GSA
Headquarters
18th
& F Streets, N.W, Room 5141
Washington
DC 20405
Please send all
comments or corrections to these minutes to Glenn Little at glittle@lmi.org.
Mr. Owen Ambur
opened the meeting by introducing himself and briefly explaining the day’s
focus:
Mr. Ambur: We may as well get started with introductions and announcements. I’m Owen Ambur of the Fish and Wildlife Service. I’m the co-chair of this Working Group.
[The attendees briefly introduced themselves.]
Mr. Ambur: Does anyone have any announcements or comments they’d like to make?
Mr. Marc Le Maitre: I have a short announcement. The XNS Protocol for digital identity management, which was presented to the XML WG late last year, has now been licensed to XNSORG, a non-profit body, on a royalty-free and open basis. The XNS protocol has been the subject of some interest by the FTC and Commissioner Mozelle Thompson. XNSORG is currently working to submit the XNS protocol to OASIS.
Mr. Ambur: I also have a few announcements:
With us today are Pat McCreary and Bob Greeves from the Department of Justice to update us on the activities of the Global Justice Information Network.
Mr.
Pat McCreary
Department
of Justice
Mr. Pat McCreary: Good morning. I’m Pat McCreary, and I’m the Designated Federal Representative at the Department of Justice’s Office of Justice Programs (OJP). The IT Initiative at OJP works with a body called the “Global Justice Information Network” (Global). They have a website that tells you everything about them at http://it.ojp.gov/global/
The unique thing about Global is that over the last few years, their goal is to help make recommendations regarding policy and standards, to help improve information sharing amongst the justice community. Through this cooperative effort, we’ve gotten to know each other, walls have fallen down, and working groups have arisen in areas including information quality, privacy, security, and infrastructure standards.
One of the great things about this group is that they’ll go to lunch, for example, and then two members will come back and say they have an XML transmission standard to help the entire country. It’s powerful, and it’s why we’re here today. Bob Greeves has been instrumental in forming this, and he’ll talk to the Working Group today.
Mr. Bob Greeves: One group that is not a formal Global partner, but which is a very real contributor, is the private sector. Our private sector working group and the IDIS institute have been very active in developing this.
Mr. Ambur: OMB was required by law to issue information quality dissemination guidelines and, in turn, OMB has required each department has to issue its guidelines. Conceptually, information quality can be specified in terms of XML metadata elements. To the extent that such metadata are common across agencies, we may be able to learn from our colleagues in the justice community. Have you made much progress in that area? For example, one indication of the quality of a record is whether it has been digitally signed.
Mr. Greeves: Not much. One area is criminal disposition.
Mr. McCreary: You don’t have to be a Global member or Subject matter Expert to be on the Working Group. Global welcomes participants. It pays for travel and expenses (not your time). Please feel free to contact me if you’re interested.
Mr. Greeves: If you go to the second page, there’s a wiring diagram.
Slide 5 [Justice XML Data Dictionary Schema Evolution Plan]: This is the model that reports everything that is, has, and will happen in the next 12 months, with respect to the Justice XML Data Definition initiative.
Look at the “RDD” column. It reports on work done to date to reconcile four communities. There’s one missing from the document. There are four communities in the justice world:
These communities are represented in the cloud in the picture. They’ve gotten together and developed a data dictionary of elements that are common to at least two of those communities. The RDD represents the aggregation of that work.
The dictionary as compiled (DJ Atkinson behind me did all the technical work) under that first reconciliation was just an Access database with a listing of tags, attributes, etc. While informative, it won’t be of value to future users if it’s not built into a schema approach. The next step is to build it into a standard schema, which is DDS2.0 (second column). We did it with the help of Georgia Tech Research Institute (GTRI). They helped convert the Access database into a recognized schema approach under 2.0.
As soon as we did that, we got information from this body [CIO Council XML Working Group] in terms of the standard approach to doing some of these things. I gather it may have origins in some of the work the Navy did—their Developer’s Guide. Once we looked at the Developer’s Guide, we identified the fact that we were doing lower camel case for tags, and yours were in upper, so we shifted from lower to upper to be in synch. That’s represented here by DDS2.1
At the bottom, you see a bi-directional arrow. It represents a translator function written by GTRI. It goes either way—from upper to lower or lower to upper camel case—for people who prefer to develop the way with which they’re familiar.
As of right now, our standards management structure, which includes Global.org, which Pat spoke about, has approved this work. We’re counseling people undertaking work in the justice community that they ought to use this core data dictionary in their efforts. At the same time that this work is frozen in time for a moment as the production version of the Data Dictionary, we’re moving forward over the next six months to move the schema approach to the Data Dictionary to an object-oriented approach. That’s represented by JDD3.0
We’re probably going to spend time at the meeting this afternoon on the technical approach. We’re looking at a six-month time line. If you drop down to Slide 6, you’ll see a pictorial of the timelines. There’s a 90-day timetable starting last Monday, going up to a November timeframe where we hope to have a version 3.0 of this available. Any of you in object orientation realize it’s non-trivial to hammer out agreements on what the structure should be. For the 60 days following, the work will be vetted to people outside of this, and some migration tools will be developed, then in the January timeframe of next year, we’ll do beta testing to see if it works with the software approaches in peoples’ applications. Then we’ll take about 30 days after the vetting to make any necessary adjustments, then have the beta version in 30 days.
