Cookham Training 1961-62

In the dim and far off days of the early 1960s, new recruits to the ICT sales force joined as trainee Technical Advisors Grade 1 and spent 9 months or so in training before being unleashed on unsuspecting customers. Most of the time was spent at Moor Hall in Cookham learning the rudiments of accounting and all there was to know about every item in the Holerith and Power-Samas ranges of punched card machines. There were two breaks from the Cookham routine when the trainees were assigned to sales offices for periods of between four and eight weeks. The final term back at Cookham was centred around producing a fully documented sales proposal in response to an invitation to tender. The culmination was a sales presentation to the Principal of Moor Hall.

The students were divided into teams and each had to make their own choice of hardware to propose, produce the sales report and make the sales presentation. The team that I was assigned to included a man who could type and who had a portable typewriter. This was a rare and valuable asset in the era before word processors. For some reason the typewriter had a Spanish keyboard but this was only a minor disadvantage and it was thought that the occasional and random addition of ~ would add a certain interest to the text - anything to distract the reader from the style and content. The team also included someone who had been, and perhaps was still, an aspiring politician and public speaking held no terrors for him. This was also thought to be a major advantage.

Come the day, our proposal was completed, our presentation was rehearsed and we, and the other teams, were ready to make our presentations to the Principal, a saintly and ever patient Scotsman by the name of George Thompson. Come the time, our team was invited to make its presentation. It was at about this time that things started to go wrong. Our lead salesman seemed to get stage fright. Instead of clamming up, which at least would have given the number two a chance to take over, he went into overdrive. He launched into tirade about the history of punched cards and nearly everything that it was possible to do with punched cards.

It was magnificent. It was irrelevant. It was non-stop. It was a disaster

After about 15 to 20 minutes of this everyone but the presenter could see that even the Principal's patience was wearing a bit thin. The Principal did his best to break the flow by offering the budding salesman a cigarette. This was declined without even a pause for breath or a break in the flow of his diatribe. A call to the telephone went ignored. Eventually the presentation was brought to an abrupt halt by the Principal and it was all over. We never did get to present our solution to his business problem.

In the subsequent post mortem of all the proposals and presentations it was a little galling to be told that ours was probably the best proposal of the three. It was however good practice for the 'real world' when on more than one occasion I had to try and look grateful when a prospect said YES ICT/ICL had produced the best proposal but NO they were not going to get the business.

--- 2 ---

Cookham Graduation 1961-62

The very last day of the initial training for Technical Advisors was a visit to the Demonstration Centre, then in Piccadilly, to show off one's expertise, or otherwise, in operating and demonstrating ICT's many and various punched card machines, to a selection of ICT sales managers. It was customary to hold an end of course dinner somewhere in the Cookham area on the last evening. Our course adhered to this tradition and also to the tradition of being rowdy and obnoxious after we had returned to Moor Hall. On this particular occasion the Principal took exception to the performance and read the Riot Act. In addition he instructed the lot of us to appear before Cedric Dickens, the UK Sales Manager, the following morning. This we did and received a further dressing down but nothing worse.

It used to be said that to get to the top it was important to get ones name known at the top. It did not necessarily matter how one got one's name known, just get it known. It does not seem to work in every case!

The Customer is always right; well almost always. c. 1965

My claim to the authorship of the least useful and least used piece of software would be for the core store sort subroutine that I wrote for the 1004. The customer detailed the requirement as part of a program specification and subsequently insisted that it was absolutely essential. I wrote the code, plugged the board and tested the program. The customer then decided that he could manage without it!

Unfavourite Location 1966-68

Over the decades ICL and its forbears have occupied a great variety of offices. To put it mildly, many of them have been a bit on the grotty side. My nomination for the grottiest would be 5 - 11 High Holborn, opposite the Prudential Headquarters. Apart from its convenience for Chancery Lane tube station it is difficult to think of any redeeming features for the building. The lift was antique, the offices were scruffy in the extreme, the furniture could easily have been retrieved from a skip and it was, and is, a total mystery how the canteen failed to give the entire staff food poisoning.

I was overseas for 5 years between 1968 and 1973. When I returned I found that London had been improved in two major respects: The Sound of Music had stopped running at long, long last and 5 - 11 High Holborn was a hole in the ground. The best thing that could ever have happened to it!

