|
|
|
Defense Technical Information Center
Ft. Belvoir, VA
June 1, 1999
WELCOME
Kurt Molholm, CENDI Chair, began the meeting at 9:15 a.m. He welcomed the members to the Headquarters Complex and to DTIC. Introductions were made.
Mr. Molholm set the tone for the meeting by reading from the Millenium Calendar for May 21. William Knoch notes that there will continue to be tension between control of information and letting it run free in order to encourage innovation. The ViewPoints feature of a recent issue of the Communications of the ACM highlighted the importance of "information distillation" when moving to the information economy. Content and not form is primary. Technology helps to deliver content better.
SCIENTIFIC AND TECHNICAL INFORMATION EXCHANGE
HPCC, NCO, CIC, NGI, PITAC, and IT2: Connecting the Dots in Computing, Information
and Communication
Kay Howell, Director, National Coordinating Office, CICC
The federal government has a large investment in information technology (IT) and a commitment to ensuring U.S. competitiveness in this area in the global economy. There are a number of federal organizations and initiatives with responsibilities in IT -- the High Performance Computing and Communications (HPCC) initiative, the National Coordination Office (NCO) for Computing, Information and Communications, the Next Generation Internet (NGI) initiative, the President’s Information Technology Advisory Committee (PITAC), and the Information Technology for the 21st Century (IT2) initiative. Ms. Howell described each of these initiatives/organizations, their interrelationships, and the status of their efforts.
The High Performance Computing Act of 1991 (PL 102-1) directed the President to implement the National High-Performance Computing Program. PL 102-1 sets forth the program requirements, which at that time were focused at high-end computing needed to support the National Research Education Network (NREN). The requirements set goals and priorities for federal high-performance computing research, development and networking; provided for interagency coordination; and established oversight for the operation and evolution of the NREN. The law also required improvements to software, acceleration of high-performance computer system development, and technical support and R&D for software and hardware needed to address fundamental problems in science and engineering. As testbeds for these fundamental problems, ten specific problems, or Grand Challenges, were selected. The law also called for an advisory committee to be established. This was finally done in 1997.
The budget for the HPCC Initiative averaged ~$1B per fiscal year, of which ~$.5B was new money. The remainder was reallocated from initiatives already underway in the participating agencies. In the ensuing years, the annual budget for the HPCC initiative reached a high of $1.2B.
The programs included in the HPCC were a loose collection of activities already underway in various agencies. If the programs were under HPCC oversight, they were included in the joint discussions of the group, and the budgets were included in the HPCC budget.
However, it was always acknowledged that significant computer, information management, and telecommunications work was being done by the federal government that was not included in the HPCC initiative, because every agency continued to have mission-specific computing activities that did not fit the objectives of a cross-cut. In order to be considered for inclusion in the HPCC initiative, the program must have wide application or the agency must be willing to modify the approach to be more broadly applicable. Projects must also be focused on research rather than Internet infrastructure and development.
In 1996, the HPCC was reauthorized without new authorizing language proposed. The program continued to exist with a constant budget. The Brooke-Sutherland Report reviewed the program. It recommended that the HPCC be broadened to include the National Information Infrastructure (NII) issues including education, curricula, user interfaces, and "high competency systems". Based on this recommendation, the CIC (Computer Information and Communications) R&D Program evolved from the Federal HPCC Initiative, while under the chairmanship of Anita Jones. (Many continue to refer to these programs as the HPCC.)
The CIC R&D Program provides a mechanism for focused long-term interagency R&D in computer, information management, and communications technologies. It coordinates the R&D programs conducted in these areas by 12 agencies and departments. The main goal of the CIC is to look for complementary and duplicative efforts where co-funding is beneficial. Joint proposals are often presented. The group provides a review of the quality and types of projects to be supported. The activities of the CIC include the Next Generation Internet initiative.
The Subcommittee on CIC R&D provides coordination, planning, budgeting and review for the multiagency CIC R&D programs. The CIC reports to the Committee on Technology of the National Science and Technology Council (NSTC). (The NSTC reports to the Office of Science and Technology Policy [OSTP] within the Executive Office of the President.) The National Coordinating Office (NCO) (www.ccic.gov) is a staff office to the Subcommittee, with the NCO Director acting as the chair of the Subcommittee. Planning activities are coordinated with OSTP and budget activities with the Office of Management and Budget (OMB). The Subcommittee reviews the programs and budgets, and tracks R&D by disciplines of interest. The NCO prepares a formal crosscut report to OMB, and is responsible for the implementation plan. The NCO is responsible for justifying the agenda and budgets to Congress. This involves a great deal of outreach. This group also provides technical assistance to and coordinates implementation of recommendations made by the Presidential Advisory Committee on High Performance Computing and Communications, Information Technology and the Next Generation Internet. The members of the subcommittee include representatives from the 12 participating agencies/departments, OSTP and OMB. An agency or department cannot be a formal member if their money is not included in the CIC budget, but all agencies are invited to meetings for information sharing.
