*************************************************************************** P I G U L K I an occasional electronic collection of news analysis, press reviews, and humor from/about Poland and the Polish community abroad ****** Oct 8. 1990 ******************************* No. 3 ***************** In this issue: Marek Zielinski: How BITNET Came to Poland Jan P. Radomski: Computers in Polish Universities Chris Heller: Computing Power for Poland, ====================================================================== HOW BITNET CAME TO POLAND THE BEGINNINGS In September 1988 a discussion group "Bitnet to Poland" [1] was formed, the purpose of which was to connect the international academic computer networks to Poland. It was an initiative of Dave Phillips and several Polish students and post-docs, who happened to be on the same network. We started night discussions on RELAY, wrote an article to NETMONTH [2] and poked into various corners and talked to some Very Important People. We learned that 3 countries from "Eastern Europe": Hungary, Bulgaria and USSR had already applied for membership in EARN. None had been yet granted; the main if not only obstacle was COCOM [3] restrictions on export of sensitive technology, especially of supercomputers. We did not know if there were the technical capabilities in Poland, besides none of Polish Universities had applied yet. We focused first on locating people who would turn the appropriate wheels and procur applications from the Polish side. Wroclaw Polytechnic was first to apply for membership; soon we were contacted by a group of physicists from Warsaw University, who were thinking about such link for some time. Lodz Polytechnic joined also and all of them sent letters to Dr. Dennis Jennings, then president of EARN. In the middle of 1989 EARN had four applications from Poland (the fourth was from Polish Academy of Sciences). KASK Poland had been developing for some time an ambitious network project called KASK. Quoting from an official document, The Polish academic computer network, called KASK, is contructed under the governmental order CPBR-8.13. A primary objective of the undertaking is to establish a computer network in order to provide the many thousands of academic teachers and researchers with reliable and easy to use communication. (...) (i) integration of existing academic computers and computers (...) (ii) design of the superstructure over the public data network, POLPAK, providing advanced technologies of communications, e.g. electronic mail; (iii) public access to data bases and information services; (iv) development of technologies for the connection of national network to international information systems. It was to be all BITNET is and more... Unfortunately, the in-house developed protocols were not compatible with EARN, and the "governmental orders" were rather tools to extract money from the communist government than serious projects. For the purpose of connecting to BITNET, KASK appears to be more a liability than an asset. Some experience in networking and perhaps wires connecting the cities could be useful (but not the "MERA" computers to be the base of KASK), while a number of people ambitionally engaged in the project would certainly create problems. POLISH GOVERNMENT In the meantime the political changes in Poland were accelerating. First the "round table talks" and then the first semi-free elections brought us the government which we could call "ours". Minister Stefan Amsterdamski was approached by the Warsaw group, who presented him with the idea of connecting Poland to EARN. Prof. Amsterdamski was enthusiastic and as a result the government agreed to provide funds for the phone line to Denmark, membership dues and some extra for hardware and software to establish the central node in Warsaw. Other academic centers were to find funds by themselves. MAILBOXES The electronic contacts with Poland were soon becoming more frequent. Although there was no EARN node in Poland yet, a number of researchers participating in joint programs with universities in Germany, Denmark and Switzerland obtained accounts on the university computers, which they could access also from Poland. The international phone calls are not cheap and low quality of connection often required batch transfer of jobs and mail using an error checking protocol (Kermit), but it was an opening to the world. This way "Donosy" were born, a daily news bulletin, very enthusiastically and solidly edited by a group of physicists from Warsaw University. It is difficult now to meet a Pole, with or without network access, who does not subscribe or hasn't heard of "Donosy". Later, Frode Greisen, the present president of EARN granted Wroclaw Politechnic an access to mailbox in Denmark, pending the permanent connection of Polish nodes. TOP LEVEL TALKS The official representatives of CREN and