Cambridge Computer Advert - November 1987

After the financial turmoil bought on by the QL fiasco - late delivery, dodgy early firmware and the curse of the Microdrive - together with other disasters like the Pocket TV and the C5, developed by spin-off company Sinclair Vehicles, Sinclair Research ran out of fiscal steam towards the end of 1985. Even a £12 million take-over by media-baron Robert Maxwell, which had been instigated by the Bank of England and and which was scheduled to have been completed in September of that year, had fallen through.

Nevertheless, creditors, which included Thorn-EMI, Timex, AB Electronics (which also built Acorn's BBC and Electron), Barclays Bank and Citibank^[1], were "generous" enough to let Sinclair see through the all-important Christmas sales period, which should have allowed the company to release the rumoured and potentially company-saving Spectrum 128, which was thought to be based upon technology co-developed for Timex US's unsuccessful TS2000 model way back in the summer of 1983, including all-important 128K bank switching, hi-res graphics and an MSX-like sound chip ^[2].

The company was also expected to release the fabled Pandora portable and the QL II^[3], however, it was all over by April 1986 and Amstrad - the brash New Kid on the Block, which had come from nowhere to conquer the market with its range of cheap, well-specced, all-in-ones - bought Sinclair's marketing and merchandising rights for £5 million (£17 million in 2023), leaving Sinclair with the debt-laden Sinclair Research^[4], which became known as SRL.

Whilst talking about the fabled machine, Personal Computer World's Guy Kewney asked Alan Sugar whether he'd bought the rights to the Pandora. Sugar apparently said "Have you seen it?". "Yes" said Kewney. "well then" said Sugar^[5].

The Pandora never saw the light of day, but the following year Clive Sinclair returned from the wilderness with a new company spun out of his Sinclair Research (SRL) holding company - Cambridge Computer - and a new machine, the portable Z88. Referring to the Pandora project in an interview in November's Your Computer, Sinclair said, when asked how much of the Pandora carried over in to the Z88 "Well, the basic philosophy is very similar, but when it came to the design we started from scratch with a clean sheet of paper. It was lucky we had the opportunity, because we were running up against a brick wall at that point".

Sinclair continued "The display technology on the Pandora was a flat screen similar in concept to the pocket television, with an optical system to blow it up to a decent size. It was very good, but it couldn't quite give us a display 80 characters wide. We went up to about 66 characters but couldn't do any better. We were just lucky when Supertwist [LCD] technology came along when it did"^[6]. Somewhat ironically, when interviewed back in 1985, Sinclair had discussed the "Spectrum Portable" and had at the time explicitly ruled out LCD technology in favour of Sinclair's own flat-screen TV^[7]. However, the biggest problem with the display was said to be that the angle you viewed the display was critial, which was not that helpful for a machine meant to be portable and which would be viewed in a variety of environments^[8].

There was much mystery surrounding Supertwist LCDs when the UK's ministry of Defence started writing to various manufacturers in order to obtain licencing agreements for the technology that it had developed back in 1982, when it was invented at the Royal Signals and Radar Establishment in Malvern, Worcestershire. Sinclair, who was using Supertwist in the Z88 was unaware of any direct approach by the MOD, as was Epson, which made the display for Cambridge Computer, although Epson's UK comms manager Edward Huggins was "aware of the MOD's concern". The MOD had secured patents in the UK, the US and Canada, and was expecting a European patent "soon" with a Japanese patent in 12 to 18 months. Meanwhile, nobody knew what, if any, effect complying with the patent would have on the Z88's price^[9].

Late again

Clive Sinclair, as illustrated for Your Computer with a Z88 reflected in his sunglasses, © Your Computer November 1987The Z88 had been released to some acclaim, although in true Sinclair style it was some three months later than the originally-promised April - the delay being blamed on writing its software, the contract for which was still unsigned as late as December 1986, with problems producing a coherent software suite being suggested. Sinclair said "It's still not complete, we've got the chunks, now they need to be merged into one"^[10]. Sinclair also pointed out that a second production line, in addition to Thorne Datatech's, was to be set up to help reduce delays. "The deal hasn't been signed, so I can't name names at the moment, but the second line will be there".

When the Z88 finally started rolling off the production line, Clive Sinclair said "We started shipping out them last Saturday [27th June 1987] and we should be up-to-date within two weeks"^[11]. Apart from being late, the machine was also £50 more than originally promised, a situation which was blamed on the strength of the Japanese Yen^[12].

