Co-chairs Owen Ambur and Marion Royal convened the meeting at 9:00 a.m. at the American Institute of Architects. Attendees introduced themselves. The chairs asked for any comments on/corrections to the January 17 meeting minutes. None were forwarded and the minutes were approved.
Announcements
Jessica Glace (LMI) reported that the following domain teams were created at the X12 Trimester meeting held in Seattle:
Presentation: "Universal Description, Discovery, and Integration (UDDI)"
This presentation is available online at the XML.gov website. A PowerPoint version is available, as well as an HTML one.
Mr. Chris Kurt of UDDI.org and Microsoft briefed the WG on the UDDI initiative.
This initiative was designed to solve some very focused problems with providing and sharing information with internet technologies.
A major challenge facing the use of XML today is the need for standards across industries. Many separate industries are working on creating different standards for themselves. What is needed is a set of standards for sharing information and a standard infrastructure behind this shared information. We are on the cusp of an explosion of XML standards. Situations like this arise whenever many separate organizations attempt to implement a new technology; for example, in the 19th century, early railroads used different standards for track width. It was not until one set of standards was widely adopted that the train system could run smoothly. Today, companies need to work together to avoid creating similar chaos with XML implementation.
The key to understanding the UDDI initiative is to have a good grasp of what a "web service" is. Mr. Kurt defined a web service as a provider of information or capabilities that is exposed on a network through a consistent set of interfaces and protocols. These services are typically lightweight, have a targeted functionality, and are independent of the consumer of their service. Web services focus on being flexible (they can be utilized regardless of whether the consumer is operating on a mainframe, a PC, or a Mac) and on being efficient.
Web services face many challenges, including decisions as to what protocol to use, how the service behaves when accessed, and how to find the information provided by the service. These challenges are all different facets of the same central question: once you put your service on the internet, what will people be able to do with it? Roughly one year ago, a group of people at Microsoft came together to address this problem.
Out of this group came the UDDI initiative. UDDI is an industry wide initiative designed to support web services. There are two aspects to UDDI: specifications and implementation. The specifications include schemas for service description and business description that are developed on industry standards and apply equally to XML and non-XML web services. However, these specifications will not do anyone much good unless those who wish to use them are given a means of implementing them. As a result, the UDDI project has released a set of implementations and provided a set of public service registries.
The UDDI initiative enjoys wide support across industries. This support has grown from 30 companies this past September to 210 companies today. All of these organizations have acknowledged the fact that there needs to be a core set of standards for sharing information. The use of the web for business services cannot be based on a fragmented architecture. As a result, there is considerable resource and product commitment surrounding the UDDI initiative.
Currently, the initiative is still learning what businesses and providers need and is trying to be nimble in incorporating these needs into its specifications. In 2002, the initiative will transition into a standards organization.
UDDI solves several problems. It helps describe services to allow broader B2B possibilities by creating a method for working with hundreds of other organizations. It helps describe a consistent way to provide web services. Additionally, it helps solve the "search problem." Major search engines such as Yahoo! do not give businesses control over where they appear in a search engine. UDDI will allow each organization to place itself into categories based on the services they provide, the geographic areas they serve, and the constituencies they serve. As a result, a web browser will receive more accurate search results when hunting for a service on the web.
UDDI provides a way for companies and web browsers to publish, find, and use web services. Entries into the UDDI registry go into a set of phone-book style pages: white pages, yellow pages, and green pages. The green pages are the focus of UDDI’s work. They provide information on the functional services available from a web service and what sets of specifications are exposed by the service. This lets a user know what he can expect from a service if he sends information to it. Additionally, the green pages will allow services to automate about 95% of their interactions so that their employees can focus on business issues.