EARN worked hard on the most important aspect of bringing Bitnet to Eastern Europe: obtaining permission from the US Commerce Department. Dr. Ira Fuchs, Bitnet founder and now president and CEO of CREN said in an interview for "Science" that the political changes in Eastern Europe convinced him that the time was right to seek such permission: "We have been trying to push Commerce to give us an answer. I figured if ever there was a time to make it happen, this was it"[4]. Also Dr. Dennis Jennings, former president of EARN as well as its current president, Dr. Frode Greisen, were very sympathetic to the idea of bringing the network to Poland. Without their active involvement in this operation, its success would be very much in doubt. LETTER FROM THE COMMERCE DEPT. At the beginning of 1990 the Bitnet Network Information Centre, EDUCOM, received a letter from William Clements, the director of the Office of Technology and Policy Analysis, US Department of Commerce. The letter was in response to inquiries of Jim Conklin, EDUCOM's director for an opinion on applicable US export regulations. The seven-page letter was written in proper legalese and its meaning it was not immediately evident. Further study by CREN's lawyer, however, yielded on February 7 the long awaited YES: "In conclusion, counsel recommends that the proposed connection of entities within Eastern European Countries and The People's Republic of China to the network proceed via entertainment of such applications for processing according to existing membership guidelines"... Almost concurrently COCOM, in a meeting in Paris which ended Feb. 16, agreed to substantially ease export restrictions to countries of "Eastern Europe", including computers, telecommunication equipment, etc. EARN soon begun considering the applications. The biggest obstacle was overcome, a large number of smaller problems remained. Starved in the communist regime, Poland lacked almost everything: computers, quality phone links, specialists. PLEARN-L The lack of experience was probably the biggest problem in effectively putting together a large network in Poland. On a RELAY meeting of the "Bitnet to Poland" group we have decided to organize a new discussion list, PLEARN-L [5], in order to bring together networking experts willing to help and those who needed advice. Dave Phillips was the one to make the arrangements and open the list in Buffalo. The list became a meeting place for the network organizers, and helped initiate many contacts which were further continued individually. It also serves as a clearinghouse for information on the status of connections to different cities and recently also to Chechoslovakia, Hungary and the Baltic countries, Latvia, Estonia and Lithuania. EARN VOTE On April 8 the EARN Board of Directors concluded the vote on new membership. The result was YES for 5 new members, Poland, Hungary, Czechoslovakia, Bulgaria and USSR. Immediately after, a meeting was held in Warsaw between EARN group headed by Frode Greisen and representatives of Polish government and interested Universities, where the decision was announced. The central Polish node (PLEARN) would be a BASF computer (clone of IBM/370) located at the Warsaw University Computing Center. The node was to be be connected to DKEARN in Copenhagen. Tomasz Hofmokl was nominated the EARN-POLAND director, and Andrzej Zienkiewicz became the Technical Coordinator. PLEARN ON LINE The ball was now in the court of Warsaw University Informatics Center. The line to Denmark was ordered, as well as software and modems, and finally PLEARN went on-line July 7. Initially the node was working 9 to 8 weekdays only, but since September 29 it is operative round the clock. Andrzej Smereczynski is the operator and a number of people from Warsaw University have already the accounts there. Five other Polish nodes are also listed in the node tables, PLWAIP51 Institute of Physics, Warsaw PLWAIP52 Institute of Physics, Warsaw PLWAUW21 Informatics Center, U. of Warsaw PLWRTU11 Technical University, Wroclaw PLWRTU51 Technical University, Wroclaw as well as Czecho-Slovak and Hungarian nodes CSEARN Czech Technical University, Prague HUEARN Computer and Automation Inst., Hung. Ac. of Sci., Budapest., none of which is operative yet. TRAINING The French office of EARN invited two people from Wroclaw University and Wroclaw Politechnic for a network training. In their report to PLEARN-L they wrote, "...we are listening to very important lectures about organisation, installation, management and administration of the EARN. It is in EARN Europe Office in Orsay. Daniele Bovio is our 'GURU' in Earnland, and Elizabeth Porteneuve our Guardian Angel." In October the "Bitnet to Poland" group is organizing training with US "gurus" for prospective node operators from Lodz and Warsaw and the International Centre for Computational Physics