At the same time, we’ll try to add in many more data elements. Currently we only have about 300—just the overlapping ones. We’ll add in some of the individual ones that don’t overlap. We’re also reaching out to state and local organizations, which are using many more data elements with overlap in these four communities. We’ll specifically target the Crime.net folks in seven states, around a standard approach to object orientation that might bring in up to 1,500 raw data elements. There’s a loosely structured IT organization representing initiatives in Oregon, San Diego, Florida, South Carolina, and the Tidewater area of Virginia, that will bring 800 more to the table. Connecticut is underway. They have several hundred. Maine and Wisconsin have some too, so hopefully six months from now, you’ll be looking at a significantly expanded dictionary.
Back to Slide 5—the fourth column is a plan, over the next 15 months, to move in the direction of an RDF. Assuming W3C finalizes its work (the Resource Definition Framework), this is the point where we build additional relationships into object orientation. It’s a moving target with W3C finalizing while we’re building these relationships. That’s it in a nutshell. Before we go on, are there any questions?
Mr. Michael Jacobs: The core objects—are a large percentage applicable to other communities outside Justice?
Mr. Greeves: Absolutely. That’s a perfect lead-in to this afternoon. Paul will show how we did “working person” and “address,” etc. The standing joke is, “How much value is there in identifying the object orientation for address if the Post Office down the line comes up with one?”
Mr. Jacobs: I think DoD did a presentation on their efforts over the last few months. They have communities in registry, and then they have an enterprise namespace that’s not yet populated. The first thing they were going to try was going to be postal address. It’s a difficult effort, that hasn’t yet come to fruition. We couldn’t find an authoritative source on where to get the postal address. It’s not the Post Office, believe it or not.
Mr. Greeves: To date, we’re still working within the criminal justice community. The Global Advisory Committee is very well structured, organized, and bought into. We realize there’ll be negotiation, but this is a good starting place.
Mr. Marc Le Maitre: Can you describe how these objects can be accessed by external parties, access control, privacy, stuff like what we’re doing?
Mr. Greeves: If it’s in an enterprise, it’s called enterprise management. You may want to sit in this afternoon, because Paul is going to go through a person object identification application. It shows relationships.
Mr. Le Maitre: My area of interest his how you create a meta object out of disparate components.
Mr. McCreary: This is well funded. We have top- notch people on it. There are many partners already implemented around the country. We do keep a database of the efforts we know of around the country. Is it worthwhile to give you the email address?
Mr. Ambur: Sure. If it is in the presentation, I’ll be posting it.
Mr. McCreary: It’s a database of Justice and state and local level public safety efforts. Many states are anxiously awaiting our developments.
Ms. Diane Lewis: These gentleman talking about state and local implementations are here as part of the federal point of view. There’s a parallel effort under OASIS. I’m a member of that with Justice (Justice is a member of OASIS). The idea is, as this goes on, court filings (I represent the U.S. Attorney federal cases) this is a concern for my customer—because if a criminal is arrested locally and charged with federal charges, this sense of identity has to be carried through the court life cycle, so other efforts are springing from this.
Mr. Greeves: Diane pointed out that we are doing a lot in partnership with OASIS. It used to be LegalXML. There’s an integrated Justice work group under OASIS. Diane is much a part of that. The state and local side of that is looking at prosecutorial reporting, judgments, sentencing, warrants, etc., using this dictionary in those areas. We’re also working with NCIC and other FBI systems to bring them into our CIO world.
Slide 7 [JXDDS 3.0 Work Plan]: This gives you an idea of what we’re pursuing [List of Justice XML efforts]. I want to close with an offer on our part—put something on the table, to form a partnership or collaboration with XML.gov, using as a pilot base version 2.1 or an upper camel version of our schema and do it in conformance with the DON Developer’s Guide and ISO 11179. We’re proposing to carry out a role like an executive agent role; operate the Data Dictionary via XML.gov. I don’t know whether it’s decentralized, or your repository. We’d have to hammer it out, and make this available to areas where there are elements that are helpful to the rest of the world.
Mr. Jacobs: Do you have your own registry/repository stood up?
Mr. Greeves: We have a Justice registry. It’s broader than just XML. It has lots of stuff in it. We’re going to put the XML in it. It’s not there yet, but will be shortly. We want to work through this group and NIST to identify namespaces to set aside for the Justice world, and work to do it in the right way.
Mr. Ambur: That’s exactly the type of activity this group was chartered to engage in. I didn’t want this group to be a check point or choke point for people like the Justice community. I do want to facilitate partnerships and build communities of interest where there are common elements that do span namespaces, legacy systems, or whatever you want to call them.
Let’s move on to the next presentation. Bob and Pat will be here for this afternoon’s meeting.
Mr. McCreary: Owen, I’d like to thank you for your support. Owen attended an industry work group meeting, where he extended an olive branch to us, and I’d like to thank him for that.
Mr. Mark Miller: HRXML is finalized in postal address, and has very robust schemas, in case anybody is interested.
Mr. Ambur: With that, the next presentation is from Shanti Rao of Raosoft, Incorporated. Shanti will be discussing XML-based E-forms, interactive XML editors, and the XML.gov brainstorming survey results.
In our dialogue, I advised Shanti of the brainstorming session following the GAO XML Report, and of the fact that we didn’t finish ranking the ideas that were raised. It seemed like an opportunity for Shanti to demonstrate the capabilities of the technology with some real data from the business processes in which we’re engaged.
The brainstorming ideas are timely and relevant, because the current Working Group charter expires at the end of next month. Next month’s meeting will focus on how to revise and extend the charter. With that, Shanti is going to give us the results of the survey.
Mr. Shanti Rao
Raosoft, Inc.
“Electronic Form System”
Mr. Rao: My company, Raosoft, has been using XML to conduct electronic surveys since 1997. In the process, we learned some things about technology and XML in general.
Slide 1 [The 21st Century]: It turns out that some issues of social engineering go into getting people to use systems. We’ll talk about that, and we’ll talk about the electronic form we use with OASIS. There are also implications of complexity we’ll talk about.
