[Help] Last update at http://inet.nttam.com : Tue Jul 4 22:50:47 1995

Extending Common Knowledge

Robert D. Carlitz
Department of Physics & Astronomy
University of Pittsburgh
Pittsburgh, PA 15260


Mario Zinga
Pittsburgh Public Schools
Pittsburgh, PA

(June 28, 1995)


Common Knowledge: Pittsburgh is a school networking project which is developing network connectivity and curricular applications in the Pittsburgh Public Schools. With its emphasis upon the curriculum and its efforts to institutionalize the use of networking technology, the project offers a useful model for other school districts to follow. The present paper describes how the project has expanded from its initial structure, delineates specific products that have been produced and indicates directions in which future expansion is likely to take place.

I. Common Knowledge: Pittsburgh

A. Goals

Common Knowledge: Pittsburgh (CK:P) is a research project based initially in the Pittsburgh Public Schools. The project began as a collaboration of the Pittsburgh Public Schools, the Pittsburgh Supercomputing Center and the University of Pittsburgh with initial funding from the National Science Foundation in January, 1993. The project's original goals [1] were as follows:

The project is now in its third year of funded activity [2], and the original collaboration has expanded to include other groups in the community. The present paper will examine how well the process has worked and make suggestions as to how activities of this sort can best be continued. This paper is meant as a companion piece to Creating Common Knowledge; School Networking in an Urban Environment [3], which we wrote to describe the project's early history.

B. Structure

The original structure of the CK:P project was designed to capitalize upon the strengths of the original three collaborating institutions. The project's educational programs were designed and carried out by staff of the Pittsburgh Public Schools; technical support was provided by the Pittsburgh Supercomputing Center; and project assessment and overall management was provided by the University of Pittsburgh. This arrangement worked well to get the project started and to develop the educational procedures and technical models required to address the goals listed above. These products are discussed in the section which follows.

II. Products

A. Research Environment

Perhaps the most important characteristic of Common Knowledge: Pittsburgh is its emphasis on maintaining a research environment for all participating students, teachers and staff. The Internet is a research tool par excellence, and CK:P has focused upon making this tool available in Pittsburgh classrooms and in providing an environment in which it can be used effectively and extensively.

Project activities are based upon a constructivist approach to education. This outlook applies not just to student work which uses networking tools; it has been extended in this project to a philosophy of teacher training and to a model for staff development for the school district's technical support staff. Examples of this constructivist model are found across the project:

B. Process Model

1. Requests for Proposals

During the course of the project, CK:P has evolved an effective mechanism for selecting new school sites. Annual competitions have been set up within the school district. School teams submit proposals which describe curricular activities for which networking technology could solve long-standing problems. These proposals are reviewed by a committee of students, teachers, support staff, principals, and school board members. The winning proposals [4] are provided with connectivity and hardware and the educational support necessary to help them complete their proposed activities.

The internal Requests for Proposals (RFP) have been successful in a number of areas:

The first Request for Proposals was made district-wide, and nearly 40% of the schools responded. In the second year of this activity, the request was limited to middle schools, and there was a 60% response. Interest in the submissions was high enough that local funds were sought for runners-up so that the project could expand its total number of new sites.

While we feel that the RFP process that has been followed by CK:P has been one of the project's most successful innovations, we cannot claim that this approach has been endorsed unconditionally by the school district as a whole. Each year there have been contentious debates on the topic, and it is far from clear whether this mechanism will ever be adopted as school district policy. There are several reasons for this degree of ambivalence, some of which reflect genuine concerns and others which may be more symptomatic of the difficulties that school districts have in accepting significant procedural change.

We feel that these are issues of the sort which arise inevitably when one is proposing significant change. Recognizing that the RFP process could fail to reach some schools in the district -- either because they didn't understand what could be provided through the competition or because they doubted that the competition could be conducted with complete fairness -- CK:P has developed mechanisms to seed activities at schools which appeared to be left out of the initial rounds of internal proposals. With supplemental funding [5] from the National Telecommunications and Information Administration, CK:P has established connectivity at the Hill House Association, a community center serving Pittsburgh's Hill District neighborhood. The Hill House site functions both as community access networking hub and a hub for seed activities at six neighborhood schools. This mechanism has worked well, with dialup activity having been successfully implemented at all of the seed sites to widespread student and teacher interest [6].