Sinclair, who was quoted from five years before as saying "If we announce a product now, it is because it is ready for production", had also got into trouble with the Advertising Standards Authority as it upheld various consumer complaints that no expected delivery date had been mentioned in the Which Computer? adverts for the machine, putting the company in breach of the Code of Advertising Practice, which mandated that goods advertised by mail order must arrive within 28 days - a clause that Sir Clive's previous company Sinclair Research had been singularly adept at failing to comply with.

Adverts for the Z88, meanwhile, had only offered a "likely despatch date" whilst still requiring that potential customers send their cheques in with their order. Sir Clive was said to have been "astonished" at the criticism levelled at the company^[13], whilst in the following week Sir Clive was on the defensive, suggesting that ASA concerns were unfounded. With history perhaps influencing opinion, given delays over Microdrives, the QL and even the C5, it was perhaps not unexpected, but Sinclair was "adamant", stating that "the software is on schedule, production of the machine should start in April, and the first customers should receive their machines in the second half of April". Only the modem was delayed, as apparently that was stuck awaiting BABT approval^[14].

By July, Sir Clive was claiming that over half the orders for the Z88 had been fulfilled and that he was expecting to be "bang up to date" by the following week. He also expressed surprise over the furore about advertising, and took particular exception to an article in Popular Computing Weekly over delayed deliveries of the machine. "The ASA had only one complaint about the advertising of the Z88 - just one, which is hardly strong criticism".

He went on "They asked us to put the '28 days delivery' notice on it, which we hadn't done before because it wasn't to be delivered in 28 days. Really, the whole thing got out of proportion". The company was also sure that the Z88 would soon be sold in retail outlets, although marketing manager Peter King wouldn't name any of them. Sinclair reckoned that "a number of retailers have agreed to stock the machine" from the end of July^[15].

The official launch of the Z88 didn't occur until the first week of September, when the machine finally made it to branches of Comet and Dixons. The launch, which was attended by Lalla Ward, who had starred with Tom Baker as Romana in Doctor Who and seemed to be there simply because she had bought a Z88 by mail order and had said "I love it, it's the computer for me", was also the point when the company announced a big manufacturing deal with the UK subsidiary of SCI. The US company, which had a turnover of $470 million (£950 million in 2023) and which also built computers for DEC and ICL, was planning to produce the Z88 at its Irvine plant on the west coast of Scotland^[16].

By September 1987, Cambridge Computer was churning out 1,000 Z88's a week, a figure which was expected to quickly double. Thoughts were already turning to the machine's eventual successor, with perhaps a Z89 or Z90, in the style of the old Sinclair numbering scheme. However, Clive seemed convinced that the Z88 as it stood was right for the market, suggesting that "The thing about the Z88 is the way we've designed it. It can go on being expanded pretty well indefinitely, so obviously there's no need to change the machine because we just plug in different cartridges and expand it".

He continued "In terms for portables, that's our statement for a long time to come. I think it's not the sort of product that needs changing". He also took issue (again) with the ASA and his previous run-in over the lack of delivery times, saying "We never said it was going to be available in 28 days, because it wasn't. We just said if people were prepared to be one of the first, that's the way to be first and those people who wanted to be first were".

In the interview in September's Popular Computing Weekly, Sinclair discussed his wafer-scale solid-state disc project over at Anamartic, which was still waiting on the signing of a funding deal which would give the spin-off the £4 million it needed to launch the 20MB silicon disk into the market. Also mentioned was the rumoured Sinclair mobile "pocket" phone project, which was being joint-developed with Shaye Communications, Timex, shipping company Fred Olsen and a certain Nokia-Mobira of Finland, the world's largest manufacturer of cellular radio products^[17].

Capitulation

First revealed at the Which Computer? Show in February 1987^[18] and initially available only via mail order, Sinclair announced distribution deals with Dixons and Comet at the end of 1987. There were also moves to make the machine more compatible with other systems, unlike earlier 8-bit machines which had barely been compatible with themselves, let alone previous models or competitors. There were utilities to transfer WordStar and Lotus 1-2-3 files to/from the IBM PC, as well as the announcement of forthcoming utilities to communicate with Apple's Macintosh and the BBC.

The latter must have been especially galling, as Sinclair and Acorn had been waging a well-publicised war-of-words for years - from press slanging-matches^[19][20] to actual punch-ups in Cambridge pubs^[21]. This had been so ever since Chris Curry - Sinclair's right-hand-man in the late 1970s - had left to co-found Acorn with Hermann Hauser, and was immensely exacerbated when Sinclair lost out to Acorn for the lucrative BBC Micro contract. Sinclair's major beef about that appeared to have been less to do with Acorn and more about the BBC's attempt to set a new standard for software. Venting in a 1982 interview for Practical Computing, he suggested that the BBC "had the unmitigated gall to think that they could set a standard - the BBC language. It is just sheer arrogance on their part"^[22]. So the biggest suprise of the Z88, and perhaps Sir Clive's greatest capitulation, was that the machine ran BBC Basic.

