When the Portuguese offered to host the key GSM meeting they asked the GSM delegates where they preferred to meet – in Lisbon or Madeira. There was no doubt in Thomas Haug’s mind that Madeira should be the venue but some delegates were firmly against it. So he asked for a democratic show of hands The result was roughly 50-50. But Thomas absolutely wanted to go to Madeira, but it was too risky to start counting votes. His old friend Marius Jacobsen then intervened and said that he had been there several times in his younger days (he had been a telegraph operator onboard a ship) and found Funchal totally without merit. He therefore thought that, because of this background, he should be given two votes, to which Thomas replied that ”Maybe you should be given your two votes, but as Chairman I should be given ten, and Madeira it is”. So the decision was taken perfectly democratically – although with some heavily weighted voting!

Madeira is a lovely island and whilst the sun was not likely to shine in February the temperature would be ideal. I decided to make a special effort and take my wife Jocelyne with me. She’d had a tough time with all my overseas business trips. We took some money from our savings to pay for her trip. To loud protests from my sons they were deposited at my sister’s place and we headed for Heathrow. My expectation was for the days to be busy but there would be time in the evenings for a relaxed time in Funchal with Jocelyne. Things didn’t quite work out that way. The meeting unfolded like a bad dream.

On Monday Philippe Dupuis told his quadripartite agreement partners that late on Friday Alcatel had got wind of what was afoot. They had got into the French Government at a very high level (I later learned that this was the Prime Minister’s office). He’d lost his negotiating freedom – at least for the moment. We weren’t to worry though.

Armin Silberhorn told me after lunch that there was also a temporary hitch for him. The French had contacted his Ministry late on Friday. His boss had been out, so instead they had contacted his Permanent Secretary. He had not been briefed and did not know what was going on. He had put Armin on hold. Again, nothing to worry about. He was trying to unblock things.

Thomas Haug opened the GSM meeting, went through all the preliminaries and started to tread water.

Tuesday and still nothing. Jocelyne and I had a very pleasant dinner with the French delegation on Tuesday night. They had no news.

Wednesday everyone started to get irritable. By mid afternoon I decided enough was enough. We hadn’t even gone around the table to see where everyone’s views were. Where would we find the time to reconcile the different narrow band TDMA variants? The quadripartite partners had at least agreed to follow a process, even if the result may now be more problematic.

Thomas gave me the floor. The UK demanded, I said, to know where delegations stood on the main issue. It was like adding a 30000-volt charge to the room. Thomas Haug agreed to my request.

He began with what he thought was an easy question – just to build confidence. “Could everyone agree that the new system should be digital”? he asked. A real motherhood statement for that assembled gathering.

There was a look of disbelief from the delegations, probably including my own, when I asked for the floor. I reminded delegations on the reserve we had put originally on whether GSM should be digital. However I would be happy to lift this reserve providing the GSM could agree now that the decision on the choice of technology would be taken on a technical basis. By this I meant that the GSM should first accept the Executive Summary table from the experts.

There was a great relief all round that this was all I was asking for. Nobody dissented. I’d edged the crucial single page comparison table up a notch.

There followed the high drama of the roll call of delegations to state where they stood. The French and Germans were isolated but there were also one or two waverers including Switzerland.

It was then proposed that a small expert group should meet that evening to tie down the narrow band TDMA technology parameters. This was following the agreed script. Wim van Eck from the Dutch PTT was given the job of convening the group. The task was going to be tricky. There were some genuine uncertainties. To this could be added some national pride, personal prejudices of experts and a generous helping of industrial interests. The one trump card was the common foe in the form of the wide band TDMA technology industrial camp. But would that be enough?

The GSM meeting moved on to some lesser issues. By then I was only half listening to other delegates. The distraction was a picture emerging in my mind of just what GSM was setting out to do with the narrowband TDMA technology – design of a huge complex system by a committee! It was without precedent. On any sort of rational basis this was a mad undertaking. How could this ever succeed?