2. Institutionalization

A second area in which conscious effort has gone to help stimulate participation is one which has to do with the training of technical support staff in the school district. CK:P has always sought mechanisms for the institutionalization of networking technology in the school district. This goal has been recognized formally by an administrative shift in which primary project funding flows to the school district, with the University and the Supercomputing Center playing the role of subcontractors for the district's networking activity. A number of mechanisms have been put in place to assist the district to gain the expertise necessary to maintain the technological infrastructure currently under development.

Despite the project's considerable effort toward the process of institutionalization, this facet remains one of the most uncertain elements in the future of CK:P. At issue is a fundamental change in how the school district views the use of technology and a fundamental change in how technological resources are procured and deployed in the district. CK:P is seeking to go beyond simply providing a pilot for the school district and hoping that this pilot might be replicated. Rather the project is actively attempting to build structures in the district necessary to sustain the elements that we have found to be key for success.

3. Curricular Applications

Common Knowledge: Pittsburgh has tried to introduce networking technology as a resource for the curriculum. The idea has been to present the technology as a tool, not to package specific lesson plans with the technology. While this approach requires more time for teachers to master, it has the potential of greater long-term pay-off, since teachers who approach technology in this spirit will be able to revise their curriculum with the aid of this tool on an ongoing basis. There are many examples of how these tools have been applied by students and teachers in the school district:

4. Publications and Presentations

Project materials, starting with the project's initial grant proposal and including current minutes of various working groups, are made available online wherever this is practical. This mechanism provides dissemination of project results to groups outside the school district, and it provides a coherent picture of the project's purpose and progress to everyone inside the school district. The constructivist approach requires that a great deal of information be made available to students, teachers and staff in the school district, and the network is an excellent way to provide this information. The hope is to reduce the compartmentalization of knowledge within the school district and thus to encourage widespread experimentation with the technology and innovation applications across the curriculum.

Among the materials created by the project are many components of its online support mechanism. This includes concept papers, tutorials, and local mailing lists and newsgroups.

5. Student Access

A key feature of the architecture which underlies CK:P is the ability to scale to reach all students and teachers in any networked school and, eventually, to reach the entire population of the Pittsburgh Public Schools. There is a desire that there be no obstacles to expanded student access and use. As resources such as the World Wide Web have become available on the project, they have been configured to allow everyone to use them. Thus there is not only universal access to electronic mail in the school district; there is also a universal capability for the creation of personal Web pages.

One place where the philosophy of universal access that is difficult to support is the area of after-school access to CK:P servers. Ideally, we would like to provide dial-up access for all teachers and students. Presently there is dialup access for PPS teachers through the courtesy of the University of Pittsburgh and Carnegie Mellon University. Efforts are under way to establish a school district modem pool, and students can use the public modem pool of the Carnegie Library of Pittsburgh to reach their CK:P accounts. The near-term limitation to this process will be the availability of inexpensive user devices for access from home. If a mechanism can be found for lowering the cost of such devices to the $100-$200 range, it will become possible to issue computers for home use much as the school district currently issues textbooks or graphing calculators. This may be a difficult target to reach, and there is a need to allow for after-school access from other sites as well. The Carnegie Library is implementing a program of Internet connectivity from the libraries of Allegheny County, and CK:P is working to provide similar access from community centers in neighborhoods around the city.

C. Network Architecture

The network architecture of CK:P is based upon a standard hierarchy of Local Area Networks (LANs) at school sites, a Metropolitan Area Network (MAN) tying together all the school sites, and a Wide Area Network (WAN) providing connectivity to the global Internet.