The Dixons and Comet deals appeared to have soured by the end of 1989, as Personal Computer World was suggesting that the Z88 seemed to have disappeared from both retailers. The reason appeared to be that the volumes that had been negotiated as part of the original deals never materialised and, although sales were perfectly respectable from Cambridge's point of view, they didn't appear to be anywhere near the large volumes that were needed to make a profit on the machine for the High Street sellers. In January of 1989, Comet started selling the machine at below cost, a move which led Dixons to believe that its competitor had better terms, leading to Dixons breaking the fixed price agreement that the three companies had, with its announcement of a price drop in the summer of '89.

Cambridge Computers, however, was keen to keep the price up, as reducing it too much might mean that its important business buyers could consider it as too toy-like and stop buying it at all. So when Dixons gave Cambridge an ultimatum - reduce the wholesale price or else - Cambridge told Dixons to get stuffed. Whilst Dixons immediately dropped its price to £99 to shift remaining stock, Cambridge's new sales and marketing manager Roger Flack swiftly moved the company's whole distribution model to a more-conventional PC style, with a few distributors selling on to around 150 dealers.

There was also some movement in the long-sought-after educational market, as a City Technology College was promising to equip each new student, every year, with a Z88 - a move that if repeated by other colleges could lead to a big increase in units shifted - at least beyond the 35,000 which it was thought had been sold since the machine's debut at the 1987 Which? Computer Show^[23].

Back in the summer of '87, Cambridge Computer ran a campaign to attract software developers to its new machine. Its "Authorised Third-Party Development Programme" was intended to give "as much technical support as possible" to anyone interested in developing for the Z88, with an option of space on the Z88 stand in September's Personal Computer World show in return. CC's marketing manager Peter King said that "Comms will be big business on the machine, but we're also hoping people will publish software with vertical markets in mind - stock control, that sort of thing. There has also been interest from companies in writing text adventures for the Z88, and things like chess and backgammon could also prove popular"^[24].

The retail price of £287 (including VAT) was about £940 in 2023 terms, and it's pleasing to see that some of Sinclair's famous clichés are still present in the new company's advert with several mentions of "coming soon..." hardware and software, like the aforementioned BBC communication program and a BT-approved modem. Also present was a bit of post-launch recall panic as Cambridge Computer announced that some mail-order Z88's would need to be returned in order to have the machine's operating-system ROM upgraded in order to handle new software that was now being written for the machine. CC's marketing manager Peter King played down the issue by pointing out that the problem only affected "the first few hundred machines shipped"^[25].

Whilst Alan Sugar, the buyer of Sir Clive's previous Sinclair Research, turned up just for the first trade-only day at the 1987 Personal Computer World show in Olympia, Sinclair himself was there for the duration at Cambridge Computer's suitably matt-black stand, which harked right back to the days of the ZX81's marketing and style. Reporting on the event, Popular Computing Weekly's Gordon North wrote "[Sinclair's] little black portable Z88 was one of the few things at the show that I found myself returning to over and over again. His stand was small compared to his neighbours and there were times of the day when the light didn't get to it, but Sir Clive has still got pulling power. I visited that stand again and again until I really began to believe that I could formulate a spreadsheet on the train without being sniggered at"^[26].

Wafer Scale

One of the post-Amstrad Sinclair Research Holding's other new companies, besides Cambridge Computer, was Anamartic - a greek word meaning "fault free". This was set up to continue the wafer-scale integration technology that the previous Sinclair had started at its Metalab research centre at Milton Hall, a building which Anamartic continued to occupy (for a while, anyway) post-buyout. 14 staff had transferred directly over from Metalab, but Clive himself was taking something of a low profile at the new company as a non-executive director.

Clive and one of his silicon wafers, © Popular Computing Weekly, December 1985When the new Sinclair announced its "new world lead in semiconductor technology", it went on a fund-raising round to seek the extra cash it needed to take its final prototype to commercial production. Malcolm Wilkinson, Anamartic's general manager said "We have a prototype. To get that into high-volume production, more testing and so-on is needed, plus the expenses of marketing it. We are looking for about £6 million, from a mix of corporate investors, venture capitalists, people like that"^[27].