Out of this dark cloud forming in my mind a light emerged. Instead of having to decide a definitive list of characteristics I proposed the concept of “Working Assumptions”. The first rule was that these were to be the best choice of characteristics in the light of knowledge at the time. But they could be changed later in the light of new evidence. Thus nobody need feel that his or her case was irrevocably lost. However the second rule was that the onus of proof lay with those wishing to displace the working assumption. Further that any evidence to challenge a working assumption had to be gathered outside of the GSM. The third rule was that the evidence had to show that the proposed new assumption was better and not just comparable to the one it was intended to replace.

To understand the value of this procedural innovation one has to appreciate that CEPT worked by consensus. Everyone had to agree. My reasoning was that if a glimmer of hope was left that the minority party could always return later and retrieve the situation they may be more inclined to accept the majority decision. Further there were genuine uncertainties. Often in politically charged environments everyone finds themselves forced to watch mistakes pass through unchallenged lest it opens all the other delicately poised compromises to re-discussion and pulling down the whole agreement. The concept of “Working Assumptions” allowed mistakes to be retrieved later in an ordered fashion. Genuine mistakes could be retrieved without unravelling all that had been previously agreed.

I described this approach as setting the parameters of the system in slow drying cement.

Philippe Dupuis had been thinking very much along the same lines and the proposal was adopted without discussion. Alain Maloberti, who took the brunt of leading the detailed design of the GSM radio system, said many years later that without the “working assumption” approach the job of producing the detailed GSM specifications would not have been possible. It would have been too complex to handle the uncertainties, different parts of the system being defined in parallel and fragmentary national/industrial pressures.

As the GSM meeting broke up that afternoon the quadripartite agreement partners plus Sweden agreed to meet in the foyer of our hotel at 8pm to take stock. The French and Germans said that they would be contacting their Administrations and would let us know the latest situation. I sorted out a back room for the expert meeting with the hotel staff. Ted Beddoes from Vodafone and Robin Potter from British Telecom Research Laboratories would attend the expert meeting.

Thomas Haug turned up in the hotel foyer with a colleague around the same time as me. Bernard Mallinder and Renzo Failli then arrived. Eventually Armin Silberhorn and Philippe Dupuis arrived looking very glum. Their political instructions from home had hardened. They had to support the wide band TDMA solution. That or nothing!

Thomas Haug threw his hands in the air “That’s it then, no agreement.” The others looked depressed. The cohesion was started to disintegrate.

The incident in London with Bernard Ghillebaert had already had me doing some contingent thinking about this very situation. It was clear that any agreement was going to be delayed. The decision would be taken in another place. But what could GSM do in Madeira to prepare the best set of cards for the new players who would take the decision?

There was still all to play for such as a unanimous technical report highlighting the advantages of the narrow band TDMA technology, a 13 to 2 result in respect of national preferences and a well defined set of narrow band TDMA system parameters. I pitched my plan of action with all the optimism I could muster. Armin Silberhorn responded with immediate enthusiasm. Others nodded. The French officials neither agreed nor disagreed. The meeting broke up. I phoned Jocelyne and we went out for a quiet dinner.

We returned to the hotel at about eleven in the evening. I said that to Jocelyne that I would see her a bit later. I wanted to check how the experts were getting on. As I appeared in the back room Ted Beddoes grabbed me and took me to one side. The meeting had not been a good idea. Everyone was at each others throats and nothing had been agreed in three hours. “Let me have a go” I said confidently.

Within half an hour I had everyone’s united on at least one thing. I was the problem.

On one technical characteristic there were three choices. There was fairly equal support for two of them. The Swedes were the sole advocates of the third. Their game was to say that they were prepared to drop their preference if one of the other two choices being proposed by their ally Norway was chosen. It seemed perfectly logical to me that we could get the choice down to just the two. It was at this point everybody rounded on me and accused me of bringing politics into a purely technical discussion.

Undaunted I pressed onto the next characteristic. I fared better on that one. It was the number of telephone channels on each radio transmission. The German expert wanted 14 and the UK wanted 8. My experts told me the smaller the number the better for hand portables. It affected the size of capacitor needed to supply the energy surge for the TDMA burst of data. It was therefore important to one of the six criteria on our list for the digital technology. I managed to isolate the German expert and the world finished with GSM having 8 telephone channels per radio frequency channel.