in Trieste announced the "First International School on Computer Network Analysis and Management" to be held in December. Chris Heller is the main mover of the Trieste initiative. The demand is big and such courses will probably be repeated. OTHER NETWORKS BITNET is a "store and forward" network with no adaptive routing. Your message (piece of mail, a file or an interacive note) is sent to the nearest node, which forwards it further along a predetermined route, but only if the next link is available. As the number of intervening nodes between the source and destination increases, the reliability goes down. A quick calculation shows us, for 10 intervening nodes that if an average mainframe is down for repairs say 6 days a year, the connection will be inoperative one day every week. (The proposed Bitnet II is supposed to alleviate this problem) For this and other reasons, Bitnet is shunned by real computer "pros" who prefer to use Internet, a collection of networks using common protocols known as TCP/IP. The purpose and general philosophy of the two networks is, however, very much different. The main feature of Internet is the remote login procedure, where a link is established between the *user* and a remote computer. It allows the user to perform work on that machine, transfer files, etc. Consequently it is used by those who have business in using remote computers: physicists, mathematicians, computer scientists... On the other hand BITNET can efficiently transfer *files*, which jump from node to node: mail files, interactive messages, sometimes also binary files, but it does not provide for a login to a remote machine. BITNET is used widely by students, teachers and scientists for general COMMUNICATIONS: effective, quick and *informal*. Poland has been granted access only to "BITNET" functions of the network, explicitly excluding the remote login possibility. Recently however, "Donosy" (9 August 1990) reported that a customs agreement has been signed between Poland and USA. Poland agrees not to rexport any American equipment or technology without exporter's consent. In return Poland will be exempted from COCOM limitations. This agreement should soon open the door to the full participation in the worldwide networks. 9/30/1990 Marek Zielinski --------------------------------------------------------------- [1] BITNET (Beacuse It's Time NETwork) formally does not exist any more; after merging with CSNET it aquired a new name of CREN (Corporation for Research and Educational Networking). Together with European EARN and Canadian NetNorth it forms one supernetwork without a name. Poland for obvious reasons joined EARN. Throughout the article we use the term "Bitnet" loosely, to indicate the whole network. [2] NETMONTH 3(3), 5 (Sep. 1988). NETMONTH is a semi-official electronic newsletter of CREN, edited by Chris Condon. For subscription send message to LISTSERV@MARIST saying "SUBSCRIBE NETMONTH Your_Full_Name". [3] COCOM is the 17-nation "Coordinating Committee for Multilateral Export Controls", set up at the height of the Cold War to coordinate the various countries' export policies and composed of most NATO members as well as Australia and Japan. The aim was to prevent the Eastern bloc countries from obtaining Western technology that could be militarily useful. [4] J. Palca, "BITNET Headed for New Frontiers", Science 247, p. 520 (Feb. 1990) [5] To subscribe to PLEARN-L send mail or message to LISTSERV@UBVM saying "SUBSCRIBE PLEARN-L Your_Full_Name" ======================================================================== COMPUTERS IN POLISH UNIVERSITIES (an open letter) Dear Friend, You have asked me in one of your previous letters about computer situation here in Poland. And I remember my lighthearted answer that on the whole it is not so bad, as we even have some fast 33 MHz 80386 clones. Since that time nothing has changed very much, but my outlook certainly did. For the worse. The cover to cover lecture of the anniversary, September, issue of the BYTE magazine had a sobering effect. The Department of Chemistry at Warsaw University is the only one among similar places at various universities in Poland, which was awarded a prestigious name of Faculty. All others including the Jagiellonian University in Cracow, and the Universities in Wroclaw, Poznan, Lodz, Gdansk, Torun and Lublin have their Chemistry included in Mat-Phys-Chem departments, or at best as the separate "Institutes". This distinction is of course nothing in itself, rather, it reflects attitudes which stems from Byzantine inclinations of comunist rule, or even from the rigid, pompous, and vainglorious structures of the past Austro-Hungarian Empire. Nevertheless our deptartment was funded and build over fifty years ago (the present building was erected especially for us on the brink of the World War II). What matters, however, are our