UDDI is being used today. Each organization registers its behaviors and is assigned a fingerprint (an unchangeable I.D.). Then implementers enter information on their business services into the registry. The UDDI registry then assigns each service and registration a unique identifier. The registry is simply a set of pointers; it does not host businesses and services. Marketplaces, search engines, and other businesses can then query the registry to find other web services and integrate with them. Additionally, this architecture can be used within a business to query amongst its in-house services.
There are several UDDI registries running at this time. IBM, Ariba, and Microsoft each are hosting registries that perfectly replicate all of the information in the main UDDI registry, providing many access points. At each site, the registry behaves the same way, even if the technology used implement UDDI is different in each company. None of the companies share technologies. One can see, then, that the technology operating behind the scenes is unimportant.
The UDDI initiative releases a refinement and enhancement of its specifications every six months. In March, it will release V2. Currently, there will not be a specification added to UDDI that takes more than 90 days to implement. The folks behind the initiative want organizations to be able to make use of UDDI as quickly as possible.
For further information on the UDDI initiative, please visit the following websites:
UDDI.orgThere is a public UDDI discussion group hosted at Yahoo Groups that WG members are invited to join.The UDDI test site
For information related to UDDI development, please visit the following websites:
SOAP/Web Services SDKOr e-mail Chris Kurt.
Mr. Kurt finished by stating that UDDI enables a P2P model in which companies can automatically locate compatible services from around the world and can deal directly with others without going through a third party.
Mr. Ambur remarked that the government has done some work with white, blue, and green pages.
[Editor's note: The White Pages are at http://www.directory.gov/; the Blue Pages are at http://bp.fed.gov/, and a summary of the X.500 Green Pages is at http://users.erols.com/ambur/greenpgs.htm.]
Mr. Royal responded that the government once used green pages to denote document sources, but the initiative to do so has not continued. He noted that the term green pages originated in the M.U.S.E. report.
[Editor's note: The M.U.S.E. report is available at http://www.iitf.nist.gov/documents/docs/government_email.html and also at http://www.fedworld.gov/pub/misc/musepart.1-d, http://www.fedworld.gov/pub/misc/musepart.2-d, http://www.fedworld.gov/pub/misc/musepart.3-d, and http://www.fedworld.gov/pub/misc/musepart.4-d.]
Mr. Kurt remarked that the government’s work with green pages; UDDI’s green pages also serve as a series of pointers, but to services rather than documents.
Eliot Christian (USGS) then briefed the group on what UDDI could provide the government. His presentation is available online in PowerPoint and HTML formats.
The goal of the UDDI initiative is to speed the interoperability and adoption of web services. Currently, there is no involvement by any government agency in the UDDI initiative.
UDDI can work for the government the same way it works for private industries. The key is to think of each agency as a business entity. Here, agency is defined as the organization that offers the web services. Agencies will register themselves and their services in the UDDI registry. As with private organizations, each agency will be assigned a unique identifier that can be used to identify its services in queries entered by citizens, other agencies, and private organizations.
Mr. Christian then showed the WG an example of the UDDI specifications. It is an XML document that is created by the end-user company, can have multiple service and taxonomy listings. Mr. Christian also showed the groups several of the APIs used by UDDI. These APIs are simple and used by all UDDI registries.
Mr. Royal asked if the identification key is modeled after a pre-existing example.
Mr. Kurt replied that the process of assigning identifiers has been designed to ensure that each service has its own identifier and that there is no collision or repetition of identifiers. UDDI makes use of a complex algorithm to assign these identifiers. Currently, a consumer cannot find a business’s homepage simply by knowing the business’s name. A different method of identification is needed to make sure consumers can find the businesses on the web that they want to find. UDDI gives an organization the ability to tell consumers precisely where its homepage is.
Mr. Christian pointed out that there is currently no registry of federal agency homepages. In fact, the only thing resembling a registry is a page maintained on a voluntary basis by Louisiana State University. UDDI could make it easier for citizens to find agency homepages.
Mr. Christian showed the group an example of the UDDI user interface.