The LAN architecture of CK:P is close to the industry standard. Manageable 10-base-T hubs provide several network drops in each classroom. Where building size requires multiple hubs, or where there are detached building at a given site, multiple hubs are to be connected with fiber links. The school LAN is populated with the following devices:

The Metropolitan Area Network [12] developed for the school district by CK:P attempts to provide whatever bandwidth is needed by the educational projects under way in the various project schools. Initial bandwidth was provided with V.32-bis modems, but this connectivity was soon outgrown. Present baseline connectivity is specified to be 56 kilobits per second, but many sites already have higher bandwidth links. A major effort has gone into identifying economical means for providing such links. Some of the successful mechanisms that we have employed include:

WAN connectivity for CK:P is provided through the project's central site, currently located at the Pittsburgh Supercomputing Center (PSC). This hub will eventually be shifted to the school district's central administrative facility, which is located on the same street as the PSC, just a few hundred feet away. A 10 megabit LADS line already connects these two facilities, and we expect the ultimate transition to be nearly invisible to school district network users.

D. Clients and Servers

Classroom user devices are either Apple Macintosh computers or Intel-based PCs running Microsoft Windows. CK:P has tried to provide an environment in which these devices can interoperate smoothly, and the choice of classroom hardware has always been a matter of individual local preference. Machine configurations, in terms of disk space, memory and video displays, have evolved with industry norms. At each stage of the project typical commodity-priced computers have been adequate for all network applications. This will continue to be the case as networked applications increasingly become a driving force for the industry as a whole.

School servers are presently Unix-based computers. This choice has been dictated by the ready availability of networked applications for this environment and by the stability that these systems typically offer. As competing operating systems, notably Windows NT, mature, we will consider a possible migration.

Among possible Unix systems, CK:P has evolved a preference for NetBSD. This is a public domain Unix implementation based upon the Berkeley Unix code. Obviously pricing is an important factor here, but equally compelling technical arguments can also be made:

CK:P servers have a number of features which make them attractive in the school environment. Some of these features are fully supported in our present software release, and others will be added as our development effort continues.

The final element of the CK:P system architecture that we should mention is how the client and server components link together. We have developed software to facilitate PC setup and maintenance. New user devices can be configured automatically from a local or remote server. The same software allows one to test for the integrity of software installed on any site PC and automatically replace corrupted components with authoritative copies.

III. Challenges

A. Project Redefinition

After two years of operation in a purely research mode, there was a need for CK:P to redefine its role and structure within the school district. This redefinition was marked by an administrative shift in which the Pittsburgh Public Schools became the prime contractor on the project. Simultaneously there was a need to shift focus from a pure research emphasis to a mix of research and production responsibilities. This shift has posed a challenge to the project's organizational structures and procedures. Increased emphasis upon deadlines, performance and reliability has placed new burdens on the project's support staff. These burdens will be offset over time by school district staff who will take over many of the current responsibilities of the CK:P technical staff. Until that shift takes place, there is not only a need to attend to the demands of running a production environment but a need to be training the school district staff who will assume these roles in the future.

The project's current phase is one in which we are actively seeking adoption of project techniques by the district as a whole. This is unlikely to be a spontaneous process and requires CK:P to package the products it has developed and make them available throughout the school district much as commercial vendors currently do.

Within the CK:P collaboration the new responsibilities and new focus have led to a re-examination of the project's components and their most effective interaction. During the phase in which initial deployment was taking place, it was easy to separate functions such as the instruction of new users and the construction of network servers. This phase was thus one in which the project's educational and technical staff could operate with a considerable degree of independence. The current need to produce more polished products requires a closer level of interaction among project staff and a convergence of interests among the various collaborators. This convergence will yield a project team capable of extending the project's work to school districts, community center and other groups across the Pittsburgh area and beyond.

B. Scaling up to Large Systems

A key issue in any school networking project (or any networking project, for that matter) is one of scalability. At the school level there is the question of how easily one can expand from a situation in which there is only a handful of students and teachers using the network to a situation in which everyone in the school is making use of the facility. At the district level there needs to be a transition from pilot projects to a full district-wide implementation. These questions raise both architectural and administrative issues, and CK:P has tried hard to consider these issues in the design of its networks.