At 2.30 am the meeting broke up. I went off to join Jocelyne and in the absence of room service the others raided the hotel kitchens. It did not prove to be a very fruitful expedition as the hotel had only left out tins of sardines. Mr van Eck got to bed at 4 am after writing his report. There were still too many loose ends. The narrow band TDMA coalition was in danger of breaking up into factions fighting each other.

At 9 am the GSM meeting recommenced. Something I’d learnt from the Foreign Office many years early when I was on the government delegation that set up Inmarsat was “The package deal”. I’d discussed this at breakfast with Ted Beddoes and Robin Potter and got their support. The French and German experts were effectively neutralised in the definition of the narrow band TDMA system since they had been instructed to only support the wideband TDMA solution. This created a vacuum that had to be filled by somebody. Why not us? Just before the coffee break I proposed “the package deal” to the GSM meeting. Something for everybody! It got general support and no opposition from France and Germany. It was game on.

There were two thing of particular importance to the UK. Ted, Robin and I decided to go for one and sacrifice the other. The one we went for concerned being able to double the capacity of the system sometime in its life (provision for half rate coders). This ensured the digital GSM solution could tick the box on one of the key UK evaluation criteria. What got sacrificed was BTRL’s full rate voice coder technology, which was a credible candidate but the German coder from PKI had wider expert support from within GSM.

There are good package deals and bad package deals. Critical to success of a good package deal is to share things out in a way that the best elements go in and not the worst. There also had to be something significant in the narrowband TDMA package for the French and Germans if Europe was to ever arrive at a final agreement.

My proposal was that the Germans would get the PKI voice coder design. It was a logical place to start a package deal. All the competing speed coders had been evaluated in a multi-laboratory testing programme under the supervision of a Speed Coder Expert Group. Whilst the overall analysis showed that no single coder could be declared the winner in all respects, the experts believed the coder from PKI and the coder from IBM France were superior. (Note: in fact the GSM full-rate coder was actually based on merging features from both these voice coders and my politically driven choice was going with the flow of what the experts had already decided). The next element was more controversial and that was that the French should have the frequency hopping feature that they attached importance to. The Nordics would get the modulation method (it had come out best in the Paris trials). There was even something for the Swiss who had to be kept on board. They (and others with mountainous terrain) were worried about long radio echoes in mountain valleys. A parameter (delay equalisation) value would be chosen to cover their interest.

When Robin Potter from BTRL got home he received a lot of stick from BT for giving up their voice coder proposal. But he could see what I was seeing – a fleeting window of opportunity for a narrowband TDMA agreement.

The principle of the package was quickly agreed by GSM but it took the rest of the morning to tie down the details.

After lunch the Portuguese had laid on a coach for some sight seeing around the island finishing up at a restaurant for a dinner. I remained behind with Mr van Eck to write up the package deal together with the characteristics agreed on the previous evening. It was slow and tedious as the good natured Portuguese secretaries struggled with a foreign language and our bad handwriting.

We finished at 7 pm. One of the secretaries had been asked to drive us to the restaurant. I’m not sure whether it was my natural courtesy or sense of survival which caused me to hold open the door of the front seat for Mr van Eck. It was only a small Fiat but was I grateful that it was Mr van Eck who had the best view of the hair pin bends coming up at 50 miles per hour. He is a big fellow and I’m sure his weight cut at least 15 mph off the car’s intended speed. She drove like a mad woman. We staggered into the restaurant. I found my wife. She was seated at a fully occupied table next to the Madeira director of tourism. To a round of applause I kissed her on the cheek and went over to a vacant seat next to Philippe Dupuis and his wife, still a little wobbly from the car journey.