The company had initially been looking to raise £50 million back in the summer of 1985 when it was launching the wafer-scale project, but by the very end of 1985 it was claiming that it at least had the first £5 million in venture capital. The modest shortfall meant that initial manufacturing would need to be farmed out, with production not moving in-house for another three years, at which point it was thought another £40 million would be needed, whilst parent company Sinclair Research was thought to be in need of another £10 million just to keep it going - a sum that Sinclair thought might be made up of contributions from "organisations and wealthy individuals". Meanwhile, Robb Wilmot, who had helped set up the wafer project, had resigned as chairman of ICL but was now ruled out of any further involvement as he headed off to concentrate on his own businesses^[28].

The first product - the solid-state mass-storage device - had been mooted for the QL, and was something that was half-way between a hard disk and regular computer memory, but which had battery backup. Anamartic's first round of investment was funded by Barclays Bank, which put in an undisclosed "large sum". This funding round was topped up with a further £2.8 million at the end of 1987, with £1 million from US computer company Tandem, and some additional spare change from SGS Microelettronica, Baronsmead, Advent and Murray Technology. This seemed to be a vote of confidence in Sinclair's technology and contrasted with US IBM-compatible mainframe maker Amdahl, which had apparently spent $150 million on wafer-scale integration before abandoning the project^[29].

Catt amongst the pigeons

Sinclair's wafer-scale memory devices were developed from the 1978 designs of Ivor Catt, a man who had made some enemies over the years by suggesting that the classical Von Neumann architecture of computers - where a central CPU controlled all the memory, did all the calculations and managed input/output, was essentially "bunk". After several years maintaining his ideas at places like Middlesex Polytechnic and running digital electronics courses, he popped up again at Sinclair. Sinclair's product - essentially a solid-state Winchester disk - wasn't quite the culmination of Catt's dream, but at least it was using his technology, albeit using blocks of RAM instead of the original serial memory design that Catt had in mind.

The fundamental problem of producing a single silicon wafer was that imperfections would mean that certain parts of the chip wouldn't work, which would normally mean rejecting the entire wafer. This was solved for conventional chips, which were cut out in their tens from a single wafer of silicon, by simply discarding any chips that didn't work, however Catt's Content Addressable Memory solution allowed the wafer to determine for itself which were working sections of memory by injecting test patterns into each section and seeing which could be read again, meaning not only that in more cases the entire wafer was usable but that the entire thing was inherently fault tolerant^[30].

The first units from Anamartic, expected at the end of 1988, were to be CMOS, but Sinclair saw the development of bipolar technology as crucial, as this required about one twentieth the voltage to drive the bit-switching transistor. The whole idea was really 25 years ahead of its time, with the rise of solid-state disks based on various forms of Flash memory only becoming commercially viable from around 2005.

However, back in 1987 Sinclair had already said "I think solid state has to dominate. I have never believed that disks were the best way to go. There is no doubt in my mind that the whole computer world will change over - not overnight, but it will happen". Sinclair was also excited about a possible wafer full of Transputers - the British-designed parallel-processing RISC chip that had been developed at great expense under the auspices of the National Enterprise Board and which was already being mooted as a processor for a new Atari machine.

Sinclair continued "Years ago I was in the States and I realised the future of computing had to be parallel processing. I had one of those brainstorms we all have and thought 'Christ, we'll never do it - the pin-out problem alone will kill it!'. Wafer-scale integration [as a way to connect multiple parallel processors without the expensive faff of having to wire them all together on a motherboard] is fundamental. It's not an option, it's a necessity".

Martin Banks, frequent contributor to Personal Computer World, offered an analysis of Clive Sinclair: "The thing that strikes me about Sir Clive is that he can actually see far too clearly, far too far into the future. And he's dead right. Disk technology is getting horrifyingly anachronistic ... but it's going to be around for years and years because there are so many people using it. He's absolutely right, but it wouldn't surprise me if it takes until the turn of the century for him to be proved right"^[31]. Given the normal track record of technological future-gazing, being only five or so years out was remarkably prescient.

Anamartic started shipping its wafer-scale solid state memory at the beginning of 1990, in the form of its "Wafer Stack" high performance storage device. Offering some 100 times the speed of regular hard disks, the Wafer Stack came with one or more 40MB modules - up to four in total, or 160MB. As well as its existing methods to route around failed components on the wafer, the company had also developed its own programmable logic which enabled each of the hundreds of 1Mbit chips on a wafer to communicate with the next and so present as a single large storage area.

Anamartic's chief executive, Peter Cavil, stated "while traditional chip manufacturing techniques often result in only 25% usable chips on a wafer, our technology allows us to use about 80% of the bit capacity on a wafer. This means that the Wafer Stack storage module can provide significantly better price /performance than solid state disks based on traditional discrete chips"^[32].