The decision to provide school servers at CK:P sites was based largely on this concern. Since the initial bandwidth for our Metropolitan Area Network links was low, a school-based server was necessary to accommodate many simultaneous users and to allow the caching of items which place particularly high bandwidth demands on the MAN links. The capacity of each school server was made large enough to accommodate all students and teachers at a given site, so there would be no reason to discourage local expansion of any CK:P project activity. The school LANs were designed with similar expandability in mind.

Scaling involves not just hardware issues; it is also a matter that raises significant procedural questions. By relegating user administration to local site coordinators, CK:P avoided a bottleneck at schools seeking a rapid expansion of their local user base. And by developing mechanisms for the automatic update of server and client software, CK:P has made it easy for central support staff to maintain an expanding Metropolitan Area Network.

C. Convergence with Practice in the School District

We believe that the technology of wide area networks offers a vehicle for change and reform within the school system. This is clear in classroom applications which open the path to new resources and new styles of work. Less clear, perhaps, is the role of central administration in facilitating this activity. Districts where this type of activity has moved quickly are typically smaller districts in which there has been a Superintendent who has personally shepherded along the process of networking the district. In larger districts this process is slower and more difficult. We can identify a few factors which are likely to be relevant:

D. Involving Reluctant Participants

This problem has been less urgent than we first anticipated -- largely because the number of reluctant participants has been much smaller than we first anticipated. District usage of the Internet is approaching the level of 40% of all teachers, and this level of participation has been achieved with LAN connectivity at only 15% of the district's schools. There are several reasons for this high level of participation:

IV. Extensions

A. Building Collaborations

Common Knowledge: Pittsburgh started as a collaboration among three educational organizations, and it is growing to be a collaboration with many more participants. A collaboration with the Hill House Association has facilitated networking activities at six Hill District schools. This activity has evolved to include a significant community networking effort, as will be described below.

Another collaboration that has been under way for the last year is one involving the Pennsylvania Room of the Carnegie Library of Pittsburgh. The Pennsylvania Room houses a collection of 30,000 photographs of historical interest. A portion of this collection [15] is being placed online for use by the community and the schools. By developing an in-depth resource of this sort we hope to be able to gauge the impact of significant new materials on instructional activities in the schools. At the same time we are learning the level of effort which goes into the creation of such a resource and the value of placing materials which were previously available to the public only in a very restricted environment out on the network for much more general access.

This process has yielded some surprises. It appears that the online exhibit of material from the Pennsylvania Room is the best way that these materials have even been shown to the public. We had expected that the online collection might be something of a teaser -- low-resolution photographs which would lure visitors into the Library to see the real thing. But the quality of the online photographs is high enough that viewers can appreciate them directly, and the navigational capabilities of the online exhibit are better than what is available with file cabinets filled with archived photographs. The result is that viewers return again and again to the online exhibit, and the curators of the Pennsylvania Room are considering a wholesale conversion to this medium.

We expect the experience of the curators of the Pennsylvania Room to influence network use throughout the Carnegie Library and its neighbor organizations - the art, science and natural history museums of The Carnegie. All of these organizations are important providers of educational materials for the school district. An increased online presence by these organizations will increase the value of network connections both for them and the schools and students that they serve.

The online exhibit is only one of several intersections of Common Knowledge: Pittsburgh with programs of the Carnegie Library of Pittsburgh. The Library is constructing Internet access points at all affiliated libraries across Allegheny County. Students will be able to connect to their school accounts through these library sites, and further access will be facilitated through a public modem pool maintained by the Library. Additionally there is a public access network, the Three Rivers Freenet [16], which will provide access for parents of Pittsburgh school-children and for others in the community who may wish to help in the development of educational resources.

B. Community Networking

The activities of the Carnegie Library of Pittsburgh are but one set of examples of how the networking of community resource providers can enrich the community as a whole. Other examples are provided by CK:P's collaboration with the Hill House Association. This collaboration is not only stimulating activity at six seed schools; it also provides a network access point [17] for members of the Hill District neighborhood and a focus for job training and commercial development in the neighborhood. We believe that successful community networking requires an economic basis adequate to sustain and expand the community's networking activities. Our collaboration with the Hill House Association is an effort to provide examples of how this can be done.