successes. Among our best groups is the Laboratory of Quantum Chemistry headed by Prof. Kolos, of international calibre and renown (one needs only to mention the Kolos' and Jeziorskis theoretical framework allowing to measure the neutrino mass - if it only has any, or Pielas method of finding global minimum of the macromolecule global energy surface). The only, but vital, tools of this lab are computers. As you certainly know, Quantum mechanics requires even for its simplest application huge amounts of both RAM and disk storage, and the CPU speed is essential as well. The best machines my department owns, however, are those 33 Mhz PC clones. We have had the two SM-4 minis (DECs PDP-11 russian clones), but both are defunct now. The last working terminal died some time ago, and besides even at their best the SM-4 were never very effective. What is the geography of our resources then? On the one pole there is my Organic Chemistry Deptartment about 50 strong staff including seven senior profesors. In our disposal are six PC-XTs, some of them even without hard disks. The two oldest are running at the mere 4.7 MHz, and with mindboggling 256 kB in 16k chips. One of these 6-year old veterans does not even include the 8087 number smasher. On the other pole, the Quantum Chemistry Lab uses three 80386 (33, 25, and 20 MHz), one 6 MHz AT, and two XTs (4.7 MHz, but I am not sure). For hard copy they have an old 9-pin dot matrix, and a HP LaserJet III. Among their best buys is external 600 Mb hard disc drive. Of course, such riches are subject of the strong envy for all of us, lesser mortals. They are generous, however, and allow us to use their hoard if we are desperate. Between these two poles you could find all kinds and variances of the PC clone theme, from Singapore, South Korea, and Taiwan. What I have forgotten to mention is that at least one of the Quantum Lab 386's is of American origin, and should it not have been so difficult, we would like to place all our orders there. Paradoxically and contrary to expectations - USA marketplace offers not only better quality, but also cheaper buys. And it is a real pity that USA vendors are seemingly not aware of our needs. What is more, despite of the softening in COCOM policy towards Poland, we all too often do hear about difficulties from US customs and other petty bureaucracy. It is kind of funny how such officialdom is always sawing and hacking at the branch on which their country economy seats on - all over the World. There is a little hope for a quick change. Mainly because of Poland monetary situation, but also because of the attitude of our own local decision makers, among whom (with a few notable exceptions) the computer literacy is not the strongest point. At other institutions of higher education the picture is more or less the same. For example, Institute of Computer Sciences, subsidiary of Polish Academy of Sciences, boast of the two VAXes 750 - the ONLY two in Warsaw (the nearest one is almost 400 kilometers westward, in Wroclaw). And these two machines constitute the best computing opportunity we have, but theit use is not cheap. Then, all of sudden, this situation is flash-lighted by the fact that tiny link connects Poland with EARN/BitNet network. At last. It is my strong conviction and hope, that importance of this link can not be understated. And I think that on the small backyard of our labs and benches its significance quite soon will be compared with the summer strikes of 1980 and their effect afterwards. Nothing is perfect, however. At the present the main and only node of PLEARN is the old BASF (IBM 370) machine at the Computing Center of our University. To connect our FACULTY we are in strong need to: a) obtain an adequate computer for the end node and file server, b) get neccessary software, c) lease the telephone line from our centralized, state owned company, which is almost as impossible as before (during the communist regime), and d) connect our PCs toogether and with (as yet absent) fileserver. Easy? You might think so, but it is not. To begin with, our PCs are anything but personal. There are periods that queue to some is several days long. Then, of course, there is the money problem. But above all - the need for a-d) must be recognized, and I expect that it will be quite a feat to explain this to our friends abroad. The second point is our complete lack of experience, nobody has done here anything like that before. There are some firms offering all kinds of solutions, but one must be extremally cautious. The honest vendor is not quite here yet, again the heirloom of the past rule. So, here you are, with complete, if somehow biased (but it is difficult to be not emotionally involved) a picture. Please, do forward this letter to all who with their experience or otherwise might help to resolve our Gordian knot. If some of you