We now live in the future. Some things we predicted have materialized; some haven’t. We don’t have flying cars, but we do have microwaves and video conferencing, and we’re almost there on the paperless office. George Jetson has a paperless office and videoconferencing. Look at what he has—special purpose hardware. That was the vision of the ‘70’s. Instead, we have general purpose hardware. Intel put all of its money into one chip to do everything. It turns out that in computer science, that’s almost always the right answer.
Slide 2 [Objectives of the Paperless Office]: In order to have continuity, we need a concept of history, and we need to add new rules so we don’t make the same mistakes again. We use electronic forms as a way of communicating. This goes from person to person, and builds on past knowledge. Electronic forms have to take this into account.
There are two broad categories of how people use computers:
You’ve seen our software if you saw the OPF Human Capital Survey. We wrote the SSW for that. This is the first step to get early adopters to use the technology.
Slide 3 [Properties of Forms]: The most handy part of electronic forms is that they pull information together and hold it in one spot; therefore they have high archival value and one searchable index, which is why we want to move to electronic forms—have access to archive data, statistics, study our past.
Slides 4-8: Skipped.
Slides 9 & 10 [Fundamental Laws]: There are some fundamental laws that apply here. Hofstadter said that any computer project will take longer than you think. There’s also Moore’s Law. Gordon Moore is one of the founders of Intel. He predicted the number of transistors, and therefore computing power, and therefore complexity of software, is going to double every 18 months. So anything you do that’s specialized now will be worthless in two years. This is where we realize you want general solutions rather than specialized solutions. The last thing is that last year’s browser won’t work with this year’s software.
Slide 11 [XML]: XML works well for most purposes. It’s good for most things, and users don’t want to generalize, so XML is the way to go. With definitions, schemas, reports, transactions, and structures, it works well.
Some things that don’t work well are things like storage and word processing. Word processing needs relationships, so doesn’t work well with XML.
Mr. Le Maitre: Can you comment on why you say it stores poorly?
Mr. Rao: I should rephrase that. It doesn’t store poorly; it makes a poor random access database.
On the Electronic Form System—as far as we can tell, the economic decision was to have a specific data type on our software for every type of human interaction decision. We thought we wanted to separate the data description from the presentation description. That’s very noble and acceptable, but it turns out to be wrong. People in the IT field who want data and presentation separated—there are many people for whom it works well, but a lot of people think that designing a form is the same as writing a Word document. There it doesn’t work well, so there’s a conflict.
A lot of people write software to do electronic forms on databases. I do that. Some people write software to see it for when it’s printed. I think adobe is winning that model.
The other thing we do is group inputs according to what individuals think of as one question. Though “Date” is comprised of “Month,” “Day,” and “Year,” people think of it solely as “Date;” just one data item.
Slide 12 [Example/Electronic Form System]: Keep it simple and successful. It’s not a big deal, but EFS is not HTML. What we said is, “If you want to add something new, you stick it in as an option somewhere in the form.” It works such that if I write a form, and send it to someone using software from another vendor, they get most of the functionality that I have. It doesn’t look the same on the Web browser, but the information is there.
An advantage of using XML
format is that you can use a searchable file system with XML files indexed, and
you can use a general-purpose editor. This is the document for the XML Working
Group survey [Mr. Rao displayed an example of the survey form]. You can edit
the text with this. It’s a general-purpose editor. I didn’t do any programming, and I didn’t pay for the
software. It works well, for without the cost of a full-fledged user interface.
[Back to Example slide] What sold us on XML was, the Air Force was doing a “quality of life” survey with 300-400 questions. How do you write a 400-question survey and test it in a few weeks? We took seven people, gave each a part, and spliced together their XML files in a text editor. It’s not quite a client/server multi-headed environment, but it saved a lot of time. Another thing you can do is write little programs to do small changes. This lets you build up the system without worrying about unpredictable interactions.
Translation is a big deal. More on that later.
There are some disadvantages to XML, though. Once you’ve chosen XML, it’s hierarchical. For example, let’s say two questions have the same set of responses. If I change one, I want the other to change. It gets complicated to do that with this data model. We decided to not do that, and instead to rely on disc compression to take care of wasted space. There’s always a tradeoff between duplication of effort and complexity of interface. Here, we chose duplication of effort.
Slide 13 [Electronic Form System/Interactive XML Editing]: It's almost always easier to write a new computer language, in which a problem is trivial, than it is to solve it in a general-purpose language. When we encounter unusual problems, we use a scripting language that reads one format and sends it to another.
Mr. Ambur: Are you going to talk about the XForms standard, which you mention on that slide?
Mr. Rao: That’s more political than technical, but if
you’d like I can come back to it later.
To sell this format, you need an interactive editor. First is that an interface with a tree works well. People like it, because the file system runs by the same metaphor: it’s very powerful. [Mr. Rao displayed an example of the Electronic File System tree interface.] On the left side, we have a list of objects in the file; on the right side, we have a property sheet of the one object you’re working with. Obviously, you want to have a property sheet. Anytime you have a new feature, it costs less to add to a 5-layer deep property sheet than to the front interface. This makes it hard to find that feature, so we have a system where the last-used property shows up on the front page. [Mr. Rao demonstrated this characteristic.]
Slide 14 [Interactive XML Editing]: Some of our most successful products have cost us the least. It pays you to do small features, for example, the feature that renumbers questions, or converts tags in the wrong color to something else, etc. If you spend development time on an interactive editor, don’t develop a really complicated one, because the hardware changes in two years. Focus on small features.