--- 3 ---

Program Overlays from Magnetic Tape c. 1966

The advent of the 1900 Range in the mid 1960s resulted in a dramatic increase in the amount of immediate access store/core store/main memory that was available to programmers. Up from a few hundred words on the 1301 or a few thousand characters on the 1500 to as much as 32,768 24 bit words on the initial 1900 models. How could one possibly want or need so much store? We very quickly found out of course! Deducting the size of Executive, deducting the minimum size of each program that you were hoping to multiprogram together and then sharing out what was left very quickly blew in a cold draught of reality.

At some sites reality was a little late arriving. One public utility ordered a 1904 with 32k and the usual range of peripherals to take over the customer billing application from a punched card installation. The systems analysts and designers set to with a will and defined a most elaborate system with a multitude of bells and whistles. It did not take too long before it became obvious that 32k was not going to be adequate for the main billing program, never mind multiprogramming it with other applications. Rather than scale down the scope of the program it was decided to overlay it. The only snag was that the site did not have any discs. In fact it pre-dated the availability of discs on the 1900. The program therefore had to be overlaid from magnetic tape. By this time the program was so large that overlays were needed to process even the basic transactions. This gave rise to near paralytic performance, excessive wear on the magnetic tape media and tape decks that objected to the continuous winding and rewinding. In the end the customer admitted defeat and bought a 32k word drum to hold the overlays followed shortly afterwards by a second drum.

It was at this point that ICT added its own contribution to the customer's problems. The program ran reasonably well after the addition of the first drum but it was still so large that some overlays had to be made from tape and this was sufficient to slow down run times too much to be acceptable. Following the installation of the second drum there was chaos. The program crashed every time it was tested but there was no obvious pattern to the crashes. It was realised fairly quickly that the crashes occurred whenever the program was obeying code that was held on the second drum. This was initially thought to be some sort of obscure hardware fault and the designers were called in from West Gorton, but to no avail. By chance, after much fruitless hardware and software diagnostic work, one of the programmers noticed that one of the crashes occurred when the system had been trying to execute data held on the first drum as program code. It was quickly established that all the crashes resulted from program overlays being made from the first drum when they should have been made from the second drum. Must be a hardware fault, more diagnostic work but still to no avail. Finally the drum software was put under the microscope. Initially that appeared to be correct until it was noticed that the drum address was masked at some stage so that only the first drum could ever be addressed. Eureka! Problem solved.

But why had the problem arisen in the first place? Due to limited hardware availability the drum software had never been tested with more than one drum. But then no one would ever write a program so big that it would need TWO 32 K WORD DRUMS - now would they!

--- 4 ---

Key to tape 'expert' c. 1967

From about the mid-sixties ICT computer customers started to show an interested in the new 'key to tape' method of data capture as an alternative to the traditional card punch and verify process and the paper tape equivalent. The new approach appeared to offer a quicker, less media expensive and generally more cost effective system for getting data into all those big, new 1900 computers that we had been selling.

ICT, with its BTM and Power-Samas background, had a big interest in the traditional card, and to a lesser extent paper tape, punch and verifier system and had tended to ignore the new technology (in the hope that it might go away?). In addition ICT made a substantial profit from its ‘own brand’ punched cards which it mandated, or tried to mandate, to all of its customers. To my public utility customers, the new technology appeared to offer a way of beating the input bottleneck as well as an opportunity to reduce costs; an irresistible combination.

At the time I was working in the Public Utilities sales unit. Amongst my customers was one of the area Gas Boards. They had taken one of the first ICL mark scanning document readers (UDR/UDT) to give them faster and cheaper data input and were looking to make other savings in the labour intensive area of data input. They approached ICT for a quotation but, at the time, we did not have anything to offer. I think that Mohawk or MDS were the market leaders at the time and perhaps even had the market to themselves. All ICT could offer at the time was FUD! I sat down and wrote an evaluation paper with cost comparisons between key to tape, punched cards and punched paper tape. Surprise, surprise, key to tape did not come out as ‘Best Buy’. The customer was so impressed with the document that he circulated it to every computer department within the UK gas industry. Before long I was receiving queries from Gas Boards around the country on the subject. Slightly panic stricken by this turn of events I sent a copy to ICT’s marketing department asking them whether I had presented the merits of the different systems relatively fairly. A week or so later I received a phone call from someone in the marketing department. The gist of the conversation was that I appeared to be the most knowledgeable person on the subject in the company and did I think that ICT should also move into the key to tape market.

