Snippings

Speaking at the cryptocurrency conference running at the World Economic Forum (WEF) shindig in Davos, I notice that Mable Jiang from Multicoin Capital in Beijing says that “It’s a kind of Cold War, but not just for crypto… Currency is the leverage”. That’s an interesting turn of phrase. Oh, which reminds me… I have a new book that will be coming out in June. It’s provisionally called “Currency Cold Wars—Cash and Cryptography, Hash Rates and Hegemony” and you can read the preface online at my web site.

xxx

The U.S. dollar is a triton among minnows. It was on one side of 88% of all foreign-exchange trades in April this year, dwarfing the renminbi’s 4%. According to the Bank of International Settlements, non-U.S. banks have $12.2 trillion in U.S. dollar assets, with China being about one trillion. These assets must be funded with U.S. dollar liabilities that frequently need to be rolled over. Controlling access to dollar funding is a huge weapon for the U.S. Although the greenback is unlikely to join the fray anytime soon, a further ratcheting up of tensions seems probable at some point.

From Despite Trade Deal, It’s Winter for the U.S. and China:

xxx

A nostalgic excursion into the history of the British computer industry triggered by some comments on Twitter, made up from a book review I wrote in 2012 and a couple of blog posts (here and here) that followed. 

It is 1975, and at the Park Senior High School (as bog standard a comprehensive as they came) a group of curious schoolboys led by a farsighted maths teacher (Pete Bayley, who subsequently went on to become the Director of Qualifications at the British Computer Society) take the bus into Swindon town centre and enter the headquarters of the Nationwide Building Society. There, we potter down to the computer room where we are allowed access to their mainframe (a UNIVAC 1109 with drum storage, if my memory serves) because they didn’t use it in the evening. I will never forget that kindness from Nationwide and they still have place in my heart for it today. It was there we worked on our first serious programming project, a system to schedule appointments for the parent’s evenings!

Can you imagine ringing up, say, Barclays Bank today and asking them if some school kids could use their mainframe in the evening if they’re not too busy? Astonishing. As Pete wrote many years later:

This same keen group were the ones that went with me to the offices of the Nationwide Building Society for several years to run, on the UNIVAC machine at their new offices in Swindon, a COBOL program (sorry app) to match staff and parents up for parents evening meetings. I doubt that I could get through today’s security there to ‘play with their line printers’ as we did then.

Having a whole computer to yourself was then a novelty. When Brian Dyer, the then-deputy headmaster of Park asked me and a few friends if we’d be interested in sitting a Certificate of Secondary Education (CSE) in Computer Studies (the school had no lessons in the subject, so we had to just read up in our spare time) my first contact with a general-purpose stored-value computer system came via punched cards sent to the University of Bristol once each week. You punched your cards, they were sent off and a week later you got back your print out. I think it was an ICL 1902A, the first IC-based range they produced (the A meant it had a floating point unit for scientific calculations), and we wrote in FORTRAN. The school got a teletype, and we had access to a GEISCO time-sharing system using Dartmouth BASIC. I think we were allowed an hour per week, or something like that. I got Grade 1, and was set for life…

Other weeks we took the bus up to the now-demolished Swindon College, where we used their Elliott 803B with a then-amazing 8Kb of core memory (you tell the kids of today that…) to write Algol programmes on 5-track paper tape. You loaded the compiler, the loaded your program and the machine produced a machine code tape (for their strange 39-bit word instruction set) and then loaded your machine code and executed. Here’s a video of someone using one of these beasts! I don’t remember much about what software we wrote, although I do remember spending an inordinate amount of time working on my football simulation that used random numbers to work out where the ball went after each kick.

Park School 1976: Me, Clive Debenham (Rest in Peace), Simon Turpin and Bob Kirby (Rest in Peace).

In the hot summer of 1976, we were also allowed, and I have no memory of how this came about, to cycle out to the Royal Military College of Science in Shrivenham and use their more advanced computer system — although I can’t remember what this was — to write in Algol 68-R. In those pre-Al Qaeda days, we were allowed to amble around the campus and wander in and out of the computer room essentially unfettered. This all stood me in good stead. When I got my first vacation job at college it was for the Southern Water Authority in Eastleigh. One day my boss asked me if I could help him with a problem. The IT department (in Brighton) were sending him the wrong statistics. Did I know anything about computers? We opened the cupboard at the end of the corridor and found a teletype connected to a 1900-series (a 1906?) running Fortran under GEORGE III, which fortunately I knew how to use. I was instantly appointed departmental IT supremo, and never looked back!

