Paul Bryant had one of the first IBM PCs at Rutherford Laboratory that were a gift from IBM. From that date in 1984 to 1993 he used WordStar for all his documents and kept them on floppy disks that have survived. Many historic documents held on the memories of computers that entered and left Atlas Lab have vanished with those machines or are archived on some media that is no longer readable. Paul's main interest was in communications during this period.
But to begin at the beginning... Atlas Laboratory was always dependent on communications. At first this was the postal services with users sending in decks of cards and rolls of paper tape and receiving back reams of line printer paper.
In the States there were quite a few experiments with interactive services and not to be outdone Mike Baylis proposed to front end the Atlas computer with a small computer linked to the Atlas via a shared disk store. The computer chosen was a Sigma 2. The project was to enable users to edit files on the disk and submit jobs and get the output back for display on terminates. The terminals were the 10 character a second teletypes. The real development work was done by Eric Thomas and John Baldwin.
The National Engineering Laboratory at East Kilbride had a Univac 1108 computer that had spare capacity and Atlas Lab were instructed to use it. This was done by Paul programming the Sigma 2 to emulate a UNIVAC card reader/line printer workstation. Although a technical success the link was not much used as converting jobs to run on the 1108 was a bit tedious and the Lab was not all that short of capacity.
The ICL 1906A computer was completely different to the Atlas having the George 4 operating system that was designed for interactive use. Thus many more character terminates were installed together with 7020 card reader/line printer terminals at remote locations. As these were expensive the Lab built its own device built of a GEC2050 computer. Rutherford built a similar device for their IBM computer again based on the GEC 2050. And Daresbury did the same but on a DEC PDP 11. Thus we found that all three SRC Labs were festooning the countries universities with these devices at great expense. At this time 1973 the Post Office (as it then was) set up EPSS (the Experimental Packet Switched Service) based on ideas arising from Donald Davies from NPL and Louis Pouzin of INRIA (France). The three Labs played about with this network but eventually decided to set up its own network but this time based on the emerging X25 protocols. The X25 switch was based on a GEC4080 coded by Andrew Dunn. Daresbury built a switch based on a PDP 11.
By good fortune this network was working by the time the Interactive Computer Facility started in 1976 as a result of the Rosenbrock report and some 40 odd GEC and PR1ME machines were eventually connected to this network.
The archive covers a time of rapid network development in which the Laboratory played an extensive part. 1984 was the year in which the X25 network developed by SERC became the JANET network available to academic sites. This was probably the first national academic network in the world. Before this time there were a number of networks across the world that met the needs of various communities. Rutherford had pioneered this work with the development of the Science and Engineering Council network that connected its large central machines as well as the 50 or so computers distributed around university sites many as part of the Interactive Computing Facility. From 1984 international academic networking was starting that underlined the need for common standards. Between 1984 and 1993 the negotiations and battles over what standards should be adopted or imposed raged but the emergence of the World Wide Web was the catalyst that finally tipped the scales in favour of the now ubiquitous Internet Protocol.
At the start of the period much effort was going into the development of the networks according to the standards supported by the Joint Network Team. At first these were the Coloured Book protocols based on X25 for the wide area network and the Cambridge Ring for the local area but was expected to move to International Standards Organisation (ISO) protocols as time progressed. During this period international networking between the academic communities started development based on ISO protocols supported by the European Economic Community as the European Union was called in those days. Progress was at snails pace due to the difficulties of developing the protocol standards and then implementing them and this was not helped by European bureaucracy. Meanwhile the European Academic Research Network was set up with IBM money using IBM protocols at a very rapid rate. This galvanised considerable opposition from the network aficionados, who were rather powerless, as they had no quick alternatives to offer. In the USA the IP protocols were becoming popular with a lot of pressure to allow their use in the UK and Europe. Thus the protocol wars erupted and as history records IP won.
Paul worked at developing the Coloured book and ISO world but also got involved with European Academic and Research Network (EARN) and became unpopular in both camps but being pragmatic attempted to provide a service to the users. It was unfortunate that in the desire for universal standards, the perfect world, providing a service in a timely fashion was easy to forget.
Paul was also active in setting up and running a local area network on the Rutherford site. Initially this was X25 but in line with UK aspirations and the Common Base Policy he installed a quite large Cambridge Ring. This worked but was hampered by a lack of products and in particular the much awaited "ring chip" that lost several reputations. So when Ethernet equipment became available the Lab moved quite rapidly to install it. The Cambridge Ring was invented when hardware was difficult to produce and the Ring protocol was quite simple with most of the work being done in software. On the other hand Ethernet was complicated and demanded chips but once developed led to cheap and reliable networks.
Post 1993 networking became a matter of installing commercial products ranging from structured cabling, through routers and bridges to servers. The fun days were over.
The documents in the collections are deficient in that many meeting papers and received letters have been lost and so quite a bit of imagination has to be used when reading letters or looking at meeting minutes. A lot of the correspondence shows the irritation with the powers that be. In particular unending negotiations for a licence to run EARN from BT and unending correspondence on sorting out the functional standards for the so-called triple x protocols. There is also documents showing the tension between those who wanted to stop EARN developing and those running it.
It is difficult not to ponder on the immense waste of effort that went into developing networks with standards that are now long departed and the anger and frustration with the protocol wars but maybe that's the way of the world.
Many of the later papers in Paul's Archive relate to meetings of the EARN Board of Directors or the EARN Executive Committee in the period 1984 to 1992 when he was heavily involved with both Committees. For the years 1989 and 1990, EARN had a more formal organisation towards meeting papers. Paul's complete set of papers for the EARN Board of Directors and the EARN Executive Committee for the years 1989 and 1990 together with some EARN Information papers are included here as sub-directories.