(My opinion was YES, because if the company did not other people would move into our accounts and there would be nothing that we could do to stop them; sole supplier tended to be the rule rather than the exception in those days. Eventually ICT marketed a product from Potter but whether my paper had anything to do with the decision I never found out.)

A Classic Sales Opportunity late 1970s

Make the opposition fight on the ground of your own choosing has always been a good military maxim. Sales training courses that I have attended have suggested that it is equally applicable to sales situations. Indeed lecturers have gone so far as to suggest that it should always be the case. In the real world, I have found usually that this utopian situation rarely exists and that the sales pitch has, to a greater or lesser extent, been 'queered' by external forces before ones own arrival on the scene.

--- 5 ---

A Classic Sales Opportunity (continued)

I did however have one occasion where I had a classic sales opportunity and was able to define the rules and set the terms for a prospective sale. I had been dealing with a major UK petrochemical on a regular basis for a number of years. On one visit to the IT department I noticed that one of the analysts annotating a roll of telex printout. After I made some passing comment about old technology the analyst explained what the data was and the associated business problem. The data was from their aviation fuels subsidiary and related to individual refuellings of aircraft around the world. Getting accurate and timely data back to the UK so that the airlines could be invoiced promptly was a major headache. The values involved, even for duty free fuel, were high and a few days loss of interest from late billing was significant. The analyst was trying to see if a telex based system for submitting flight refuelling data would be an improvement on their current paper based system.

At the time ICL was marketing the 1500 System * that it had acquired as a result of buying the Singer Business Machines computer operation. The 1500 came with a large variety of software including some packages specifically for data capture. One of these, Complex Date Entry (CDE) was chosen and a demonstration program was written quickly. Some basic validation was performed on the input data and as an additional check the name of the airline was generated from the flight number by means of a small look up table. The demonstration was first shown first to the IT department, who gave it their blessing, and then to the users. The uses liked it and thought that generating the airlines name from the flight number was magic. Users were much more easily impressed in those days and far less cynical! Despite being a better than expected match to the business requirements, the rules said that it would have to be subject to competitive tender. Because the tender was drawn up very much with the 1500 in mind it was a one horse race and ICL got 'the nod'.

That was when the going got really hard. The users liked it so much that they wanted to extend the scope of the project to cover additional airfields and countries. Instead of installing the system in just two or three European countries the customer now wanted to put systems into about ten of them, as well as sites in the Middle East, the Far East and Africa. This should have been good news. WRONG! ICL did not even operate in some of these countries and in others it only had agents. Even an ICL presence did not necessarily mean that the 1500 was marketed locally particularly where there had not been a previous Singer presence, as in the Arabian/Persian Gulf States. It very quickly became obvious that the "International" part of ICL’s name was more of a wish rather than a reality. Each territory appeared to ‘do it’s own thing’ in it’s own way. In practice it meant we had to obtain quotations from every relevant ICL subsidiary and agent and try to get them to agree the same level and style of support that the customer was used to in the UK and expected elsewhere. It very quickly became a multi-lingual and multi-cultural nightmare. Was I glad when I got moved on to another project! I did feel slightly guilty about leaving behind the consequences of my classic sales opportunity exploitation.

(See page 6 for notes on the background to the 1500)

--- 6 ---

* Note

The ICL/Singer 1500 should not be confused with the rebadged RCA301 sold as the ICT 1500 computer system in the early 1960s. The ICL/Singer 1500 was the predecessor of the DRS20 Range. It originated (I think) with a company called Cougar who invented or were pioneers of MOS computer store in the early 1970s as a replacement for the earlier ferrite core store. Singer either bought Cougar or, more likely, just the 1500 and marketed it as a range starting with the basic entry level 1501. In many ways the 1500 was ahead of its time. Different processors could be linked together using a proprietary LAN (it predated Ethernet) and individual processors could act as print, file or communications servers for the network. The smallest model, the 1501, was probably one of the earliest desktop computers where the user could move the system around without having to involve an engineer. Although one person could move it, it would be difficult to describe it a portable; staggerable would probably be more accurate. It also came with a computer game, Grand Prix, which we thought was pretty good at the time. Such innocence!