The Subcommittee also oversees the five Program Component Area Working Groups and the Federal Information Services and Applications Council. The five Program Component Areas (PCA) are High End Computing and Computation; Large Scale Networking; High Confidence Systems; Human Centered Systems; and Education, Training and Human Resources.
The real collaboration happens at the program manager level where proposals are reviewed and coordination of similar proposals occurs. To quality, the work must span multiple agencies. A project may not be easily pigeonholed into a component area, but each is assigned to only one area in order to count it consistently in reports and budgets.
The Next Generation Internet (NGI) (www.ngi.gov) is an initiative under the Large Scale Networking PCA. It is a multiagency R&D program to develop new and more capable networking technologies to support federal agency missions, create a foundation for more powerful networking, form partnerships with academia and industry, and enable the introduction of enhanced networking services in the private sector. A subset of the CIC agencies are involved with the NGI initiative, including the Defense Advanced Research Projects Agency (DARPA), National Science Foundation (NSF), Department of Energy (DOE), National Aeronautics and Space Administration (NASA), National Institutes of Health/National Library of Medicine (NIH/NLM), and the National Oceanic and Atmospheric Administration (NOAA). The proposed funding for FY2000 is $103M.
The NGI includes an emphasis on metrics and, therefore, the improvement goals have been stated in terms of end-to-end performance improvements of 100 - 1,000 times the 1997 Internet rates. Quality of service, including security, are also important. The first goal is to build on current federal high-speed networks, including NSF’s very high performance Backbone Network Service (vBNS), NASA’s Research and Education Network, DoD’s Defense Research and Engineering Network, and DOE’s Energy Sciences Network. As of March 1999, they have connected about 150 sites in all 50 states.
Work is also proceeding on the second goal, a 1,000x testbed (the SuperNet). Approximately 20 sites have been connected as of March 1999. These connections are 1,000 times faster than the 1997 Internet. Many of these connections are to regional high-speed networks that include national laboratories and federal research organizations.
The third goal of the NGI is to conduct R&D applications that require high performance networks. These include collaboration technologies, digital libraries, distributed computing, remote operation and simulation, etc. Disciplinary applications are also considered important. These may include specific applications for crisis management, education, the environment and health care.
The President’s FY 2000 budget request was reviewed. There are obvious changes in the assignment of monies. DARPA’s budget for R&D has been declining, as it shifts from long-term R&D toward shorter term prototyping. The DOE ASCI has been added to the CIC budget for FY 2000. It involves a large simulation effort to certify the nuclear stockpiles. This is a major computing-intensive activity so, although the CIC budget appeared to grow, it was really because the ASCI money was included this year. There hasn’t been a decision yet on what this cooperation with ASCI will entail. There is concern about the time frame because the simulation may be moved forward too quickly for information sharing and coordination to occur.
The President’s Information Technology Advisory Committee (PITAC) is composed of 26 non-federal members established to provide advice and independent assessment of IT issues to NSTC and OSTP. The NCO is responsible for setting the PITAC agenda, scheduling speakers, helping with the reports, etc.
The final report from the PITAC was released in February 1999. The report found that federal IT R&D is inadequate given the growing importance of IT to the nation. The PITAC report recommended a strategic initiative in long-term R&D, requiring an investment in long-term, high-end R&D that the private sector won’t do. The investment should be increased by $1.4B per year by FY 2004. Four priority research areas -- software, scaleable information infrastructure, high-end computing, and impacts of IT on society and the economy -- were outlined. PITAC originally discussed a federal department of IT R&D. However, a distributed model was thought to ensure more diversity. The need to re-establish NSF’s role as a leader in computing R&D was emphasized.
The final IT initiative discussed was the Information Technology for the 21st Century proposed in the President’s FY2000 Budget. IT2 responds to recommendations from the PITAC by increasing investments in fundamental IT research (computers that are 100-1,000 times more powerful, capable of five to 40 trillion computations per second by 2003), advanced computing for science and engineering, and research into social issues and support for training and retraining an IT workforce. The initiative calls for the establishment and funding of multidisciplinary teams working on the most challenging problems including climate change prediction, weather prediction, genetic functions, modeling the evolution of the universe, and simulating combustion. The four priority research areas are the same as those identified by the PITAC. The Administration’s proposed FY2000 budget for this initiative is $366M.
