered eGov, Roy noted that government employees are citizens too and that it is currently too hard for them to discover and obtain standards that are applicable to their applications. He suggested the IC GIF may be a means to enable them to do so more easily. Barry Schaefer observed that citizen involvement/participation, such as in regulatory decision-making, requires different logic than just providing access to information, but the dichotomy has not been addressed effectively yet.

Steve’s presentation is available at http://xml.gov/presentations/ic/index.htm A white paper containing additional details on the IC’s methodology for Total Business Integration is available at http://xml.gov/presentations/ic/tbi.htm

Bentley Roberts, Clerk of the MSPB, briefly discussed the board’s records management responsibilities and noted Tim Korb’s expertise in both MSPB law as well as technology.

Tim introduced Ipedo’s presentation and briefed the group on MSPB’s activities, which fall into two categories: 1) studies and reports, and 2) the adjudication of employee appeals. Those activities involve and generate a lot of documents, and those documents are not as accessible as they should be since they are maintained in stovepipe systems with limited and varying search capabilities. His office spends a lot of time copying documents from one repository to another and converting from one format to another. They are seeking cross-repository search capabilities and a non-proprietary format that affords confidence of accessibility into the future. Thus, the issue became how to get their documents converted into XML and that led them to Ipedo, which has the capability to convert virtually any format into XML while preserving the presentation of the original.

Ipedo stores the XML files in a native XML database and can pull metadata out of the legacy systems for indexing in the new system. Output can be produced in multiple formats, while the original versions remain undisturbed unless the board decides to retire them. In addition, the legacy applications can be synchronized with the new one. Unlike the U.S. Courts, which requires documents to be submitted in PDF, MSPB plans to accept any format via eFiling into its case management system. The system will greatly facilitate responses to FOIA, Privacy Act, and Congressional inquiries by taking advantage of both full-text as well as fielded search capabilities.

Owen asked if the board had considered using a proposed XML-based standard like XFDL to preserve the presentation of documents for record-keeping purposes. Alice Marshall asked why the courts require the use of PDF. Owen explained that the integrity of records considered as evidence must be assured, and that they must look exactly the same tomorrow as they did yesterday. Otherwise the reliability of the evidence may be called into question. Jim Disbrow noted that RTF used to be a reliable format for exchanging files between Word and WordPerfect but no longer works between them.

Tim’s notes are available at http://xml.gov/presentations/mspb/introduction.htm

Via telecon and Webinar from Ipedo’s office in Redwood City, California, Kam Thakker briefed the group on Ipedo’s XML Intelligence Platform (XIP). He said it took them about two days to set up the demo using MSPB documents. XIP provides Enterprise Information Integration (EII) by leveraging XML. It provides real-time access to legacy databases in rapid time, e.g., 2-3 days or a week. It creates a virtual data layer and can pull in unstructured documents as well as data from databases, using either message/push or Web Services/pull capabilities. It supports and uses XQuery and is available outside as well as inside a firewall, with security. The three layers of XIP are defined as Assemble, Analyze, and Visualize. It does not “touch” the legacy databases and does not perform data synchronization. Instead, it provides a vitualization layer that makes applications format and schema independent.

XIP is integrated with both .NET and Java. Key features include XML views, XQuery builder, XML rules processing, XML pipelines, and XML “accelerators” (templates). The Integration Manager is a client/server application and can provide bi-directional synchronization of data, if necessary. The XQuery capability supports EII and about 200 functions are provided in the XQuery/XPath library. XML schema management functions are also provided and Ipedo strictly supports open standards, including WebDAV for views of Ipedo document collections. Owen asked if XIP supports the metadata capabilities of WebDAV, and Chetan Patel said the next version will.

Kam then demonstrated the query/retrieval capabilities of XIP using .doc and HTML renditions of MSPB documents. Diane Lewis asked if the documents are “intelligent” and Chetan said XIP uses .doc styles to glean intelligence that is implicit. Tim took some credit for the fact that MSPB’s documents are well and consistently styled, thereby facilitating the capture of such intelligence in more usable XML format. Brand Niemann, Jr., noted that his employer, Tax Analysts, is doing something similar using a competitor’s product. Someone asked what happens when new documents are added in a legacy application and Chetan responded that synchronized indexing can be set up. Kam also noted that existing SQL queries can be reused.

Kam’s presentation is available at http://xml.gov/presentations/ipedo/xip.htm

Approximately 20 individuals attended part or all of the meeting. A number of MSBP employees did not add their names to the attendance list. Among those in physical attendance who registered their presence on the xmlWG meeting roster were:

Bruce Altner, NASA/SAIC

Owen Ambur, Co-Chair

Jim Disbrow, DOE

Ian Foster, Software AG

Cheryl Holmes, Deque

Steve Hamby, Software AG

Timothy Korb, MSPB

Diane Lewis, NHTSA/DOT

Alice Marshall, Presto Vivace

Roy Morgan, NIST

Frank Napoli, LMI

Vicky Niblett, SAIC/NASA

Art Saenz, NSF

Barry Schaefer, X.Systems

Tim Bornholtz, DoEd

Among those participating by teleconference were:

Bob March, eCourt filing (retired)

Brand Niemann, Jr.

Please convey any additions or corrections to Owen_Ambur@ios.doi.gov 

Alice Marshall’s account of this meeting is available at http://technoflak.blogspot.com/2004/08/federal-xml-work-group-july-21st.html