The last day was ghastly. The French didn’t like the element the Germans were getting out of the package deal. The Germans and Italians didn’t like the bit the French were getting. The Swiss and Germans didn’t like the way I’d defined the parameter dealing with their mountain echo problem. The Swedes didn’t think the French should get anything the way they had behaved. Even my own delegation was up in arms about some detail on data transmission. On top of all this the French started to get cold feet in allowing narrowband TDMA to be quite so dominant in the emerging GSM report.

Thomas Haug gave me a back room and it became a sort of doctor’s surgery. I had a perfect rapport with Thomas who ploughed on with the GSM meeting dealing with some of the lesser issues.

Early visitors to the back room were Bernard Ghillebaert and Philippe Dupuis. Bernard had received new detailed instructions from the head of CNET (the French PTT research laboratory) – an old hand in CEPT technology battles. The narrow band and wide band TDMA systems had to be given absolutely equal treatment. If a set of working assumptions were being drawn up for the narrow band TDMA solutions then a list must also be drawn up for the wide band TDMA solution. Both had to be agreed by GSM and placed side by side as equals.

>This carefully balanced approach had been the CEPT custom and it always led to an impasse. My diagnosis was that the more unbalance in favour of the narrow band TDMA solution the better and flatly refused. Philippe Dupuis looked at Bernard Ghillebaert and I could see a resigned shrug passed between them. From this I deduced that Armin Silberhorn had probably told them that he would only give German support if they persuaded me first. They may also have held back as, in their hearts, they believed the narrowband TDMA was the right outcome.

The German position was confirmed to me later in the morning. Robin Potter had asked a very senior executive from British Telecom to find out from Armin’s chief Mr Haist what was going on in the German Administration. The message we got back from Germany via these informal channels (Note: which may not have been from H Haist himself) was that we should establish the evidence at the GSM that the narrow band TDMA technology was better.

It was a very politically complex situation. SEL was now French owned. This largely freed the German mobile operator to look at the technology choices much more on merit. There was less political pressure to wrap a German flag around a piece of technology. This gave Armin Silberhorn a little bit more room to manoeuvre in Madeira. But the political agreement tied him closely together with France. The French PTT experts had always been more open-minded and were leaning towards the narrowband TDMA solution. But Alcatel had effectively wrapped the French flag around the wideband TDMA. The French PTT were also responsible for industrial policy…hence the instruction to support the wideband TDMA. The French GSM team therefore had absolutely no room to manoeuvre but were conflicted between the instructions they had to follow and what they would have liked to have done if Alcatel had not made its high level political intervention. By default the UK had become the rallying point for the narrowband TDMA at Madeira and left reading the script that at one time Philippe and Armin had been a party in drafting together with Renzo Failli. Had the French and German delegates believed the wideband TDMA technology was the right solution for GSM I would not have had such a relatively easy run in putting the right politics in place behind the narrowband TDAM statements coming out of Madeira.

Armin Silberhorn acted in a very statesmanlike manner with one of the final proposals to the GSM Madeira meeting that the GSM record a request that the two dissenting countries should be asked by the middle of March to say if it were possible for them to join the majority. This was straight off the agreed script.


Figure 22 – The GSM Madeira Meeting in February 1987 ends in 13:2 impasse

It may have been a meeting of a standard group but at the same time it was the stuff of theatre, with Thomas Haug, the theatre director, struggling to control the flow of the play without sight of the script and a group of prima donna actors meeting from time to time off stage to sort out the latest crisis.


Figure 23 – GSM actors temporarily leave the stage

(Note: Mr van Eck, who wrote-up the narrowband TDMA package deal, is the man holding his hand to his head)

At one level the GSM Madeira meeting had failed – there was no agreement on the choice of technology. But at another level the GSM meeting had just adjourned and the theatre company moved to another place. What they parted with was a technically good narrow band TDMA system having been defined, a unanimous report stating in simple terms that on all counts it was better than the wide band TDMA system and a 13 to 2 majority in favour of the narrow band TDMA solution. On the face of it was not a bad outcome. But as Donald Cameron from Plessey said to me later that week “13 to 2 sounds impressive until you look at who the two are!”

The next phase of the standards battle would take everyone into completely uncharted waters.