[Back to the tree example] There’s a lot of stuff on the right [The screen was split into two sides—the tree on the left and tree descendents on the right.], but the left side is fairly clean. If you look at the way people communicated 50 years ago, e.g., a Winston Churchill speech, we instead like PowerPoint slides now. We’re conditioned to look for three-point bullet lists.
If you have an XML editor, you’re working with a database. It’s not the same as a word processor. In a word processor, you have “undo,” in a database, when we delete, we send it to a trashcan rather than wiping it out. Instead of undoing in order, we save to a trashcan. It also means you can put it back anywhere else. One big problem in XML is that it’s not really defined on how you talk about character encoding—for example, the line feed on a character tag. It’s not allowed, but it’s what the customer wants. How do you embed one XML into another XML? It’s ugly. It pays to think about it ahead of time.
Slide 15 [XML Editor/Systems Engineering]: I mentioned that XML files can be edited in a text editor, and are therefore “field-serviceable.” The text editor gives you a lot of flexibility. Once you have an application, it’s easy to make little programs that let you do whatever you want. I predict that, because of how websites work, if I wanted to reproduce this form editor as a Web window, I’d have to reconfigure to be able to handle the interactions. It’s difficult to do through a website, so I predict that platform-specific applications are going to be continue to be popular for the next few years.
There’s a tradeoff between complexity of interface and duplication of effort. It’s always true. For example, everyone here has a laptop that wants DC power, but the wall gives you AC. On military aircraft, you have a power panel with DC voltages of every type. Commerce has found that lots of transformers will sell, but a three-pronged plug works well. That’s a tradeoff there.
You also have to ask how long the software will last. You can only reasonably expect two years. You have to plan for obsolescence.
Slide 16 [Human Interface]: How many people in America speak Persian? In Southern California, the Persian Yellow Pages are very thick. There’s a huge community of former Iranians. They have their own community, and an alphabet that reads the long way. It’s not mappable onto ASCII. Up until now, they’ve been economically disenfranchised. We didn’t care much. Times are changing. There are entire cities in Orange County where they only speak Vietnamese. They’re part of the economy. We have to provide forms with which they can interact.
The initial claim of XML was that it could handle all languages. The problem is, “How many languages can I write in a file?”
Slide 17 [Example/XML as a Database]: I’m going to have a “Thank you” page in several languages. Japanese doesn’t work as well because XML isn’t a Unicode system.
Mr. Brand Niemann: What did you say?
Mr. Rao: Almost all XML processors are ASCII. It’s getting there, but if you start at the beginning, you have to commit to one way or another. XML Spy is useful, but a lot of tools aren’t. Even if you do have Unicode, it’s not necessarily appropriate. For example, Russian has a complicated system that determines which word’s appropriate. Cultural idioms weren’t put into the XML database. What’s appropriate at the beginning might not be good later. We’re better off having many files with a single language than one with many. Because of cultural issues, and the fact that no one can speak all of the languages in one database, it works better with many.
So XML is a database. There are make changes:
In both cases, it means one program has exclusive access to the database, so an XML file doesn’t make a great database.
Unidentified member: There a re a number of XML databases on the market today
Mr. Rao: And they have a number of problems.
Mr. Niemann: There are a number of advantages. The largest is in the U.K., containing 400 million objects. It’s clearly the best solution.
Slide 18 [Complexity and Robust Control Theory]: Whenever you have an automatic procedure algorithm, OAW of Electrical Engineers says, “The more clever and complex, the better it works. An example is hotel cases. The more you can take into account the more reliable it’ll be. The tradeoff is that, as you improve your performance, the robustness decreases, because of the consequences. If something unexpected comes along, it performs worse. The better it performs, the worse it is when it fails. It’s a fundamental aspect of nature.
Another example—if you’re a purchaser and you need more pencils, you buy pencils, and you have an expectation as to when they arrive. You gauge your ordering upon a combination of the expected outage date and the delivery schedule. Then the delivery truck breaks down, and the pencil ordering system goes awry. It’s called a “Feedback Loop. Its success depends upon how well it’s built, disturbances, time delays, and measures.
For the next four or five years, if I have an XML database that makes decisions, and I publish them to another database, then someone gets results from that and makes a decision, which feeds back to the first database, and on and on. It’s possible to have a very large problem. Vicious cycles are something we always worry about, but we haven’t really addressed network interaction problems before.
Slide 20 [Robustness/Prevent Computers From Running Amok?]: How do you prevent computers from running amok? Have a database that keeps a record of information in and out, and a guy who uses your information has that same record; then there’s a way to make a map of where the information flows. Simple pieces are better than complex, and it’s easier to find unexpected interactions.
Mr. Niemann: Are you talking about the requirements you’d like to see, or do you have a commercial one—and how does it compare to something like XMetal or SoftQuad?
Mr. Rao: These are decisions I’ve put into XML editors on our products. They’re the kinds of things you should be looking for. It might be possible for the government make ad hoc interfaces for whatever interface you have. These are things to consider.
Mr. Niemann: Create an interface to an editor, or to forms?
Mr. Rao: When you create an XML editor these are things to consider.
Mr. Niemann: Forms as the front of the database?
Mr. Rao: That’s true, yes.
Mr. Ambur: Your mentioned something about your comments regarding XML forms being more political than technical in nature. What did you mean by that?
Mr. Rao: My complaint is that they haven’t finalized the standard. XForms attempts to put into it many features of electronic forms that many of us have had to solve by ourselves. My competitor has as the goal of XML forms to commoditize electronic forms; however, the companies in the business have continually added features, data formats, input devices, and ways of organizing. Unless XForms keeps pace, you’re going to lose out on innovation
Mr. Niemann: The W3C process says you have to have two successful implementations, so we invite vendor input.
