SatelLife was originally conceived by the Nobel prize winning International Physicians for the Prevention of Nuclear War (IPPNW) as an east-west collaboration to address the increasing north-south disparity in current health information. It was recently chartered as an internationally directed, not-for-profit organization to link medical centers and health providers with critical sources of information throughout the world. The Anglo-Polish Information Retrieval Service has become one of its first operational projects.

Ready access to available information is taken for granted by researchers in all developed countries through telephone and electronic mail to their colleagues, through online access to computerized professional and bibliographic databases, through membership in professional societies, through participation in professional meetings, through journal subscriptions, and through access to professional libraries which in turn have access to a comprehensive network of cooperating institutions. Such is rarely the case in the developing world. Severe lack of funds and lack of a well developed or functioning technological infrastructure may make even a telephone call a difficult task to complete. Lack of hard currency funds and economic belt tightening imposed by such organizations as the World Bank and the International Monetary Fund have further aggra-vated this situation. While United States medical school libraries have collections of 2000-3000 journal titles, only 23 of the approximately 250 medical schools in Latin America hold more than 300 serials, and many hold none (Bruer et al, 1981; Saracevic, 1988).

At the 8th World Congress of IPPNW in Moscow, it become clear that lack of ready access to current literature and research results was becoming a major handicap for researchers also in the Eastern European countries. The Central Medical Library of Poland is the equivalent of the US National Library of Medicine. It maintains a significant collection of foreign publications. How-ever, this collection was significantly compromised by the recent economic crises in Poland. The CML has had access to international computer-based information retrieval services (MEDLARS) by telephone. But even this service had to be discontinued recently because of the lack of foreign currency funds. Currently, the CML uses various computerized bibliographical services available on compact disk, such as MEDLINE on Silver Platter. However, providing access to full text documents is a continuing problem and is often impossible.

To aid the CML in this difficult time, SatelLife has obtained a grant from the A. Jurzykowski Foundation to explore means of improving the CML's access to the international medical literature.
 
 

2. THE ANGLO-POLISH MEDICAL INFORMATION RETRIEVAL SERVICE

SatelLife has obtained the cooperation of the Library of the British Medical Association in London, England, to provide the CML with services similar to those provided to members of the BMA. Requests from Polish researchers to the CML, which cannot be satisfied by the CML, are forwarded to the LBMA via telex, fax, electronic mail through PEACEnet via Sweden and what-ever other means that are currently available and operable. Plans were to return requests for copyable materials via fax from London to Warsaw, allowing a potential turnaround for requests of one to two days. However, the lack of operable and reliable telephone circuits in Poland made this a frustrating and unacceptable option. It would often take hours to obtain a telephone connection and then the connection could be interrupted before completion of the whole transmission. Other methods for document transmission had to be explored.

While mail and courier services are useful and available, SatelLife was particularly interested in exploring electronic transmission media, which could take advantage of the excess satellite capacity over Europe. SatelLife was also interested in a system that could be extended easily for communication to other Eastern European countries with similarly constrained library and commu-nication systems. A donation by Microspace Communications Corporation of free access to a satellite broadcast channel, with broadcast capability to Eastern Europe, provided the opportunity to develop a fax-type transmission system using broadcasts by satellite.
 
 

3. SATELLIFE'S FAX BROADCAST SYSTEM

An initial design criterion for the fax broadcast system was to make it operate similar to a nor-mal fax system, to use off-the-shelf hardware and software as much as possible, and to provide for completely automated transmission over the satellite. Normal fax operation involves a standard protocol, which includes establishment of the telephone connection (e.g. dialing and line monitor-ing), negotiation with the answering fax machine of transmission characteristics and error control (e.g. transmission speed and requests for retransmission of faulty transmissions), and termination of the telephone connection after completion of the fax transmission. Inherent in this protocol is the requirement for two-way (duplex) communication, in which each fax machine can respond to requests from the other one. The telephone provides this type of a connection. In contrast, the satellite broadcast channel is inherently one-way (simplex), from the satellite to the receiving ground-station only, disallowing direct connection of fax machines to the broadcast channel at either end. These conditions required development of specialized software to work around these difficulties.

Figure 1 depicts the current prototype system as a block diagram. In "London", the output from a normal fax machine is intercepted by a microcomputer and stored as a computer file, each faxed page as a separate file. These files are then transmitted from "London" to the satellite uplink station in "Raleigh" for broadcast, using standard computer communications procedures. In turn, the broadcast files are received at the destination ground station in "Warsaw", stored again as computer files and then processed for transfer to an appropriate output device such as a laser or dot-matrix printer.

Custom software has been developed for the "London" station, which allows normal use of the fax machine for scanning of material. Transfer of saved files to the satellite uplink station is automatic at the termination of input. In addition to monitoring the fax machine, the software modifies the received files to allow easier transmission and reception at the destination. (In the future, address information about the designated recipient also will be added at this step.) Soft-ware at the uplink station in "Raleigh" stores the received files until broadcast, which is initiated at the end of file reception or at another designated point of time. The receiver in "Warsaw" must be switched on and operating at the beginning of each broadcast. Again custom software has been developed which continuously monitors the satellite receiver and preempts the computer for storing incoming files whenever there is satellite transmission. (In the future, this program will check the designated recipient information and skip over transmissions not destined for the respective groundstation.) Lastly, another program prepares the received files for output on a number of computer devices such as video screen and printers.

