"Sloppy" Networks

by Jim Britell

Summary:   A communication network appropriate for grassroots activists may need to violate many conventions and be sloppy, god-awful redundant, overlapping and massively parallel.

A computer network's design should complement the organization it serves. As we go about developing grassroots communications, we should avoid, consciously or unconsciously, trying to create an amateur version of a corporate on-line network. Most existing networks are designed to serve corporate, hierarchical, top down organizations; but even corporations with far greater resources than grassroots environmentalists, find it difficult to establish and maintain communication networks. The environmental movement, since it is non-hierarchical, and lacks a central headquarters or a single goal, seems to violate almost every known design rule for organizations. A communication network appropriate for grassroots activists may need to violate many conventions of established telecommunication network design.

The grassroots, disbursed, autonomous and likely to be poorly trained, is not really subject to direction. It is quixotic to assume that it could ever pull together to implement one large project as a corporation might. The only way to implement a national network is to take the disorganized and chaotic nature of the grassroots as a given, even a strength, and use it as an opportunity. A communication network appropriate for the environmental movement will be sloppy, god-awful redundant, overlapping and massively parallel. It will need to perform at 100% efficiency even when 80% of it is inoperative, and it must be dirt cheap, easy to use, state of the art and technically supportable with minimum staff.

Complicating our problem is the fact that most networks do not have sworn enemies determined to destroy them. But networks and electronic forums created by controversial movements such as abortion rights, feminist and anti NRA have experienced persistent and sophisticated problems along this line. Our network can expect the same treatment. Given the openness of the environmental movement we can expect that our network, and even material such as this paper that discusses creating it, will receive continuous scrutiny by the far right. Our network must be built to withstand natural and artificial catastrophic events such as jamming, viruses and message flooding.

This paper proposes a new type of design, but to understand how that design would work one needs some background on:

  • How existing networks function and how they are different from one we might construct.
  • Problems people have using their computers and the implications of this for our network.

Differences Between Commercial Networks and Our Network
Corporate networks, with thousands of users, are usually designed to communicate among and between a central office and many users. These networks must operate at near perfect reliability and so deploy proven, tested, fail-safe technology. For example, if a large company is communicating a price change for a product to thousands of dealers, that message may need to be received and acted upon simultaneously in a thousand locations. Commercial network design seeks perfect 24 hour reliability across the entire installation, made secure from sabotage, monkey wrenching and unauthorized intrusion with elaborate and expensive "fire walls" and other security features.

The existing grassroots organization is very different from any commercial organization: it has many heads in DC and the states communicating the same or similar messages by fax, email and written alerts to grassroots groups which may have more than one "receiver". When something happens that all the grassroots need to know, several DC and one or more regional groups often communicate that information simultaneously to thousands of grassroots groups. This material finds its way out to local groups through many channels: fax, phone, email, web sites, and paper mailings. It is now, and likely to remain forever, completely open and totally unsecure.

The enormous communications redundancy within the grassroots can be viewed as an opportunity for the planners of a network, rather than as a problem. A network that is designed to communicate its essential information through many modes may appear raggedy, disorganized, and redundant, but if it performs its primary job to keep local groups up to date on what they need to know and is cheap and bulletproof it will nevertheless be an elegant design. Additionally, a grassroots network that is essentially a pandemonium or of a "sloppy design" will be hard attack.

If we set out to establish perfectly reliable 24 hour communication links with a grassroots group, like a private company might, we could easily spend $10,000 in start up costs for hardware, software, site preparation and training, and then another $2500 per year in maintenance and training for each group. But if we establish two or three alternative channels of communication with each group, perhaps a $250 fax machine, and one or two email connections for $1000 each, the overall reliability of these two or three seemingly redundant, quick and dirty connections (from an overall network design perspective) may be as great or greater than one perfect connection, and allow us many options for hardware, software and training not available to conventionally organized networks where individual "nodes" must work at 99+% efficiency all the time.

The many users of our grassroots network can create varied and unique computer applications depending on local needs and interests. The only thing every organization must be able to do is to reliably obtain the voting records of their local elected officials and hot button issues from groups monitoring the legislatures so that they can apply constant pressure on their local officials. If that information flows to a grassroots group through many different and redundant channels the network that supports them can be built very cheaply and each individual organization will be able to develop local capability to fill the needs of its own particular niche.