Unidentified member: They’re also trying to integrate the data and logical presentation
Mr. Rao: My prediction is, we’ll have and XForms program because that’s what everyone wants, but I predict it won’t be enough
Mr. Le Maitre: At the Mortgage Bankers Association, lots of paperwork has to be transformed into X format, using this gentleman’s process.
Mr. Rao: Philip Greenspun, an MIT professor, proposed that years ago. It takes a long time to standardize.
Mr. Barry Schaefer: The entire industry may be missing the fact that the form is an artifact of the limitations of a rectangular page. We have a complex forms industry because they have to be on paper. I wonder whether the way to look at information on forms is, when we go to a screen without the special limitations, many of the complexities go away. We’re all getting used to forms we fill out, where the form is only a vertical series of questions. I wonder whether, rather than focus on forms, we should focus on bindings between information we store and a way of representing it visually?
Mr. Rao: That’s a very good comment. For example, it takes a completely different way of creating the form for a palm pilot from the way it’s done on paper. The data model is different.
Mr. Schaefer: Maybe model the content rather than the form, It’s up to the deliverer of the form to model how they want it, so we focus on how the content is organized logically. For example, a PDA can’t display much at the same time. It’s up to the owner of the form to decide how it shows on the PDA.
Mr. Rao: I think that’s where the industry is headed, but there’s one economic obstacle. There’s only one person who buys software to design the form. People in the industry want to have the control features to have alignment, sentence breaks, etc. The way we originally designed the electronic form system is the way everyone would want, but it’s not what the market supports.
Mr. Niemann: The XForms scanner is more complex. It separates content from presentation and functionality, without scripting of HTML that makes it browser-specific.
Mr. Rao: Right, and I’m behind it, but I don’t think it’s going to make money.
Mr. Ambur: It won’t unless the market demands it, and I think we should demand it. To the degree that I have any ability to influence the outcome, I’m going to insist that government agencies use standards-compliant products.
Mr. Niemann: There’s a contract underway to bring that technology to the Air Force.
Unidentified member: It’s a matter of critical mass. When it comes with it, they’ll use it.
Mr. Ambur: As a matter of principle, if government managers are buying into the notion that taxpayers should be locked into using proprietary products, they should look elsewhere for employment.
Mr. John Dodd: Even your survey shows that interoperability is what people want. Don’t you have to follow the survey?
Mr. Rao: I follow the market.
Mr. Dodd: All the economic growth right now is in the federal market. They’re all coming to the federal market because it’s the only place where there’s growth.
Mr. Ambur: Even in the private sector, look at the scandals that have made so much press lately. The basic problem underlying all of them is poor record management practices and systems, and one of the challenges that GAO recently identified in its E-records management report is the use of proprietary file formats.
OK, we need to go to the results of the survey.
Mr. Schaefer: One final comment—one place that isn’t looked at much is the world of structured information that started 30 years ago. Many of the concepts, here and in Shanti’s presentation, the SGML world has dealt with for a long time. There’s a good schema that thousands of companies used to use in about 40 languages. Because XML is new, we don’t want lose sight that XML is a step in the chain of structured information for which there are many cogent solutions. It would be a mistake—like forgetting how cars have been made for 30 years if we have a new type of engine. It would be a mistake and cost the government a significant amount of money. XML editors are new, but structured information has been around for a while. We need to look at it longitudinally, understanding that some of the final processes will be different for XML.
End presentation.
Mr. Ambur: Shanti, can you give us the results of the survey?
XML.gov Brainstorming
Results
Slide 3 [Survey Summary]: The summary broke the questions into categories; you saw the initial results on the LMI website. There is something to them, because some categories are more popular than others. I took all the questions in all categories, averaged them together, and this is the percentage of people who responded to each item as a high priority. For each question, respondents have the option of categorizing the priority as “high,” “medium,” or “low.” We counted the “highs.”
Mr. Atkinson: Not the percent who were favorably inclined, but the percent who felt it was a high priority?
Mr. Rao: Yes. I need to explain that.
Slide 4 [Governance]: This shows, for each category, the percent of people who felt it was a high priority.
Slide 5 [Governance, continued]: The next slide is a breakdown of the high, medium, and low priorities for each category. It appears that the “XML Seal of Approval” is widely regarded as a low priority, and a standard set of XML guidance is considered a high priority by many.
Slide 6 [Registry]: “Develop project plan for XML registry” is the most popular. It seems to be a kind of cutoff point, using the combination of high and medium as identifying things we should do. We didn’t tell the respondents what high, medium, and low meant. We were hoping that the average respondent would decide that.
Slide 10 [Survey]: The “survey” category was low all around, but there was a big medium. Many gave it a medium priority, but not many gave it a high priority, with the exception of identifying best practices.
Mr. Dodd: Is there any way of identifying the type of correlation between the responses for similar questions?
Mr. Rao: Yes. [Mr. Rao displayed a sample correlation on the screen as he performed this interactively with the survey database.]
Unidentified member: There were 18 respondents?
Mr. Rao: The total was about 20, but not all answered every question.
Mr. Ambur: We need to recognize that this certainly was not intended as a scientific study, nor is it binding upon us in any way. However, hopefully, it provides some useful information to help us focus our thinking.
End presentation.
Discussion
of Brainstorming Ideas with reference to XML Registry & Extension of
Charter
Co-Chairs
& Participants
Mr. Ambur: As you’re aware, our charter lapses at the end of next month. Presumably, we’ll be rechartered. Meanwhile, the Architecture and Infrastructure Committee is undergoing reorganization. Lee Holcomb has gone to the Office of Homeland Security, and Debra Stouffer and John Gilligan, CIO of the Air Force, have agreed to co-chair the committee. They may have some direction for us at the committee’s regular monthly meeting tomorrow.