Rep. Sensenbrenner, Chair of the House Science Committee, has announced his own recommendations for IT R&D expenditures. In some cases, his proposal has advantages over the Administration’s proposal. President Clinton is not willing to commit to a budget beyond the year 2000, because he wants to solve social security funding first, but Sensenbrenner has called for a five-year program which tracks very closely to the PITAC report. He also leaves more flexibility in applications areas for the money. It does zero out DOE’s money and moves it to NSF. Almost all of this increase is new money.
The IT2 initiative will bring increased coordination among the initiatives and organizations in federal IT. In FY2001, the HPCC/CIC Programs and the IT2 will be integrated and reported as a single program.
The attendees requested a copy of Ms. Howell’s slides.
Discussion
Members of the group asked about the relationship of the NGI and the Internet2 initiatives. Ms. Howell explained that the NGI is a multi-agency federal initiative, while the Internet2 is a group of U.S. research universities. There is a cost associated with joining the Internet2 group. There are corporate partners who contribute resources. Quest, Nortel and Cisco have donated high-speed connections at no cost for a year.
The Abilene Project is an initiative of a group of the Internet2 organizations that have formed the University Corporation for Advanced Internet Development. The goal of the project is to support development of an advanced backbone network that connects regional network aggregation point of presence, or gigaPoPs. These aggregate points would then connect to the high performance networks such as those in the NGI. These aggregates provide scaleable high-speed connections. Abilene is peering with other high-speed research and education networks including the vBNS. There is overlap between the members of the Internet2 and those receiving grants from NSF through the NGI funding. Some organizations are using their NSF grant money to pay for connections to the Internet2 networks.
New Technologies, Architectures and Applications: DTIC Project Developments
(Kurt Molholm and Staff)
"Proactive Customer Advocate Program (PROCAP)"
B Rex Ridenhower
A key question in the current environment is "Who is the customer?" The need to develop more knowledge about customers and their needs has led DTIC to develop PROCAP, a proactive "outbound" call center. Based on the concept of a "listening post", representatives call long-term, new and potential customers for DTIC’s products and services via telephone to determine customer satisfaction and to enhance long-term relationships with customers. PROCAP allows DTIC to conduct market research, gaining what they believe to be more quality data from the phone than from limited focus groups or mail surveys.
HQ establishes the call scripts and tests them. The scripts are then used by the account executives at the DTIC Regional Offices. The information gathered from the interviews is entered into the Service Quality Information System (SQIS). The files are transferred to DTIC where they are processed and a variety of qualitative and quantitative reports are provided for various levels of management.
New users are contacted within 60 days of becoming customers. During this conversation, the customer representatives ask questions to ensure that customers are making effective use of the product. This system is also helpful in collecting information during new product testing.
A customer representative talks to each DTIC customer at least twice a year. This not only establishes a relationship with the customer, but it provides a means of describing new products and services, and ensuring that e-mail and mailing addresses are up to date.
PROCAP is used in addition to more traditional research approaches such as mail surveys and focus groups. DTIC believes that multiple approaches provide more valuable information than a single approach.
"Defense Learning Demonstration"
Pete Suthard
As part of its Strategic Plan for Distance Learning, DTIC is developing a Web-based user training center. It is multimedia-oriented and allows enormous flexibility and customization of content. As a demonstration, DTIC has developed a Web-based course on scientific and technical information (STINFO). This is used in conjunction with a three-day, full-time class. The STINFO course emphasizes scientific and technical publishing, particularly acquisition, communication, dissemination, and the importance of scientific and technical information to the research process. It features audio and video clips from various members of the user community. There are also links to the regulations that define and authorize the STINFO program.
The STINFO course is modular and begins by allowing the user to select his/her "role" from a group of people on screen. For example, the user can see the program from the view of a R&D manager, scientist, or information manager. This, then, tracks throughout the interaction. An estimate of the duration of each module is provided to give the learner an idea of whether or not he can complete the module within the time he has available. Some individual modules are as short as 5-10 minutes in length. The system tracks those modules the learner has already accessed. Online evaluation forms provide feedback to DTIC. The system will even print out a certificate of course completion. Users can sign up for a forum discussion list related to the topic.
Discussion
Mr. Suthard was asked about the authoring tool and the cost and time to develop this web-based training package. The authoring tool is ColdFusion with additional software used to prepare various pieces of the multimedia input. The product, developed with the aid of a consultant, took over a year from inception. The Marketing Department created the actual content that the contractor then used to create the final product. This took approximately eight months. The cost was approximately $30,000 - 40,000 in addition to DTIC staff labor hours.
"Defense Virtual Library "
Marcia Hanna
In conjunction with the Center for National Research Initiatives (CNRI) and DARPA, DTIC has developed the Defense Virtual Library. The DVL uses the Handle system, developed by CNRI for digital object identification. DARPA provided some start-up funding and advice. This project gives DARPA a good interoperability testbed. Technical advice was also provided by the Library of Congress’s National Digital Library Project.
