CrossFile was the name given to the first GEIS-wide email system designed in 1976. Previously, there had been multiple systems on a per department basis. The (very small) team was lead by Malcolm Davies, who was my manager at Mark Century in London. George Feeney (then CEO) tapped Malcolm because of the simplicity, reliability and performance of the Interrnational email system that Malcolm maintained and enhanced, which had been prototyped (I think) by Dick van den Berg in the Netherlands.

The following story came from Malcolm:

"GEIS established international computer time sharing access, with a circuit from White Plains to Goonhilly, in December 1969. From the inception of Mark II in the UK, there were requests from companies to establish basic communications applications to reduce the costs associated with Telex. The Common Carriers strenuously opposed applications where “message content’ was the sole purpose. The CEPT cartel, which licensed the carriers, also declared that this was illegal. [CEPT was the controlling committee of the governmental post and telegraph organizations which gave licences and set rates for international postage, telex and telephone calls. They also decided how much of the cost of such communications would be shared between governments e.g. if you called from England to France it determined how much of the price went to the GPO and how much to the PTT. CEPT, founded in the 1870’s in Paris, used the Gold Franc as its internal unit of currency] CEPT finally had to address the fact that transatlantic (and in 1973 pacific) private data circuits were an established fact and decided merely to impose a tariff of 10 gold centimes per 1k characters of ‘textual message’ transmitted which was not modified by ‘substantial’ data processing.

By 1976 it was apparent that this was a failing rearguard action to preserve their monopolistic telex tariffs. The main challenger was a little private communications provider in the US called Microwave Communications Inc (afterwards known as MCI). Back in 1971 GEIS Europe had installed (quite legally since it was internal to GE) a messaging system to link their European offices and personnel. This system, COMCOM, was written by Dick van den Burg in Holland (he just retired from GXS in September 2004) and ran very well and saved thousands of dollars in telephone calls across Europe.

Additionally, in the US, Operations had established a communications system called Network Messenger (NM) to link computer centers and NCC’s.

In May of 1976 George Feeney thought that MCI’s activities in flaunting the common carrier rules would quickly open up the whole field of computer aided communications to commercial operation. I was summoned from London to Rockville at short notice (literally 30 hours) to give George a presentation on messaging systems. The meeting lasted about 2 hours with Hench, Leadley, Castle and Marshall attending. He decided to go ahead and asked me to set up a development and marketing team. This was another of George’s experiments to see the effects of having a multi discipline team under a single leader develop a product. The team was myself and Roger Dyer as designers/developers, Brian Garnichaud and Roberta Jankowski to handle Marketing and Logistics and a Documentation and Testing person (anyone recall the name?). We agreed on a specification by June (when I was allowed to return to London for the first time for my wedding) and we had the system code complete in August. By that time word had come down from GE Corporate Legal (Schlotterbeck?) that whereas a small company like MCI could flaunt CEPT regulations, a highly visible company like General Electric could not afford to do this. [Don’t forget that the phases of the moon court case in which GE colluded with Westinghouse to fix bids for power generation equipment and 7 VP’s went to jail, was still fresh in Corporate minds]

The project went ahead as an internal system to replace COMCOM and NM (there was also Memosys [RJD]). George had intended this phase anyway as a means of testing and stretching the system. We had all the logistics plans (to create a company wide directory of user id’s and distribute them), and conducted training sessions world wide in September and October 1976 when CrossFile or XFL (as the product was first called), went live. It was a great success and stood up to the load well. By December we had our first spam with big diagrams of Xmas trees crafted from asterisks, plus signs and other special characters that took 5 minutes to print out at 10 cps. We added a ‘delete unopened’ feature straightaway. The system was strictly character text messages; there were no screens or graphics in those days. By 1980 it was obvious that international data lines and usage had long made the arcane rules of CEPT obsolete and XFL was launched as the commercial product QuickComm."

To add to Malcolm's story - XFL was originally designed with a maximum of 1250 users, with 4 character addresses and pw's. Due to the growth of GEIS the address list was significantly expanded - I think the last version could take probably 4 times that number - and had encryption added in about 1979-80 as there was concern that operations people could read executive "secret" emails. I came up with a multipass, randomized key word and randomized packet size encrytion algorithm that was light on processing (we were expected to minimize resource use for internal projects), yet difficult to break. This hurt me when XFL processing was moved to Amstelveen and I had to spend time with the lawyers who were wrapped around the axle about exporting "restricted" technology - which is how any encryption algorithm was viewed even one for internal use.