In addition to the direct economic benefits that networking might offer a given community, there are several arguments which suggest that community networking should be an integral part of all school networking efforts:

C. Parallel Activities

As networking activity in the school district expands to involve all schools and all classrooms, it is inevitable that we will encounter recalcitrant and uninterested teachers and principals. Realistically, one must admit than some of these individuals will never take part in the activity except as might be mandated by the district (for example, to receive e-mail announcements relating to district policy or to fill in online forms required for administrative matters). Here an important feature of network architecture is likely to lessen the impact of this situation. This feature is parallelism -- the ability for individuals on the network to function independent of any controlling entity. Students can locate resources on the network whether or not they have a teacher who is actively encouraging them to do so at every given minute. All that's needed is a functioning infrastructure and widespread knowledge of how to use this infrastructure.

It is likely that over time the number of teachers who refrain from any network involvement will be small. This expectation has to do with the fact that the network is a two-way communications medium. As schools develop more and more online resources, an increasing fraction of all parents will begin to use the network to see what's going on in their children's schools. Through this mechanism teachers who maintain no presence on the network will appear to be inactive in their schools. Involved parents are likely to question this and will provide a further nudge for these teachers to increase their level of involvement.

D. External Resources

An important lesson that can be drawn from the various collaborations with which CK:P has been involved is that there are valuable resources and significant expertise available in organizations outside the school district but available to the district via the network. This single point could probably be used to provide an economic justification for extensive school networking programs. While participants of CK:P have all profited from the resources present in the various participating organizations, it is clear that the sharing of these resources cannot be expected to occur on a totally spontaneous basis. Different organizations have different associated cultures, and the mere fact of connectivity between two organizations in no way guarantees a significant sharing of resources between these organizations. What we have found is that when one makes a conscious effort to share resources, the network does offer a superb mechanism for carrying this out.

E. Creativity

The issue of resource-sharing and collaboration invites a question as to what motivates educational collaborations and what gives them the vitality to endure. An obvious answer lies in the word education and in a shared desire that many groups might have to provide an environment for continuing education. But this word is broad in meaning, and it's unlikely that organizations can be drawn together by so broad and vague a premise. We prefer to emphasize the word creativity and to invite organizations to provide educational environments in which creativity can flourish. We regard wide area networks as a significant stimulus to creativity and a marvelous mechanism for the sharing of creative ideas. In this spirit we invite expanded collaborations of the type which have marked the growth of CK:P, either through direct affiliations with Common Knowledge: Pittsburgh or through similar efforts motivated by a kindred spirit.

A logical focus for these collaborations is provided by the schools. This requires a commitment in terms of resources and personnel, but it offers significant new resources for the schools through the groups it brings to the schools and through the services that these groups offer. Networks build community, and school-based networks can help bind the community together and strengthen support for education. We believe that the future of public education is closely bound up in its ability to serve the community. In this sense successful school networks may be a key to the livelihood of public education itself.

Footnotes and References

[1] These goals are listed in the original grant proposal for CK:P, which is available online:
[2] Funding for CK:P has been provided by major grants from the National Science Foundation and the National Telecommunications and Information Administration of the United States Department of Commerce. Local support has been provided by the Buhl Foundation and the Vira Heinz Endowment. Vendor contributions have been provided by Apple Computer, Digital Equipment, Sun Microsystems and Tele-Communications, Incorporated.
[3] This paper is available online:
[4] A description of the RFP process is available online:
[5] The NTIA grant proposal is available online:
[6] Vann Elementary has already constructed a World Wide Web presence:
[7] An outline of the system administration course is available online:
[8] Examples of the German language lessons are available online:
[9] The beginnings of an online presentation for Kennywood are given in:
[10] Examples of student research are available online:
[11] Students' poetry from Liberty School is available online:
[12] The paper Stages of Internet Connectivity for School Networking is available online:
[13] The Cablelabs RFP is available online: http://www.cablelabs.com.
[14] The URL for the CK:P Web server is http://info.ckp.edu.
[15] A trial version of the online exhibit is available at
[16] Online access to the Three Rivers Freenet is available at the URL
[17] The Hill House Community Access Network can be seen at: