Disintermediation

There are two main themes underlying the information revolution. The first is a continuation of the industrial revolution, most things got better/faster/cheaper/smaller unless they required skilled manual effort. And the class of things that require skilled manual effort is shrank.

The second main theme is “disintermediation” which is fancy term for cutting out the middle man. Increasingly information flows from source to sink electronically without the need for someone’s specialist knowledge to interpret or summarise it. People who need to make decisions get the information they need directly from its producers.

The next stage of disintermediation started about 40 years ago when new notation systems were created to describe the process of managing and using information. Waterfall methodologies like SSADM enabled system users and system designers to share a common language, sort of. Although it took special skills to develop diagrams like DFD (Data Flow Diagram) and ERD (Entity-Relationship Diagram) it took only a little training to understand what they showed.

The latest victims of disintermediation are the systems analysts and project managers who managed communications between coders who wrote software and the business users who specified what software needed to do. The gap between the two groups is gradually narrowing. Business users have PCs at home and are used to how software works and are getting better at describing how it should work. Programmers no longer wear white coats and live in air-conditioned basement computer-rooms. They talk to business users every day and often have a better idea of how the business really works than those who think they are running it. This new way of designing and building software has a name “agile methods” which is a clue to what they are intended to achieve. But that’s a story for another day.

Shy, Retiring

Last week I flipped my LinkedIn status to semi-retired, acknowledging a state that has actually existed for some years. Three quarters of the UK’s senior (age 50+) IT workers are unemployed. As the years pass and the gaps in employment get longer it gets harder to find a recruiter who will even return a phone call.

This isn’t a new situation but there is a sign that it can’t go on much longer. In the past the solution to finding (cheap) new technology skills has been to import them. In my lifetime this started with importing West Indian transport workers and more recently Polish plumbers. In IT we have been relying on India and Eastern Europe to provide new graduates.

The growing resistance to immigration, if it continues, will restrict UK companies’ freedom to import workers. Paradoxically Brexit will make the situation worse because we will need to import several hundred thousand IT workers to make it happen just when the brexiteers expect immigration to go down.

In theory we have age-discrimination legislation but it isn’t enforced. For instance it is still quite common to see the requirement for a degree in job specifications. As the probability of having a degree correlates strongly with age this is covered by the regulations on indirect discrimination and has been since 2008. Adding “or equivalent experience” to the specification makes it legal but we really should be taking active measures to expose employers who don’t implement that change.

T’interweb

Back when Methuselah was still in short trousers I had an idea. It happens now and then. I had been reading, in Byte magazine, about a system used in France called Minitel. Homes were given little terminals that were attached to the home phone line. They replaced the printed telephone directory but people were finding new things that could be bolted on to deliver other services too.

At the time I was working for the advertising agency Ogilvy & Mather. I was training people throughout the company how to use a computer, at that time this included telling people what a mouse was and how to use one. One of my trainees was Pierre, who was going to be handling the new IBM account in Europe. I mentioned to him that there was a new Minitel like system available worldwide. Not as advanced as Minitel but showing some promise. I had access to it from home but our work network wasn’t connected to it. I suggested that he look into the possibility of using it for advertising.

I wonder whether anything came of that?

Olympics 2024?

Watching the news today it struck me that some changes to the lineup of sports included in the Olympics might be called for. Anyone who knows me can probably tell that I have a limited interest in organised sports: “Built for comfort, not for speed.” But I can’t avoid what is effectively a shutdown of the TV networks for the next few weeks. One of the news items today was on the sports that are going to be added for the 2020 Olympics.

My opinion, that nobody ever asked for, is that unless the results of a sport can be quoted in SI units it isn’t a sport. I know I’m not alone in this. As far as I am concerned any competition that awards marks for “artistic merit” isn’t a sport, it’s an art-form. When the Olympics can offer medals for an objective test of “Best Cute Cat Picture” I’ll accept that it’s also time to award medals for gymnastics, synchronised swimming and ice skating. Not until.

The thought was triggered by the mention of the inclusion of rock-climbing as an Olympic sport from 2020. How can a climber’s performance be compared to another’s? Time to climb a standardised climbing wall would do. But for results to be comparable year-to-year the walls each year would have to conform to a published specification. Hand/footholds of the same geometry and coefficient of friction ever time.  Contestants might be divided into different classes based on their height and armspan.

That thought led on to thinking about what other sports could be legitimately included in future years. What other competitions are a test of skill that can be measured in numeric values. Digital values. Obviously, computer games! Easy to standardise from event to event, unambiguous scoring. Where do I sign up?

No More Maths

It is axiomatic that it is important for our school students to be mathematically literate; to be numerate. But that axiom needs some closer study. There’s an assumption there that needs examination, possibly for the first time since Pythagorean geometry was state of the art.

This isn’t just my idea. I was at a meeting of volunteers on the Computing At School project. I had a conversation with a professor of mathematics that crystallised my thoughts. Let’s drop mathematics from the school curriculum!

The axiom that privileges mathematics equates mathematics and numeracy, and it’s a false equivalence. Numeracy requires the ability to spot an arithmetic error in a supermarket bill. Numeracy does not require an understanding of trigonometry. Numeracy is important, mathematics is not in and of itself important.

The apparent importance of mathematics dates from the time when the mark of a gentleman was his ability to read Homer in the original language. Mathematics in schools doesn’t seem to have progressed much since the 18th century.

My idea is that at that time mathematics was the only tool for taking observations about the physical realm and taking them into the abstract realm. In the abstract realm concepts can be transformed in various ways, then mathematics handles the conversion of the transformed concepts back into the physical realm. That’s important but three centuries later we can do that same job better.

Let’s ban the News

There is an argument that the Internet’s harmful effects outweigh its substantial benefits. Back in the early 90s it was argued that media fragmentation would lead to balkanisation. It has.

Let me unpack that a bit. First, media fragmentation. When I was in short trousers there was one TV channel in the UK. In the playground you could tell what day of the week it was by listening to the chatter. Children of families that had TV all talked about what they saw the previous night. Later we got a second station and now there were two different conversations. We now have hundreds of channels. There aren’t hundreds of conversations though. Instead there are a handful because TV viewing clusters around a relatively small number of popular channels. Some people get their news from CBS and others get it from Fox.

Balkanisation is the division of a group or place into separate pieces that owe more allegiance to themselves than to the whole. Serbs and Croatians, Republicans and Democrats, Jews and Palestinians, Hatfields and McCoys. It’s not uncommon on a small scale, and by small-scale I mean Palestine or Northern Ireland.  In those two cases balkanisation is due to religious and/or racial differences leading to separate cultures. Now in the UK we have Remainers and Brexiteers both pointing at news sources that conclusively prove their mutually incompatible views.

With Republicans and Democrats in the US balkanisation has hit the big-time. There is some light at the end of a very long tunnel. Solving balkanisation in the US might throw some light on how to build a society that can cope with groups having radically different ideas of how things should work. And that won’t just help the Hatfields and the McCoys.

Brexit complications

The text of an email I have just sent to my MP:

let me first add my voice to those asking their MPs to do whatever is necessary to avoid The UK leaving the EU.

Next let me mention an issue that may affect the decision. Article 50 states that the EU may conclude negotiations and force exit two years after invocation. It seems to me that the practical implications of that have not been discussed.

To execute a withdrawal the government will need to review every department’s IT system as will every other organisation with a significant IT system, such as the NHS. The review could go ahead immediately if necessary, and probably should be started in weeks rather than months. A (very) rough guesstimate is that this will require an additional 100,000 skilled IT workers over and above those we already have and take five years. The likely source for those extra workers is India unless China enters that market. Some of this work will probably involve legacy software and that will require some existing UK IT staff to be brought out of retirement.

Negotiation of exit arrangements can to some extent run in parallel with the review. The next stage will only start in earnest when the first heads of agreement are available. This stage is modification of existing software to implement the new business processes that put negotiations into effect, and then testing it. If managed well this phase of the project will probably require another 200,000 immigrants. It will also require the active involvement of substantially every existing civil servant. It will also take five years, if managed competently. During this stage every new piece of legislation (foreign and domestic)  should be examined for its likely effect on the progress of the rewrite and if necessary deferred.

I should add that expecting this project to be managed competently may be a triumph of hope over experience. Even if it is managed well you should not expect any change from your hundred billion pound note.

Another fine….

We are finding out more about the consequences of brexit. Here’s another issue that I haven’t seen mentioned elsewhere.

Untangling the UK’s government systems from its involvement with the EU is non-trivial. Pretty much every government department and company will need to look at the final negotiated agreement and work out how its administrative systems will need to be change. If we assume that the EU is willing to enter into negotiations before Article 50 is invoked then that shouldn’t take more than five years or so. That is a bit faster than a normal EU trade negotiation but both sides already understand the system.

Knowing what the legal requirements are is the starting point for re-engineering IT and business processes. It is important to understand the scope of the work. Most government departments will have to analyse their systems so that they understand which of them will need changes. Most of this work should be easy because building modern IT systems generates the necessary documents. But some of the systems are not modern and so will have to be documented. That shouldn’t take more than a year or two and we can start right away.

Once we understand what we need to change and what changes we need we can get started on the main task. The first step will be to halt any other legislative changes that might interfere. Then we need to hire enough analysts and developers for the job. By the time this kicks off there may be enough of them in Europe but if not we will need to import them from either India or China. This will happen at the same time that industry also needs them so there will be competition for the best people. My rough guess is that we will need around half a million new immigrants, best case.

Worst case will be to hand the development work to the jokers that “built” the Universal Benefit system for DWP. Best case the work required will cost around £100 billion and take five years. Worst case it will be £500 billion and ten years. That is for the government half of the job. Industry’s bill will be about the same size.

I suspect that whoever wrote the two year time-limit into the Article 50 clause was well aware that it effectively makes leaving impossible.

It’s a fine mess you’ve gotten us into Stanley

Or rather David. The web is alight with messages about Brexit. I voted remain, and we lost. Rumour has it that lots of the people who voted for exit only wanted to register a protest and didn’t expect it to win. Whatever. We’re here and this is what we are stuck with. Time to quit the recriminations and plan to play the hand we have been dealt.