Benefits of using UDDI in populating a registry of federal services include the ability of each agency to self-describe its services and the fact that UDDI can provide sources of locator records, including directories of agency locations, facilities, descriptions, and services.
Currently, the Global Information Locator Service (GILS) can provide similar information. By merging GILS and UDDI, the government can create a gateway that will make information immediately accessible to citizens via the web. Thanks to GILS, the data needed to merge with UDDI has already been compiled; now all that is left is to connect the two technologies.
Theresa Yee (LMI) asked Mr. Kurt what, if any, profit Microsoft hoped to make from working on UDDI.
Mr. Kurt responded that Microsoft has acknowledged that a registry such as UDDI’s must exist. Ultimately, Microsoft will benefit from the information that is in the registry; it will be able to make more efficient P2P and B2B transactions, and will be able to automate many transactions, leaving personnel free to work on other projects. Additionally, the work on this project will give Microsoft an opportunity to learn how to apply this technology to support its own web services.
Richard Campbell (FDIC) asked how the UDDI initiative is dealing with authentication issues. How will the consumer know whom it is he is dealing with?
Mr. Kurt replied that the core principle of UDDI is to create a core structure for providing information about a service. The actual invocation of that service is left up to the designs of the individual agency. UDDI is merely an introduction service; it does not provide a medium for service invocation. So, authentication is the job of the agency. However, UDDI version 3 will include the ability for agencies to add digital signatures and certificates.
Mr. Christian added that UDDI provides a view of the kind of information that is available on an organization’s web site. UDDI does not care about what is actually maintained on the web site in question.
Mr. Royal then asked how UDDI plans on maintaining the integrity of the data in its registry.
Mr. Kurt responded that in UDDI Version 3, users will be able to search only by companies with valid certificates. Additionally, users will be able to search by registrar so they will only retrieve information that they know has been registered by a certain person or agency. Also, search results can be forwarded to other validation services. The core of the UDDI registry cannot do this validation directly, but that is not its purpose.
The WG then broke for fifteen minutes.
Presentation: "Tagless Editing/Markup of OMB Circular A-130, Management of Federal Information Resources"
This presentation is available online in both PowerPoint and HTML formats.
Owen Ambur introduced Keith Thomas of i4i. Mr. Ambur reminded the WG of the challenge he issued on the listserv inviting any interested company to render OMB Circular A-130 in XML format, and to relate the potential of XML to the requirements for the management and use of information technology as set forth in the Circular.
Mr. Thomas briefed the WG on i4i’s work on developing this DTD. The process of converting this circular, or any circular, into an XML DTD involved four steps: identifying the target environment, developing the DTD, creating and deploying the content of the DTD, and providing a framework for managing the content life cycle and re-use of the DTD.
XML is usually oriented towards one of two things: transactions or documents. While most of what we hear about XML concerns transactional XML, in this instance, we are concerned with document-oriented XML. Document-oriented XML can provide a whole new way to manage content.
Documents are part of a network of references. A document can reference and can be referenced by other documents. XML can make this linking structure dynamic.
Every document is an aggregate of information components that can change over time. In order to build a DTD based on a document, one must consider the document components to be the primary pieces of information, not the document as a whole. These components can change over time, can be used in other documents, and have their own life cycles. Components can be graphics, text, spreadsheets, and databases, amongst other things. This focus on the component represents a new approach to content management. In this view, a document is just another component in a web of content. The question for the developer then becomes: if this is a web of content, how can he manage, structure, and control the components over time?
XML gives one the ability to embed meaningful content semantics inside a document. These semantics can be wrapped around other components, allowing machines to process the information contained within.
The use of content semantics provides opportunities for dynamic and user sensitive rendering and improved indexing and searching. Additionally, reference links can be used to create a semantic web of content, and containment links allow the user to move his data management and creation processes down to the component level. But in order to achieve these improvements, one must begin using XML.