A "sloppy network" for grassroots activists is similar in some ways to the kinds of relationships that are growing up between the small high tech companies that are prospering in the Internet environment. There, the most successful companies are small, independent, and enormously creative, and pursue their products absolutely independently of each other within their own unique niches. They have no formal relationships, but they do cooperate fully on common technological standards for all their products.

Some Thoughts On How Corporations Use Technology
There is a certain lead time for absorbing technology. At any point in technological development, mainstream organizations will be found busily implementing technology that is two to five years old. Large organizations must go through a procurement and testing process and bring their technical people up to speed. Only when they can be sure that it is safe will they commit to a given technology. This reality is expressed in such unwritten rules as, "No systems manager ever got fired for buying IBM equipment" and "Never be the first to implement a new technology." Managers in large corporations won't jeopardize their jobs if they stick to established, tested technology.

Large organizations are able to use technology effectively because they use tested systems, prepare their own training materials, invest heavily in hands-on training and maintain on-site systems staffs who prepare custom applications and teach employees how to use them. Half of all major companies develop their own software. A company may budget eight hours of training for each employee just to teach them a new version of their word processing program, and they employ people full time troubleshooting their networks. The busy executive with a mobile phone and on-the-road communication capability is never more than a phone call away from his technical support when problems arise. Without this kind of intensive backup he would not long survive.

Companies Give Their Employees Technical Support
Average people lack the technical support companies give their employees. Not all citizens are coping well with the challenges and difficulties of getting their computers to navigate the information highway. The help lines of computer manufacturers are swamped with calls from people who cannot get their computers to do what they are supposed to do. Mastering communication technology is a rough road of systems crashes and hangups, conflicts and frustration. Even the experts are frustrated by the difficult technical problems that must be solved to function on the information highway.

As an example: I am involved in an email conference which consists of "experts" on grassroots networking technology. Ordinarily when someone receives email they click on the "reply box in the incoming email message, type in their response, hit the "send" button, and the reply goes directly to the person who sent the original message. The other day someone making a routine reply included an offhand remark quite critical of someone on this same conference. What the sender overlooked was that the incoming message was sent by the conference's electronic manager, so the response went to the electronic manager and was routinely and automatically routed to all the list members including the criticized one. The incident was embarrassing for the sender. It points out the problems in mastering this technology and the larger security problems inherent in fast electronic communications.

Grassroots Network Design
Paradoxically, problems such as the one described above may actually be an opportunity for a grassroots network. Although we may not be able to replicate the existing technological and rhetorical infrastructure of the right (and even if we did they would just advance another level while we caught up), we may be able to leapfrog them and build a better one.

For most technology, whether VCR's, computers or modems, the training and instructional materials that accompany the equipment are usually produced as an afterthought. Manuals are written from the point of view of the individual device, not from the standpoint of the user, and not with much consideration of the overall configuration in which the device will be placed. Since most of the problems that arise in on-line communications are created from the interactions of the various devices in the configuration, by standardizing on a model system we can write instructional material from the users point of view, not just give our users a bookshelf of incomprehensible manuals.

The technical resources of the grassroots environmental community are sufficient to create an entirely new type of user friendly computer training materials written specifically for our model system. While we may never be able to impose discipline upon the entire grassroots organization (nor would we ever want to) my earlier experience with other on-line networks has convinced me that it may be possible for a small group of highly disciplined trainers and instructional writers to create training materials of such a breakthrough quality that people will be able use their computers easier than is now generally the case. If our system was easy to use, its greatest strength would be right where most systems are the most vulnerable.

Three design principles and three strategies that may be appropriate for a network designed for grassroots activists:

  • A revolutionary "Sloppy Network" design should be the conscious and deliberate foundation for a grassroots network.
  • Technical support and training must be deliverable in disbursed settings under adverse conditions to users of widely varied sophistication.
  • The computer/modem/software configuration chosen should be simple to operate, cheap, and packaged with instructional and training material that allows easy system's operation by the average non-technical person.