The DVL includes a variety of document and format types. The Army Band contributed a series of audio clips that have been cataloged and indexed. The DVL also includes photographs, oral histories, and a limited number of technical reports. There are fewer documents from DTIC than originally anticipated because they needed room on the server for the video files. The project has studied the collection development issues related to image collections. Other issues included the digitizing of photographs, storage, metadata, and how the results should be presented. When working with sound, they found that the oral histories were exceedingly long. Therefore, DTIC took the approach of LC and edited and excerpted from the oral histories.
The DVL is intended to be very easy to use with simple search tips. For example, relevance ranking is available but is not described in detail because it is considered difficult for users to understand. Despite the fact that MARC tags are used, the user-oriented fields have been moved to the top of the display, with the more detailed, library-oriented fields at the bottom.
Discussion
Mr. Molholm indicated that the Handle system is compatible with the Digital Object Identifier. The DOI is the theme of the latest ICSTI Forum, which Mr. Molholm recommended as a good overview and status report on the topic.
"The Channel DoD"
Carlynn Thompson
The Channel DoD system is a project to provide customized portals for DoD users. A portal has several characteristics. It has content channels that are directories or guides to information that are grouped in a logical way based on the needs of the user/audience. A portal also has the ability to perform special searches, such as searching in a certain language or for a particular document or format type, based on the content and user characteristics. A portal also has customization features that let the user personalize the way content is presented.
The Channel DoD design is based on customized browser designs such as "My Netscape". It provides a personalized Internet starting page that acts as a doorway to the most relevant information, such as newsfeeds, commerce, databases, etc. The Channel DoD will initially focus on scientists and engineers, but it is also appropriate for other customer groups. Unlike the commercial portal sites, it will deliver DoD-specific newsfeeds; search tools such as STINET, Technology Navigator, and Directives and Instructions; and tools such as calendars, distance education, a per diem calculator, and listservs.
One of the issue areas when developing the system was that of access to commercially licensed databases. DTIC will be conducting a survey later this year to determine if users are interested in an enterprise-wide license for such databases. Defining the access group and determining who pays and how much are key issues negotiated with producers.
Channel DoD has been an 18-month effort. The technical framework was developed during the first 12 months. The prototype has been developed over the last six months.
Discussion
Mr. Pedtke indicated that the portal concept is also being utilized in the intelligence community. However, he is interested in how to encourage people to update their portals once they have been established. No one knew of any work in this area, but all agreed this would be critical to the ongoing usefulness of a portal service.
"AF Online Newspaper"
Carlynn Thompson
The USAF Online News is an online newspaper aimed at improving communication between senior USAF leaders and USAF people. The focus is on people, missions, and resource news. The primary goal is to inform USAF people, improve morale, and provide news that will improve their readiness and positively affect their decisions. It is also aimed at promoting public trust by openly sharing USAF news.
The online paper includes editorials, commentaries, notes from the field, regular features, pictures, and an events calendar. The paper is published every Wednesday, but supplemented by breaking news on an "as needed" basis.
There are correspondents world-wide who contribute to the newspaper electronically. The USAF community may also submit contributions, and the general public may submit letters to the editor. This requires a timely publication system, but one that has well-established work-flows and enhanced security features.
Online News is the first major system that DTIC has developed that employs a three-tier architecture -- Oracle and WebObjects with Java-based network applications. The infrastructure separates the data sources from the application server and web server from the browser. The application and web server sit as middleware, limiting access to the back-end Oracle database. Permissions are always locked down through this "gatekeeper".
Templates have been created that allow authoring and editing to be done more efficiently. A form has been developed for submission of letters from the public. Content is stored in a back-end Oracle database, where it awaits processing and release by the editors. The editors have access to the editorial system where they can view and modify the information that has been submitted, as well as access to back-issues so they can see how and when a topic may have been covered previously. The editors work in a template with the various news sections already arranged. They review and edit the text and integrate photos and graphics as needed. Metadata tags are created to accompany the full text. When the editor marks the item for publication, it waits in a queue for the next issue. The system is also capable of providing hot-news flashes to its audience, bypassing the actual weekly issue publication.
Issue processing involves the creation of static HTML web pages. Static web pages are more reliable than creating the pages "on the fly." They are also efficient in providing the content to the mirror server in California. A combination of the Oracle fixed-field and Verity full text search engines are used to create the content for the HTML pages.
The security afforded by this tiered architecture will allow the system to become more complex in the future. There is already a desire to post separate versions of public and closed events.
Other DTIC customers have requirements for similar systems. The use of WebObjects will shorten the development cycle for these other systems because reusable components, pre-built components, and graphical development tools can be utilized. More information on WebObjects is available from http://www.apple.com/webobjects/overview.html.