The DTD plays a very important role. It defines a grammar for a family of documents (here, the A-130 is part of a family of circulars, many of which are similar). In this case, the developer would want to use valid XML, since he actually wants to be able to manage the document. Valid XML requires a DTD.
Mr. Thomas stated that the DTD he would present for the OMB circular is not meant to be used at this moment. Rather, the purpose of the A-130 DTD is to illustrate the major principles involved in using XML for federal forms.
Mr. Thomas then showed the DTD to the WG. He pointed out the names of various elements and the information that they carry.
Mr. Thomas then addressed the various methods by which one can create and edit XML. An organization can train its document creators in XML. This is the approach taken by most of the XML editors on the market. However, several barriers to this approach exist, including user resistance to change, training costs, and the cost of the inevitable learning curve. Another approach is to apply XML after the document has been created based on the styles provided with the text. This approach also has drawbacks, the primary of which is the cost of correcting errors.
i4i advocates a third approach: tagless editing. XML tags are hidden behind a familiar interface, such as Microsoft Word, so that the user needs only concentrate on the logic of the document itself, not on the XML behind it.
Mr. Thomas then showed the WG sample screenshots of the tagless editor employed by i4i. The user interface is Microsoft Word. However, there is an XML engine lurking behind the interface that inserts the appropriate XML tags. Should the user so choose, he can turn on the tags so that they are visible while he edits the document. Additionally, the standard Microsoft help program offers XML-based tips within the document to assist the user. This application has been designed to be easily customized to various XML documents and to be easy enough for non-programmers to use.
Mr. Thomas then went on to speak about the management of content life cycles and re-use. There are two questions that rise out of the "web of content." First, how can information creators understand the content web and manage their work? Second, how will legacy resources be included in the content web?
Both of these questions can be answered and content management accomplished through the use of content workspaces. i4i is currently working on a new project that is based in part on the notion of a content workspace as a visual metaphor for representing the components of a task. The content workspace itself is a virtual document supported by a metadata server. The actual content resides elsewhere.
Mr. Thomas showed the WG a slide depicting the architecture behind a content workspace. Enterprise servers maintain pointers to a variety of services where the metadata in question actually resides. The server propagates a user’s request and captures data as it flows through it.
In summary, XML can deliver a semantic web of content. It can be deployed enterprise wide without an army of markup technicians, by means of tagless editors that allow analysts to write in XML without seeing the XML. XML can be deployed incrementally to protect an organization’s investment, and as XML itself is an open standard, an organization’s investment in it is protected.
For further information, please contact Keith Thomas or visit the i4i website.
Mr. Royal had a question regarding the content workspace architecture. The slide depicting this architecture showed a specific client for the user and a server that managed the content along multiple sources. Is the content workspace product a stand-alone product? Or, does an organization need to get i4i’s server as well?
Mr. Thomas replied that the Enterprise product requires the server.
Mr. Royal then wanted to know if the server has any knowledge management functions.
Mr. Thomas responded that the server handles schemas that say how to deploy information from objects as the objects pass by. Additionally, the schemas say how to package and label objects in terms of the information they convey. In many cases, the user will find that there are well-established vocabularies that make this easy to do.
Mr. Royal asked how difficult it is to design a DTD in the tagless editor.
Mr. Thomas replied that a second year computer science student with a two hour briefing in XML was able to write the A-130 DTD in two days using the tagless editor. The editor makes it quite easy to write DTDs.
The tagless editor includes a tool that can validate XML instances. Validation does not occur while the user is authoring the document.
Mr. Christian remarked that on the policy side of this issue, legal documents are based on consistent objects that anyone can view. Would the mechanism used in the tagless editor be canonized so that the full markup of the document would be a part of it? This would let the public know that the stylesheet has created the legally approved view of the document.
Mr. Thomas replied that he was not conversant on the legal issues at stake, but he believes that certain things could be published with digital signatures for authentication.