Three strategies are necessary to implement such a network:

  • "Trailing edge technology" (i.e. computer platforms two or three generations behind the current line) because such platforms are quite able to perform the required tasks and are very cheap ($600-$1000).
  • "Cutting edge technology" (i.e. todays technology) should be the standard for the operating systems (software, modems, Internet, email and fax applications.) Fortunately these are the low cost, even the free parts of telecommunications systems.
  • "Bleeding edge technology" (i.e. prototype technology still being tested ) should be used to prepare and distribute training materials.

Constructing an Advanced "Sloppy" Network

  1. As much as possible each site on the network should be functionally equivalent so it can be supported without contending with a crazy quilt of non-standard hardware and software. The network must promulgate standards so sites will be nearly identical from a maintenance and support perspective, although the individual computers found in particular sites may vary in cost and performance. This means that the non-computer parts of systems that create most of the problems, like web browsers, modems and communication software, should be standardized.

    Fortunately, computers with widely different functionality are more or less technically identical from a maintenance and support perspective. What drives up the cost of computer configurations is the ability to process complex graphics and scientific applications, not the uses we care about. The applications we need (email, Internet etc.) can be performed just as well on a $1000 configuration as on one that costs ten times that much. (It may seem unreasonable to think we could all standardize on one model system in every corner of the country, but I think we can. This subject will be covered in depth in a future paper.)

  2. Use creative people within the movement to develop an instructional infrastructure for the network that is more advanced than that found in commercial installations so that activists can maintain and troubleshoot their computers much easier than is now possible. Instructions for users must be focused, not on how the computer works, but on exactly what the user needs to do. We must take advantage of new technologies which allow computers to produce and distribute production quality instructional material directly to users. These include the production of audio/visual materials such as video tapes to orient new users not yet computer capable, and the advanced web technology of the Netscape browser which allows a technical person to create real time radio programs, newspapers, and even television.

  3. The model system must contain processors that are proven and reliable, perhaps two generations behind current models. These should allow the system's cost to remain low, but still provide all the email, faxing, and web browsing capability required. Modems and software, the least expensive components of the system, must be the latest versions and models.

  4. Model systems must be standardized so a few people can monitor the on-line forums of the manufacturers of the equipment and prepare technical bulletins which will allow the users to maintain their systems without requiring each of them to keep track of the continuous small upgrades which arise because of the continuously evolving nature of on-line technology.

  5. New users who have not yet bought their first modem need an automatic "fax back" system so they can call in and from a voice menu order material which will explain in simply easy to understand terms how they can begin to get on the information highway. Since huge numbers of grassroots groups do not even have email this should be the first order of business, not simply creating more web sites which are useless to those who lack the communication tools to access them.
  6. A few users and consultants should be set up as test sites to experiment with future technology such as: two way audio/video; on-line, off the grid production of Internet "television"; and long distance voice calls over local phone lines. This will ensure that two years from now, when these applications have become mature technology, we will have obtained enough experience with them to incorporate them into our network.

Fighting the right may seem difficult enough without taking on the simultaneous task of making computers usable to the average activist, but in the American revolution our forefathers had to redefine the idea of how war was fought before they could beat the British. Our generals did not set out to invent guerrilla warfare, they just used the available tools in a fashion appropriate to the task.

Corporations tend to create networks and communication models appropriate to their hierarchical structure, and are therefore not a good model for us. They are like the British army: technologically timid and conservative, and institutionally incapable of applying new technologies till years after they are proven elsewhere. The future belongs to virtual everywhere-and-nowhere organizations that adopt new technology. They will defeat old fashioned hierarchical ones, just as our minute men were able to defeat the British in the Revolutionary War.

I believe we have the ability within the grassroots environmental movement to create a revolutionary communication technology; but we must keep the needs of thousands of small, technologically unsophisticated grassroots groups foremost in minds at every step of the way, and be very clear about what a grassroots network is for and what we expect it to do. We will need to be ruthless in our effort to render computers perfectly mundane.

©2000 Jim Britell
All rights reserved.
May not be reproduced without permission.

Home | Consulting Services | Biography | Index of Writings | jim@britell.com