Our discussion today will better inform me to prepare for tomorrow. As I mentioned, the survey and brainstorming were to better-focus our decision-making. I forgot to ask Shanti whether people used the comment fields for additional comments that weren’t originally mentioned during the brainstorming session.
This is an opportunity to be heard on the scope of what we should do. The brainstorming results might provide a good framework or outline for our discussion. I briefly scanned them, and it’s interesting to look at the white bars as well -- things people thought were of low priority. [Mr. Ambur identified several that many people had graded as low priority, including the “XML Seal of Approval” and the notion of establishing a full-time centralized management activity to include XML. He also related the background of the events leading up to the survey, including the GAO Report and the subsequent brainstorming session.] With that, I’m open to any suggestions.
Mr. Ken Sall: I have a comment on identifying best practices—things coming out of the Developer’s Guide on specific aspects of XML development. It would be helpful to break it down into two best practice areas, like a general area describing when and where XML is appropriate, versus, XSLT, namespaces, etc. I’m sure there are government projects focused on some, but not others. Some people might have particular expertise on a subset, and be in a position to suggest best practices in those areas.
Mr. Ambur: So for this Working Group to add value, we’d have to identify who those people are, identify expertise, and ask them to draft best practices.
Mr. Roy Morgan: The overall goal of the Group being helpful is to ask, “Who needs help, and what do you need?” Some in professional management positions might need top-level guidance—around the policy level. Then there are people who need to put software on a disk, compile it, and make it work. That’s at a different level from the management folks. It’s helpful to think of whom we help and what they need. We can hold workshops; write policy; think of a variety of the things we can do, and think of the pieces we can contribute now or soon.
Mr. Ambur: I’ll be at the Architecture and Infrastructure Committee meeting tomorrow, at which the primary stakeholder audience will be John Gilligan and Debra Stouffer, so that frames my context a little. Then there are the people who attend this meeting, along with those who visit the XML.gov site.
We have the registry pilot underway, and Booz Allen is drafting the business case for the operational version of the registry. Regardless of the role of the Working Group, the need for the registry will continue to exist. The question for us is, “How can we best contribute?” I noticed that setting up a separate sub-group for Web Services came out as a relatively low priority on the survey, but only 17% of the respondents said it is of low priority. Incidentally, Brand has agreed to lead a Web Services group, and that will be discussed at the AIC meeting tomorrow.
Mr. Niemann: One comment I had regarding the results is, I know how important the wording of the question is. For example, if the word “discussed” were changed to “do,” the answer might have been different.
Deb Stouffer has two principal tasks at EPA:
Kim Nelson [EPA CIO] is also very supportive of our XML work.
The third thing to mention is I’ve provided a lot of training. My best experience so far is with the Small Business Administration, with Diane Gannon. She called me up for training on the Business Compliance One Stop. I worked with her. She said, “Who should be there?” I said, “All your key people.” She had 25 people there all three days.
I said it was a good experience, and they were very pleased with that, so I think there’s a real need for education and outreach, starting with why it’s important. People begin to use on it on their web pages and other areas, and they figure out how they can use these Web Services. If you want to build interoperability without replicating databases, and you don’t know how to, this outreach and beginning needs to be replicated. The only way is to find people to step up, then work with these E-Gov initiatives to make it happen. That’s what I’d say if I were talking to the CIOs. Each Agency needs at least one representative to take it on.
Mr. Lex Poot: As separate counsel for the small agencies, we need to do some outreach there as well.
Mr. Niemann: At FedWeb in the spring, we did a tutorial. We’re planning a fall thing. I think the more we can do, the better we’ll be.
Mr. Ambur: It sounds like the training is almost a workshop, since your take their data and put into a Web Service.
Mr. Niemann: It’s not driven by my things. I want them to set the agenda with the topics they want.
Mr. Ambur: I’m glad to hear that you’ve conducted training for the Small Business Administration. EbXML can be summarized as EDI for anyone, and it is small and disadvantaged businesses that can benefit the most. So it’s good to know that SBA is looking into how best to capitalize on the potential of XML to assist small businesses in this country.
Mr. Le Maitre: I have an observation and a question. We responded to E-authentication RFI from GSA last week. I wondered whether the XML Working Group has had involvement in that. It’s a specific manifestation of a project where adherence to best practices is imperative. The problems of not having a common schema and naming convention causes problems farther down the road, but if you can’t get into the system in the first place, you don’t even get a chance to see whether objects cause a problem in other repositories. I wonder whether people from this group should be in there to participate.
Mr. Ambur: Absolutely. Incidentally, I have been in communication with Judy Spencer from GSA with respect to the possibility of scheduling a forum on SAML in relation to the E-authentication project.
Mr. Le Maitre: Part of Bob Greeves’s initiative will be aligned with the E-authentication, in questions like, “Who can be given access to the data?” It seems to have a connection there as well.
Mr. Ambur: You mentioned that you have to authenticate to get in, and we also have to have appropriate standards to keep people out. The lack of standards to enable both functions is a risk factor, because people have to set up their own means of meeting those requirements.
Mr. John Kane: Can we look at the enterprise architecture model?
Mr. Niemann: We’ll present something at the Federal Architecture Work Group on the 17th of September. We have a pre-presentation at this point.
Mr. Ambur: John is heading up NARA’s E-records management schema project under the Adminstration’s eGov Strategy. One of the two proposals that led to the formation of this group was the potential to use XML metatags to classify and manage E-records Governmentwide. The question is, “How can this Working Group best contribute to projects like that?” The XML registry will have a close relationship with the E-records management schema, for example.
Mr. Kane: The intent is to get involved in XML uses, particularly metadata.
Mr. Dodd: Is there a way this group can map to each of the 24 E-Gov projects, and Homeland Security, to say, “These are the projects we’re influencing, and these are the agencies and players at each level?” If you’re getting more coverage and you’re growing, then it’s good.
Mr. Ambur: My first thought is that OMB has tried to identify solutions architects for the 24 projects. They’ve had trouble getting that many. Marion is one of them. It seems like the solutions architects might be good candidates to map the potentials of XML to the particular requirements for the E-Gov projects.
Mr. Dodd: The first thing is outreach.
Mr. Niemann: We’ve linked to four of the 24.
Mr. Ambur: The first time I saw the list at an Architecture and Infrastructure Committee meeting, I said, “From my perspective, this is a set of XML schemas.” When the registry software is proven, it seems like it would be reasonable to expect the project sponsors to register their schemas and make them available in the repository.
Mr. Dodd: I was concerned that in the Industry Advisory Council, they didn’t want to address data- and information modeling. We took that as a challenge. We said, “You have to address metadata. You can’t address integration if you don’t know what to share.” My concern is these are the tough issues; that this is what you have to address.
Mr. Ambur: The Federal Enterprise Architecture folks have released the high-level Business Reference Model, but what value does it add other than some academic sense of the big picture? Until you get to the data elements that support the business process, it doesn’t seem to be that you’ve accomplished very much. If you look at the Federal Enterprise Architecture website, they have a placeholder for the Data and Information Reference Model, but they haven’t yet established a time line for producing that model. Thus, one question is, “Can the registry become the data reference model?” Or, in other words, “How can we best support the FEA effort?”
Mr. Dodd: It’s actually a management thing, because people have to address it. Like Bob Greeves said, getting people to agree is what we need. It’s getting the trust that’s important.
Mr. Ambur: Face-to-face contact seems to be an important part of building trust, but it is also difficult to involve everyone who should be consulted. So we should strive to facilitate and leverage consensus-building, such as through use of the registry.
Mr. Greeves: One of the most effective things for us is, when the face-to-face had negotiation, we reduced it so there were only two or three people from each organization involved. By the way—they agreed, and the rest took what they said. On the 24 Quicksilver projects and agencies, I encourage you to think on a broader scale—the State and local level. Almost all agencies have something at the State and local level. State CIOs are engaged on their own. They approached us on how the enterprise architecture should be done. We’re looking at the differences between the federal and State approach. Hopefully we can hammer it out. There’s a need for us to think about XML.gov on a national basis, rather than a federal basis.
Mr. Ambur: Absolutely, not just federal but government at all levels. And not just in the U.S. but internationally as well. The issued is how best we can focus our efforts to leverage what’s going elsewhere -- “thinking globally and acting locally” while at the same time thinking within the scope of our resources. In that regard, I encourage this group to think expansively. We won’t get resources unless we ask, specifically, for example, for the registry.
Mr. Greeves: We’re interested in collaboration, because the visionaries in our world are realizing that they can’t operate well without interfaces between [the Departments of] Justice and Education, Social Services, Health and Human Services, etc. So it’s important in the long haul to develop standards in collaboration with other communities, because we have to cross those boundaries.
Mr. Ambur: That summarizes the essence of this Working Group and particularly the registry. The issue is how to make them as useful as possible, with as little needless overhead as possible.
Mr. Greeves: You have to identify a subset of leaders to come to the table and talk about these things. You’re going to have to get other Agencies to come to XML.gov or some table.
Mr. Ambur: One of the brainstorming ideas was to require each Department or Agency to designate an official liaison. In the past, I haven’t felt that we’ve reached a sufficient stage of maturity to ask for or to expect that. However, at the point when the usability and scalability of the registry have been proven, it would make sense to have a policy to require each Agency to register and/or subscribe to the XML data elements and schemas used in their business processes.
Mr. Greeves: When we were involved, we found that it wasn’t the XML experts that needed to be there, but the people who had the information that had to go into it.
Ms. Susan Turnbull: Maybe we should think about “XML-izing” project information. In effect, it would help us find the people who are the key movers in the projects. It would probably be used as an introduction to XML, as a way of finding the projects that’ll help people find one another. The person that spoke at our workshop yesterday from another country said that’s what they’re doing.
Mr. Ambur: That reminds me of the [OMB Circular] A-11 budget process and the 300 Exhibits that are required for major investments. One of the brainstorming suggestions referenced the A-11 process. OMB could have a role in that. Any project with data should have an XML schema.
Ms. Susie Adams: We’re [Microsoft] XML-enabling the 300’s and 353’s. You’ll be able to submit any XML document.
Mr. Ambur: While we’re at it, Susie, do you have any updates on BizTalk and Microsoft for this Working Group? I see that UDDI is going to OASIS.
Ms. Adams: The general perspective is to conform with standards groups—not just pick a single group, like with UDDI. Anything to do with XML will be standards-based. We hope to jumpstart things, like with SOAP, etc. Our hope is that when you communicate with XML, no matter the product—whether Word, BizTalk, or our database products, it’ll work with schemas. We think that in the future all the communication will be done through XML schemas.
Mr. Ambur: So if we in the government specify our XML schemas, you’ll be ready to implement them in your products and compete in the marketplace to convince consumers that you can help them do their work better than any other vendor can?
Ms. Adams: Yes, like all the others, but the world is going to product-independence, and we’re going that way. We’re trying to leverage resources from different organizations so as not to reinvent the wheel. Lots of the government has a huge misunderstanding on how XML can help.