Mr. Christian then stated that his understanding of the semantic web of content is that there are existing semantics that describe the distinctions between objects to the user. He wondered if i4i had done any work addressing this.
Mr. Thomas responded that it has not, and that such work is not part of i4i’s focus.
Discussion of XML.gov, Upcoming Meeting Agendas, and Strategic Plans
Mr. Ambur then opened the floor for discussion of the XML.gov website, upcoming meeting agendas, and strategic plans/goals of the WG.
Several WG members discussed XML and paper reduction initiatives that they knew of within their agencies.
Mr. Royal commented that when he started with the GSA, he heard of an initiative to eliminate all of the faxes that went from OMB to agencies.
Daniel Bennett (Citizencontact.com) remarked that he has been working on setting up a technology for sending company information to Congress and working on setting up messaging standards for agency-Congress exchange. This initiative did not get very far, but he is now working on XML-like standards for the same process. He will post any developments on this work on the WG listserv.
Brian Cole (NSF contractor) remarked that the NSF is using XML in many ways, primarily to tackle archiving. Currently, the agency must take electronic data out of its databases, put it on paper, and send the paper to NARA. They would like to archive without using any paper at all. NSF is just beginning this initiative, and any advice from WG members would be appreciated.
Mr. Ambur commented that next month’s agenda includes a presentation from BroadVision demonstrating the use of XML in formatting electronic records for capture and management in DoD-certified E-records management systems.
Mr. Ambur then introduced Brian Hopkins of Logicon, whom he asked to advise the WG on developing a strategic plan.
Mr. Hopkins is currently assisting the DoD’s efforts to implement XML. He stated that the WG must focus on identifying the key problems it wants to solve, how it should solve them, and what path it should take on the way to solving them. When reviewing the WG’s charter, he spotted many key points and some key holes. Currently, the charter focuses on tactics, the actions the WG wants to take to implement a government-wide XML repository. But there are many steps this (or any) WG must go through before it can take action.
First, the WG must define its vision. It must decide what the federal CIOs envision XML doing for them, in a perfect world. Second, the WG must define its mission. What will the WG do to achieve its vision? The WG charter currently defines its mission, but not its vision. Third, the WG must define its critical success factors, those things it can to do help it succeed. One of these success factors is already defined in the charter: the focus on highest-payoff areas. To define these success factors, the WG must have a clear understanding of the ways in which XML can be used in general and for government-specific problems.
What will make XML work government-wide is the willful adoption of standards by the agencies. The WG is going to want to provide something everyone who wants to participate. It must instead focus on visibility and ground-up implementation. In a perfect world, the WG could mandate standards, but in reality agencies with enough money will develop their own tags if others are not readily available for their use. Thus, the WG must focus on a ground-up strategy. In his work with the DoD, Mr. Hopkins has noticed that the agency has gotten wrapped around the wheel in trying to mandate standards. The WG must not fall into this sticky situation. Instead, it should give agencies a reason to collaborate. Financial factors could serve as a good motivator. The WG should try to use the federal CIO Council’s power to create or recommend standards, not its own debatable power. As such, the WG should start small and simply, but be able to grow with ease.
These critical success factors can be followed down into specific objectives, then into strategies, and then finally into tactics.
Mr. Campbell pointed out that there are several trailblazers with regard to government use of XML. The United Kingdom has officially adopted XML. Unfortunately, he has been unable to find any history on the politics of this adoption. But Oliver Bell of Microsoft was there for the adoption and has offered to give a presentation to the WG based on how it took place. This could serve as a good example for the WG.
Margery Reynolds (Microsoft) remarked that Microsoft has extended an invitation either to take some WG members to the UK or to bring some of their UK counterparts over to the US to work together on government adoption of XML. The co-chairs told Ms. Reynolds that they would appreciate having the opportunity to speak with people from the UK.