have in yours lofts or backyard sheds any equipment in working order: old PCs, expansion cards, drives or controllers, monitors and all that, which you consider to be old junk but are hestitant to simply throw away - SEND IT TO US !!! Even if it will be of no use to make serious science on, we could use it for training, for ourselves or for our kids. I am looking forward to hearing soon from all of you. Greetings from postcomunist Poland, Jan P. Radomski. or PS. If you could help me to get in touch with three BYTE readers whose out-look for the future I find much more visionary than that of many experts, I will be really happy. Here they are, all from the page 36 of that September issue: Jeanette J. Bieber-Moses, Billings, MT; Richard Crews, San Raphael, CA; and Charles Bridgeland, Urbana, IL. ========================================================================= COMPUTING POWER FOR POLAND Let me first make some corrections to factual errors in Radomski's letter, and then continue with more general remarks and a proposal. 1) The Department of Chemistry at Warsaw University is not "the only one among similar places at various universities in Poland, which was awarded the prestigious name of Faculty". Approximately 10 years ago Institute of Chemistry at Jagellonian University was transformed into separate Faculty of Chemistry. On the other hand I do not see what is the connection between this and computers... 2) I do not think that "quantum mechanics requires even for its simplest applications huge amounts of both operating memory and disk storage, and the CPU speed is essential as well". Quantum chemistry, on the other hand, probably does. 3) I find it hard to believe that the situation in computing resources is so desperate as the author portrays. Even now at the Institute of Physics, Jagellonian University, Krakow, several Definicon boards (which outperform VAX 750 machines as far as number crunching is concerned) and two 68030 based Cromemco machines are at work, not to mention hordes of 386-based PC's. Purchases of superworkstations (like IBM RISC/6000 series) are planned. I do not know about other institutes or universities, but I cannot imagine that institutions based in Warsaw have less money that their counterparts in Cracow. The above remarks notwithstanding, I agree with the general conclusion of the letter: people in the universities need bigger and faster computers in order to be able to perform calculations vital for their research. Present and coming machines will not substitute Crays, and I doubt that any Polish educational institution will be able to afford one in close future. But there is some hope for an almost immediate relief, and this is where my letter comes to its conclusion: International Centre for Theoretical Physics, Trieste, Italy is now working on a project where people from developing countries can submit programs for execution by means of e-mail (including X.25 connection) to the supercomputer in Trieste. The software for the project (first release) is nearing completion and pilot facility should be operational within two months. At present the weakest point is the computing power available; there is only one Convex C-210 supermini (50 Mflop, 128Mb memory) working for the project. But when the facility proves useful (which means many inquires from interested parties, hint, hint) some additional funds could be allocated, and maybe even special center created for this project only. This could mean extending Convex (by adding more processors and memory) or even buying an entry-level Cray, if there is justification. This is at present only in a very preliminary phase, but I think that enquires could help to prove that this project is indeed very useful. 'Official' letters should be addressed to: Prof. A. Nobile, ICTP, P.O. Box 586, 34100 Trieste, Italy. E-mail inquires to myself: HELLER@ITSICTP.bitnet. Krzysiek (Chris) Heller HELLER@ITSICTP.bitnet ========================================================================= PIGULKI - Editors Jerzy Klimkowski (Incommunicado, USA) email: jzk@vanity.ncat.edu Dave Phillips (Cheektowaga, USA) email: davep@acsu.buffalo.edu Jacek Ulanski (Lodz, Poland) (email: pending, really, pending!) Marek Zielinski (NYC, USA) email: zielinsk@nyuacf.bitnet Note: PIGULKI is distributed free of charge to masochistic readers who request it (send requests to davep@acsu.buffalo.edu or to zielinski@nyuacf). All signed articles are Copyright (c) 1990 by their authors. PIGULKI may not be copied or retransmitted without prior permission by the editors and notification of your local public health authorities. Your articles, letters, threats, denunciations are welcome; please send them to any editor you can find who'll admit being one. We reserve the right to edit for brevity. ======================================================================== .