But wait. Let’s go back to Nationwide for a moment. Why was one of our largest financial institutions using this American system? After the invention of the general-purpose computer during the Second World War, Britain led the world in the technology. I’ve already mentioned the ground-breaking Elliott from 1963. Yet by the time I was in high school, America dominated. What happened?

The universal automatic computer (Univac) series in use by Nationwide had been developed by US entrepreneurial pioneers the Ecket-Mauchly Computer Corporation, which ended up part of Sperry Rand. Why weren’t they using an EMI computer? If you want to know why, you should pick up Tony Gandy’s “The Early Computer Industry”, which is a terrific study of the post-war, nascent computer industry’s evolution. And if you’re wondering what happened to Britain’s EMI and its early lead in the computer industry, I refer you to my favourite quote in Tony’s book:

Although this was the fastest printer in the UK, its reliance on bicycle chains was a source of some concern to EMI engineers.

OK, so maybe not EMI then. But why not English Electric (EE)? Or Elliott? Or ICT? Or why not Ferranti? The world’s first stored program general-purpose computer was the Mark I developed at Manchester University in 1948. Manchester was the home town of the electronics giant, who went on to win the word’s first export order for a mainframe computer when they sold a version of the Mark I to the University of Toronto.

Britain led the world in the graphene of the post-war world. Indeed, in 1963 it had a trade surplus of half a million quid in computer technology. Two decades later, it had a trade deficit one hundred times greater. How did this happen? Why did the computer industries in the US and the UK develop so differently?

The truth is that by 1963 it was already too late. The structural differences were already manifest. The US was developing computer companies, whereas in the UK computers were sidelines for electronics companies. IBM had already defeated the electronics companies who had pioneered the technology. When it came down to it, what IBM had that they did not was experience of commercial sales of equipment. By the end of the evolution of “first generation” computer technology, everyone else was already a rounding error compared to IBM. As Tony makes clear, not everything IBM did was a success. Big Blue didn’t have a crystal ball, but it did have a relationship with commercial customers through a sales force that was able to sell them computers.

For a truly English take on all this, by the way, look no further than Gillian Ferry’s “A Computer Called LEO: Lyons Tea Shops and the world s first office computer” which tells the story how the Lyons Electronic Office (LEO) computer was developed by a restaurant chain and went live in January 1954 processing their payroll. It was then subsequently developed by English Electric after it was acquired in 1964. In 1965 they called in McKinsey, who recommended a reorganisation, and went on to launch the System/4. They sold/10 of them and English Electric vanished three years later after being merged into ICL in 1968.

When I originally reviewed Tony’s book for Financial World magazine back in 2013, I said that the book’s subtitle should have been “I’m from the government and I’m here to help” rather than “Limitations of scale and scope”, because the story it tells is of how British state support of the nascent industry doomed it. Tony concludes that the US government’s strategy of purchasing large numbers of computers seems to have been a much more effective route to support than the British government’s policy of directly supporting development in the wrong firms the wrong time. We had a Ministry of Technology (led from 1966 to 1970 by the noted technologist Anthony Wedgwood-Benn) that directed the formation of our national champion ICL, a champion that inherited a legacy which meant it could never develop a successful strategy despite having one of the ranges that IBM actually saw as a serious competitor, which was that old 1900 series machine that I used in my very first paid programming job.

*wipes away tear*

If you want to post your memories of the Elliott below, I’d love to see them!

xxx

The government’s attempts at digital identity have failed – it is time for a new, modern approach

From Implementing a 21st century approach to digital identity:

xxx

xxx

The accuracy advantages of UWB are clear: UWB can measure distance and location to an accuracy of 5 to 10 cm, while Wi-Fi, Bluetooth, and other narrowband radio systems can only reach an accuracy of several meters. What are the other tradeoffs between the different systems?

Modern-day UWB systems provide a data rate of 6 to 8 MB/s, which falls between Wi-Fi and Bluetooth. Previous-generation UWB delivered much higher data rates, up to 100 MB/s, but power requirements and distance range made it impractical for use in mobile devices. Though today’s UWB may not be sufficient for streaming HD video, it’s enough for sensor data, security camera streams, instructions, and the like.