That might have been so, but a 40MB system with SCSI interface still rocked in at $11,680, or about £20,300 in 2023.

Chips and MIPS

During the summer of 1989, news of another of Sinclair's chip developments leaked, at least slightly, in an edition of BBC's Horizon, which was about Sinclair himself. The new chip - the PCG-1 - was co-designed with Chris Shelton, the man who had designed one of the UK's first - and for a while most popular - micros, the Nascom 1. The new chip was already the subject of much gossip around Cambridge, as it was rumoured that it was going to be very fast, clocking in at around 150 million instructions per second, or MIPS. Or rather, not clocking in, as another rumoured feature was that it was to be a "free running" chip, in that it had no actual clock - unlike pretty much every other conventional processor - and would execute instructions as fast as it could, at least until it had to do stuff like communicate with the outside world.

Writing about a previous encounter with Shelton, Personal Computer World's Guy Kewney recalled how the Nascom designer had suggested that computer scientists all went about writing programs back to front. Instead of starting with a language and coercing it to do what they want, they should write out the program instructions and write an interpreter to run it, which was how Shelton had worked for may of his projects. The design of his new chip for Sinclair had this in mind, with the chip created to be able to do lots of new things and only when it could do them would they think about stuff like emulating boring old 8086s.

To help achieve this, the new chip came with 256 bytes of fast ROM into which customers could write their own instructions - with Sinclair then supplying a bespoke chip with the custom instructions built in, a bit like how modern FPGA chips are "programmed" to do particular things. Shelton reckoned that all this would be available within the next 12 months, but before that Sinclair also had an MS-DOS-based computer to launch as well as a "bicycle in a briefcase" project, which was apparently a real thing and not a joke, unlike rumours of an electrically-powered plane - an idea which is actually now being realised, some 30 years later^[33] - where the only drawback was said to be the 100-mile mains cable^[34].

When asked to elaborate upon the ultra-portable bicycle, Sinclair was said to have suggested that he was thinking of naming it "after Norman 'on-yer-bike' Tebbit" - one of Maggie Thatcher's cabinet ministers who had suggested that unemployed people get on their bikes to look for work, and who was also seriously injured in the Grand Hotel IRA bomb, where Acorn's Chris Curry was also in attendance.

In the same interview with Guy Kewney of Personal Computer World, Sinclair was quizzed about plans for a bigger Z88, which had just gone on sale in the US via merchandising company Diversified Foods. Instead of answering, Sinclair replied "I've often wanted to do a bicycle, one which you wouldn't know you have". However, he eventually conceded that there was something going on, saying "what we're working on is not a replacement for the Z88, and it won't be out this year". He continued "Who's telling you this anyway? We've only spoken to two people about it. You're way, way too early"^[35].

Future imperfect

Whilst Sinclair often had a spookily-accurate grasp of the future, he didn't always get it right, for instance when he claimed in an interview for Practical Computing's July 1982 edition that the ill-fated Sinclair Microdrive was "miles ahead of what anyone else is doing", shortly before the 3.5" floppy cornered the market. His dream that expert systems - now generally re-branded as Artificial Intelligence, or at least Machine Learning - would replace doctors and schools because computers "will teach better than human beings because they can be so patient and so individually attuned" is also still some way off, although it's certainly not ruled out.

However he did offer some interesting prognostication on what might be the impact of the Internet and social media, some 20 years before Facebook launched in the UK. Discussing the threat to normal personal communications and the "de-socialising influence" of expert systems, AI and networks, as highlighted in a report published in New Scientist earlier in 1982, which suggested that networking buffs "became withdrawn from their everyday lives and preferred to communicate with their on-screen pals", Sinclair said, in what could easily read as a prediction of the death of the High Street at the hands of Amazon and the like, "Yes, I am concerned with this. We have to watch very carefully that you do not remove the rituals of things like shopping or banking. Sometimes it is possible for something to disappear before people realise what it is they want to keep".

The same interview discussed how advances in computers could lead to experiences that were better than normal life, like those available on ultra-expensive military flight simulators, or in other words Virtual Reality. Sinclair clearly didn't forsee a world where VR could be achieved by lashing a smartphone to the front of one's face by using an absurd goggle-box covered in polyester fur, but instead suggested "a three-tube projection TV with a 50" full-colour CRT display [or where] the optics and electronics could be fitted into a shoe-box-sized unit projecting onto the wall-mounted screen"^[36].

See also Sinclair.