The system is operational as installed, however it does not include error correction of
 

transmissions over the satellite. Error free transmission of binary files between computer systems is routinely handled by common communications software using well established protocols, but all these protocols require again duplex communication channels. Work is currently in progress to include forward error correction, which will increase significantly the overhead of transmissions. Only operational experience with the system will determine the frequency and severity of actual transmission errors.
 
 

4. ECONOMIC FEASIBILITY PROJECTIONS

A major motivation for the development of the fax broadcast system was the ability to reach locations which do not have any or only unreliable accessibility by telephone or other electronic means. In such situations, the use of satellites eliminates all needs for local services except electricity. The components required for local operation of such systems are becoming increasing-ly available and affordable. The basic groundstation consists of a small dish antenna, a book size receiver, and a microcomputer with printer. The antenna and receiver cost between $200-400, depending on the size of dish required (marginal reception areas require larger antennas). Appropriate microcomputer systems may cost less than $1000.

The following cost projections are based on currently published rates for commercially offered broadcast services to Europe from PanAmerican Satellite #1 (PAS-1). The uplink to the satellite is located in Raleigh, North Carolina. The satellite broadcast can be received anywhere within the satellite's footprint (Fig. 2). Satellite channels can be leased at the standard BAUD rates

for asynchronous serial communication. Transmission rates of 2400, 4800 and 9600 BAUD were selected for the cost comparison, higher BAUD rates permitting proportionately greater through-put. The satellite transmission cost estimates include telephone charges for transmitting material by telephone from Boston to North Carolina. In addition, the calculations include a 25 percent over-head allocation allowing for the future inclusion of error correction. However, these costs do not include the amortization of the groundstation equipment.

If reliable telephone service were available, the telephone cost of fax via telephone provides a benchmark for cost comparison. The telephone cost included in Figure 3 represents the average cost for the transmission of one page via telephone from Boston to Poland, assuming uninter-rupted, error free transmission. As Figure 3 indicates, the satellite alternative becomes competitive for transmission rates above 300 pages per day and becomes very competitive for higher rates, approaching $0.50 per page at 1500 pages and above. Table 1 summarizes some of the assump-tions and estimates used for the calculations in producing the graphs of Figure 3.

The cost projections above include a number of worst case assumptions in terms of costs and use of the technology. Published rates of identical services for broadcast over the United States are half those quoted for Europe. Thus, there may be room for significant cost decreases in the future. Distribution of non-pictorial material via fax is extremely inefficient. The projections assume the transmission of the faxed image of a page. The faxed image of a page is typically at least ten times larger than the equivalent text representation. A full page of text or data contains between 3000 to 4000 characters, but its fax image will be between 30000 and 60000 characters. Thus, transmis-sion of material as text or data would reduce required transmission capacity by a factor of ten or

more. However, the most important cost reduction will come from "appropriate" use of the

transmission technology. The cost projections assume only transmission of material to individual recipients, replicating the capabilities offered by fax for telephone. The advantage of the satellite broadcast technology lies in the capability to distribute material simultaneously to any number of recipients. Every groundstation within the satellite footprint can receive any broadcast. In situa-tions, where the same material is to be distributed to several recipients, only a single transmission will be required and the cost of this transmission would be shared by all recipients. In this situa-tion, the distribution by satellite broadcast becomes highly economical.
 
 

5. NEXT STEPS AND FUTURE APPLICATIONS

Transmission of the first pages of material have proven the technical feasibility of the fax broadcast system. Simplification of the batch and script files which implement the automated operation of the system will make the system more robust and transportable. Addressability of transmissions to designated recipients as well as inclusion of some form of forward error correc-tion are being explored and will be included in the future. Current plans also include the imminent extension of the Anglo-Polish project to other Eastern European countries.

The development and use of the current system was strongly influenced by the potential for donations to SatelLife. Improvements in the infrastructure of countries in Eastern Europe may reduce the need for the current service in the future. However, this type of technology is ideally suited for distribution of common materials to many other areas around the globe. Similar projects could be developed using such technologies as videotext for distribution of material as for example replacement of the repetitive broadcast of the typical videotext material to television receivers with one time broadcasts to microcomputer supported receivers. Such a scheme would allow use of existing and often unused broadcast frequencies for transmission of text, data and pictorial material.

While such systems are eminently feasible and cost effective, as the above projections demons-trate, there is though one major obstacle to their effective use. Current copyright laws severely limit the distribution of copyrighted materials to individual recipients and prohibit the distribution of copyrighted material to multiple recipients without prior approval. In the case of SatelLife, it is hoped that such approval can be obtained for selected material which is important for scientific and humanitarian reasons and which cannot or would not be distributed by any other means, as in the case of most developing countries.
 
 

REFERENCES

Bruer, Goffman & Warren, "Selective Medical Libraries and Library Networks for Developing Countries," American Journal of Tropical medical Hygiene, 30 (5): 1133-1140 (1981).

Saracevic, Tefko, "Selective Medical Library on Microfiche,"Bulletin of Medical Library Association, 76 (1): 44-53 (1988).