UWB consumes significantly less power than Wi-Fi, although Bluetooth 4.0 also uses extremely low power. That, of course, is a key factor for mobile devices and other small electronics in terms of battery life and practical usability.

Obviously, one tradeoff against UWB is that Wi-Fi- and Bluetooth-enabled devices are better able to interact with today’s smartphones and tablets. Nonetheless, some companies are implementing UWB, building devices that use both UWB and either Wi-Fi or Bluetooth to get the best of both worlds.

What’s Next

While smartphones and tablets garner most of the attention these days, electronics design is moving more and more to the next generation of smaller devices, which comprises wearable devices and the Internet of Things. Devices are steadily shrinking, from appliance-sized to pocket- or wristwatch-sized, and correspondingly, power and size requirements are getting increasingly important. At the same time, data-rate requirements are often less significant—smaller devices tend to send and receive sensor data and instructions, not the HD video that typically goes to bigger screens.

Increasingly, designers of wearable and Internet of Things devices also want to be able to precisely track location. Some devices are designed to be found by other devices; some move around (like robots) and need to know where they are so as not to bump into walls; some carry out various actions based on their location (such as turning on a radio when entering a room); some are designed to know how far they are from other devices; and some want to track their locations for other new and innovative reasons. Whatever the reason, UWB technology gives devices the ability to track their locations more precisely than Bluetooth or Wi-Fi, with lower power and size requirements

From What’s The Difference Between Measuring Location By UWB, Wi-Fi, and Bluetooth? | Electronic Design:

xxx

xxx

According to Iranian English news platform Financial Tribune, the country is set to offer tax holidays to bitcoin mining firms. Iran’s National Tax Administration (INTA) says that bitcoin miners who repatriate their foreign earnings will be eligible for tax exemptions.

From Bitcoin Miners to Enjoy Tax Holiday in Iran – BTC Investor:

xxx

xxx

Lawmakers in New York are trying to correct this with a new bill that would create a “public Venmo” system designed to include more people in the formal economy and stimulate local economic growth. In November, New York State Assemblymember Ron Kim, Senator Julia Salazar, and Cornell law professor Robert Hockett announced their Inclusive Value Ledger (IVL) proposal. If passed, it would create the country’s first publicly owned electronic banking platform, as well as a digital currency that can be exchanged for goods and services within the state.

From New York Is Proposing the Creation of a ‘Public Venmo’ – VICE:

xxx

xxx

In the end, a shift to a basket of world currencies might be the healthiest outcome. If the euro, yen and yuan were to join the dollar as international reserve currencies, it would mean a safer and more stable global economy. And it would enable the world’s major economies to compete on a more-or-less level playing field.

From World Is Stuck With the Dollar as the Reserve Currency – Bloomberg:

xxx

xxx

A previous release earlier in October 2019 demonstrated that a public blockchain could be secured by running a secure “level 2” protocol on top of the Ethereum blockchain providing confidentiality to transactions while also allowing full auditing by a designated third party.

This confidentiality is possible through a cryptographic technique known as “Zero Knowledge Proofs”. It’s not without its limitations, however. One drawback is that the mathematics involved in the cryptography are complex and therefore computational expensive. In the world of public networks, where participants pay to carry out transactions, that means cost; $8-$10 per transaction, in fact. That’s an acceptable amount if the transaction involves something of large notional value, such as a house, but not if you are paying for a pint of milk.

From Ernst & Young Doubles Down On Its Bet With Ethereum:

xxx

Alan McIntyre of Accenture

You are right that the vast majority of their income comes from the interchange and payment transaction fees, and those are declining. More and more of these types of transactions will move to account-to-account payments rather than go over the Visa and Mastercard rails. That will make the economics tighter and tighter. So I think over time the water drains out of that revenue pool.

From Why Fintech Challengers May Not Conquer Banking After All:

xxx

xxx

I think the [challengers] at the moment are going to struggle to be profitable. Ultimately these guys will either need huge scale at relatively low margins or they’ll have to build more of a traditional balance sheet model where they build the assets and liabilities and make the spread off of them.

From Why Fintech Challengers May Not Conquer Banking After All:

 

xxx

I’ve long taken the view that “challengers” wasn’t really the right name for the new financial institutions that use new technology but essentially the same business models to fragment the customer base, which is why I’ve been using the term “niche banks” rather than “challenger banks”.