Mr. Dodd: It takes making a position, so the Federal Enterprise Architecture people and CIOs don’t get into a battle about .NET, etc. It’s about the interfaces. One of your initiatives has good examples in there. When you see these big and small companies signing up, you can see that we should be moving in that direction. It may not be there yet, but it’ll be there in a year or so. There will be a whole set of companies moving there, so the Federal Government should go there as well.
Ms. Adams: As we move toward this big target, we can hopefully do it together. All the major vendors will say they support these 15 or 20 different versions of the specification, so you can communicate with all the vendor products. Can we do it today? No. There’s a lot of proprietary code.
Mr. Ambur: At our last meeting, someone made the comment that the next version or perhaps even the current version of the [Microsoft] Office Suite can save files in native format. Is that true?
Ms. Adams: It’s not XML-enabled, the way people would like. You can write code to save things the way you’d like, but it’s a band-aid. We haven’t announced a product. For the last four years, the federal practice at Microsoft has been working with the Word people, especially on forms, so if that gives you a clue…I don’t know how it’ll be packaged, but it will give you capabilities to mask from the business user that XML is behind the scene. Today there’s lots of code. Developers write on how to use XSLT, etc. In the future, all that will go away. All vendors’ products will support it. It will be something that looks like Word. You open it, create a document, and exchange it. That’s kind of our vision for the future. When you hear .NET, that’s the vision.
Mr. Niemann: We’ve had a couple of experiences—one with States, and another with a government Agency without an E-Gov effort. We gave a presentation in February to the head of West Virginia environmental protection. He said, “We’re committed to E-Gov, and doing it with EPA. How do we get over the hump?” They need to participate, to get paid in the future for the data they deliver. I said, “Where’s your data?” The answer was that it was in Oracle. I showed how to link with XML Spy version 4.4. Within 30 seconds, we had a schema for all 600 tables in their Oracle database. They then just had to decide which to make public for part of the EPA registry. It seemed to be a breakthrough for them to see that they didn’t have to build a separate system. Once they understand, they can move ahead rapidly.
The other is that you have to do outreach. There’s no substitute for sitting with people and letting them see. Another is with Agriculture. They don’t have E-Gov going on in their Agency. They said, “What can we do with XML?”
I told them about VoiceXML, and they said, “That’s it!” They have critical information from their website that many can’t access. They’ll put it on the Web with VoiceXML, advertise its availability to rural America, and hope it’ll be a success. Right now, they have considerable marketing and outreach expenses.
Trying to XML-ize a whole Agency is different. You have to start with an early success experience, then show that we can link and chain XML services across Agencies. To do that, you work on the vocabularies, then standardize schemas, etc., but you’ll start to build a group of supporters in your Agency. You put it into XML, deliver it, then build on it. You get a few going, then talk about all the work that goes into to harmonize schemas and vocabulary.
Mr. Jacobs: What’s the status of the Web Services sub-group?
Mr. Ambur: That’ll be discussed tomorrow at the AIC. Martin Smith, who is the cofounder of this group, suggested that it would be good to form a Web Services group, perhaps as a task group under the XML Working Group. I don’t care about the reporting lines, but rather the leadership and participation. Now that Brand’s agreed to lead it, I believe we’ll have a good group, and I don’t particularly care about the reporting lines so long as we coordinate our activities closely with each other.
Mr. Morgan: At NIST, we having overlapping areas. One is “naming conventions.” If anyone is interested in participating, please let me know who you are.
On the governance model on the survey, fifth block on the right, it says, “Create a sub-group on draft policy.” What was the policy?
Mr. Ambur: It was not further specified in the brainstorming session. It could parallel the DoD policy, which at this point merely says that XML artifacts used in DoD must be registered.
Mr. Jacobs: At the Navy, we have a draft that’s more extensive than just registry. You could look at that if you’d like.
Mr. Ambur: I have long thought that if and when the registry is proven to be usable, scalable, and so forth that we should recommend that OMB make it mandatory for projects seeking funding in the budget process.
Mr. Morgan: There’s clearly strong support for that. I was curious whether there was a policy.
End meeting.
Attendees:
|
Last Name |
First Name |
Organization |
|
Adams |
Susie |
Microsoft |
|
Ambur |
Owen |
FWS |
|
Atkinson |
DJ |
NTIA |
|
Bargmeyer |
Bruce |
|
|
Bjornsen |
Terry |
Booz Allen |
|
Boyle |
Carrie |
LMI |
|
Callahan |
Jack |
Sphere |
|
Campbell |
Richard |
FDIC |
|
Cocos |
Dena |
LMI |
|
Dodd |
John |
CSC |
|
Embley |
Paul |
PRG |
|
Greeves |
Bob |
DOJ |
|
Hamel |
Debbie |
Discovery Logic |
|
Jacobs |
Michael |
Dept of Navy |
|
Jordan |
Misty |
Booz Allen |
|
Kane |
John |
NARA |
|
Le Maitre |
Marc |
|
|
Lewis |
Diane |
DOJ |
|
McCreary |
Pat |
DOJ |
|
Miller |
Mark |
Booz Allen |
|
Morgan |
Roy |
NIST |
|
Niemann |
Brand |
EPA |
|
Pittman |
Ken |
BAE Systems |
|
Poot |
Lex |
DTS |
|
Rao |
Shanti |
Raosoft |
|
Reagan |
Tana |
Software AG |
|
Rogers |
Rick |
Fenestra Technologies |
|
Sall |
Kenneth |
Ken Sall Consulting |
|
Schaefer |
Barry |
|
|
Turnbull |
Susan |
GSA |
|
Weber |
Lisa |
NARA |
|
Yee |
Theresa |
LMI |