Stephen King (Interior) commented that, in terms of developing a strategic plan, it is often difficult to talk about focus and vision in a large group. He suggested that the WG establish and executive council of four or five people that would work on defining the group’s vision. The council would then present the WG with its recommendations.
Mr. Ambur replied that this was a good idea, and he would consider how best to create such a council.
John Cochran (GSA) introduced himself to the WG and discussed his work with trying to write XML tags for the GSA. He asked the WG if it was aware of any approved methods for writing tags and getting them accepted.
Mr. Ambur referred Mr. Cochran to Lisa Carnahan at NIST. One of the issues addressed on the XML.gov website is whether or not the WG should create a government XML registry.
Mr. Hopkins added that the DoD wants to use XML tags as a public interface. He would like to work on ground-up implementation by creating consortiums of like-minded organizations (agencies, their customers, etc…) that would work on designing an XML registry.
Mr. Christian felt that working from the bottom-up would limit the opportunities for interoperability. Each community will define what is important to it in different ways, resulting in many different kinds of tags that all say essentially the same thing. However, most agencies have data dictionaries lying around. The WG could begin its work with those.
Mike Sinisgalli (XML Solutions) added that there are many pre-existing cataloging projects. There are also many pre-existing EDI data streams. These catalogues and data streams are prime sources for defining the data that the XML registry will have to contain.
Mr. Christian remarked that it is often difficult to peel information from.
Mr. Hopkins stated that his belief in the need to work from the bottom up stems from his work with the DoD. So far, all of its attempts to mandate tags have not succeeded.
Mr. Ambur pointed out that XML will happen whether the CIOs do anything or not. So the issue is what they might do to make it happen more efficiently than will otherwise be the case.
Ms. Glace added that EWG and X12 have come together to tie in XML core components. This work is taking some time, but the tags are being worked on. The next EWG meeting will be in McLean, VA. WG members can find more information on the EWG and X12 websites.
Mr. Royal asked if this would be an ebXML registry.
Ms. Glace replied that she is not sure where the registry will be.
Mr. Hopkins recommended that links to these groups be placed on the XML.gov website.
Mr. Royal commented that the NIST website has a page that discusses its web-based standards roadmap. WG members should visit it to learn more.
Mr. Ambur pointed out that there is a "comments" option on each page of the XML.gov website. He urged all WG to feel free to use it to submit comments and ideas.
Next Meeting: March 14.
| Owen | Ambur | Interior-FWS |
| Bruce | Bailey | ED OCIO |
| Naba | Barkakati | GAO |
| Oliver | Bell | Microsoft |
| Daniel | Bennett | CitizenContact.com |
| Richard | Campbell | FDIC |
| Eliot | Christian | USGS |
| John | Cochran | GSA |
| Brian | Cole | NSF |
| Mike | Douglass | Sequoia |
| Jessica | Glace | LMI |
| Karl | Hebenstreit | GSA |
| Bruce | Hoglund | DLA (Consultant) |
| Brian | Hopkins | Logicon |
| Jim | Hunt | GSA |
| Dan | Jansen | NARA |
| John | Junod | Navy |
| Stephen | King | Dept. of Interior |
| Jun | Lee | UNISYS |
| Marsha | Misenhimer | U.S. Senate |
| Bob | Moe | EDS |
| Joseph | Morey | OGE |
| Jeff | Poisson | i4i |
| Lex | Poot | DTS |
| Margery | Reynolds | Microsoft |
| Marion | Royal | GSA |
| Gretchen | Schlag | DTIC |
| Chris | Sciortino | XML Solutions |
| William | Selfridge | Data Networks Corp. |
| Mark | Shaw | GAO |
| Mike | Sinisgalli | XML Solutions |
| Keith | Thomas | i4i |
| Steve | Vineski | EPA |
| CC | Wang | USCG |
| Jan | Wendler | GSA |
| Theresa | Yee | LMI |
Certified 508 Compliant 03/30/01