What VW emissions scandal teaches us about losing a race

So the car emissions cheating scandal gets Worse.  The thing is it seems that for a great many years there have been government regulations that simply have been beyond the reach of some, many or even all manufacturers.  In simplest terms VW (or Mitsubishi) could not make a Diesel engine that met standards at a profitable cost point.  Now if it turns out that all manufacturers have been cheating then the fault is ours – regulations were set beyond technology and we simply needed to take a different approach to lower emissions (more congestion charges, more public transport, different city planning guidelines – you know, hard expensive decisions)

But if some manufacturers could do it (and it seems the higher end engines from BMW did make it) then the issue is one of company structure and culture.

So let’s imagine we are a middle tier car manufacturer looking at our current capabilities. We find our competitors can meet standards but at a higher price point – but we don’t think we can do it.   We have several options

1. Abandon that market segment – it’s not profitable
2. Stay in market segment as a loss leader
3. Improve our engineering capabilities so we can do it

1. And 2. Are interesting – they are essentially strategic business decisions.  They have little or nothing to do with the technology, other than having an honest and open appreciation of the technology realities. (This is a very hard thing for an organisation to have)

3. Sounds simple but is by far the hardest – it is however linked to that “honest and open appreciation of technology realities).  A technology led organisation needs to be able to pass accurate balanced good and bad news up and down the chain.

To do so means psychological safety as a prime goal. 

But if you don’t have that, how do your employees pass that kind of news up to you?

You have to assume everyone is frightened to give you the bad news and work accordingly 
Otherwise you won’t know enough to know to decide between 1. and 2.

The 40 year hurt – and Europe

The 40-year hurt is driving America into Trumps waiting arms, but it is only now that I recognise what drive the hurt, the American dislocation and dispersal out of the American industrial heart and south east and west.

It is what has been playing out in Greece – the need to devalue a currency to even out trade imbalances or to deflate the human wages.

Perhaps it’s inevitable. Perhaps we need to try harm reduction, to build better communities explicitly instead of hoping they will just happen
But either way, Europes future looks a lot like Americas recent past – but no federal assistance programs to help

Government software sweet spot

My journey began some years back when I was fortunate enough to be paid to develop Open Source Software.  It’s kinda cool to realise that if your boss wants to know what you have been upto all they need to do is look at the GitHub timeline.  It’s also kinda cool to realise that if I was being paid by say the taxpayer (instead of via grants from Bill and Melinda or HP – thanks guys!), that the taxpayer (my boss) could also look at the same GitHub timeline and have the same insight into my work as my boss.  That’s going to be revolutionary for management practise Real Soon Now.
But what is not cool to realise is that the Texan University I was working for was 5,000 miles away and was more enlightened in software terms than my local borough council who had people who knew what open source was, but it could never get in the front door.

I banged my head against the idea.  I wanted to spend a career developing open source software to solve practical government problems.  

There is so many services government is obliged to do (link). All of which must become digital soon.

But how to get in?

Luckily the government digital service had the same headache from the other side of the door.

History of Agile and open source in GDS

The under-served middle ground

There seems to me to be a middle ground of open source government “apps”, everything from case management for child services to public house licenses.

All of these need to be developed once, extended many times and installed and maintained everywhere.

This is to me the new middle ground.  A pot of best practise software that is brought in and maintained by changing suppliers.

It’s only logical innit

The sooner we fall behind schedule, the more time we will have to catch up.

Scott Adams advice

I made a list of the skills in which I think every adult should gain a working knowledge. I wouldn’t expect you to become a master of any, but mastery isn’t necessary. Luck has a good chance of finding you if you become merely good in most of these areas.

 

In his unusual book “how to succeed by failing almost all the time” (ok that might not be the title, please google) Scott makes a claim that the Venn diagram of several skills at competent levels makes success more likely than a single circle Venn diagram at mastery levels – that it’s easier to be good at several complementary skills than a master of one.
He lists a set of skills, and I am creating my own bucket list here. These seem more as abilities to be deployed at marginal cost than skills per se.  I may be missing Adams’ point

• Public speaking
• Network “marketing”
• Self publishing
• Stats and reporting and measurement 

A similar list of abilities is necessary for my upcoming CTO manual book.

Living a coherent life

So listening to the founder of soylent, whom I happily dismissed as a crackpot in the first five minutes, I slowly realised that I was not only not living my desired future life (y’know, eat healthily, exercise, time with kids, think deep thoughts, live by beach, do something meaningful) but that my desired future life was … incoherent, and even if I did it, much of it is the wrong choice.

A lot of this is occasioned by the birth of our third child, Lottie.  One tends to re-evaluate ones life at these times.  But even so I should be more clear on my goals.  That’s true for us all, and all our goals.  It’s even a pretty good project management idea.

I need to do some figures on this but let’s say that organic farming is the right healthy choice.  We all see the TV programmes, yes?  Now let’s say that the 1860’s were the right level of organic farming – that’s how productive our arable land will be, when we are proper organic.

Can that farming sustain ten billion people by 2050? Probably not. Cannot sustain ten billion people who are vegetarian? Still, no.

What’s the shortfall? Because that is the moral choice we make as organic users – which 3 billion people need to starve to death to ensure we can all eat organic?
Or which methods of food production can feed 10 billion people – and do so healthily.
Oh and while we are at it, plants take most of their nutrients not from the soil but from the air blowing past them – so to have healthy farmed plants we need organic soil and organic I polluted air.
These are what I call 21C choices.

How to be humbled as software professional

Yesterday my wife gave birth to our third child – as a C-section. A team of medical professionals, surgeons, nurses and anaesthetists, all started and stopped their procedures with rather boring and mostly obvious checklists.  Are you the patient, what’s your date of birth, how many scalpels did you use.  And they all answered loud and clear. No, snide play along. No eye rolling. This was taken seriously – by people who knew it was at some level silly, but at another level, when you did not count the scalpels, bad things happened on that one bad day.

Checklists as best practise

I am a software “professional”. And we have checklists too. But some are just make weight – vanity metrics, things that can be measured  other ways or worse have no value anyone can spot.

The problem is bad checklists diminish the respect for the good ones.  When you have a checklist and you don’t follow it people die, the list gets some respect.  When the consequences are manager A has difficulty completing the weekly compounded you get … Less respect all round 

Why Cloud is different to virtual

So OpenStack (the open source alternative to Amazon AWS) has some 350+ different repos on GitHub. It’s a complex beast.

And those are mostly services that the developers realised they needed to build *after* the main Infrastructure-as-aService parts were built.

Some years ago Steve Yegge (a commentator and developer at Amazon and later Google) wrote an internal Google article (which was later released publically by mistake) about what makes Amazon and Googke different.  Google was better at almost everything (engineering culture to not having Jeff Bezos as a boss) but, what Amazon did well was platforms.

Google does not do them well

Bezos decided in mid 2000’s to make sure all of Amazon was running a Service Orientated Atchitecture

(The defintion is good and worth quoting)

And they learnt some hard lessons – lessons similar to those being learnt by the OpenStackndevelopers

In short – running a cloud takes a SOA approach and a huge amount of ancillary services – services that need to be automated and aware of each other but not brittle.
It is a good blueprint for how to build a programmable company. 
And it means cloud is a lot more than just hiring  ritual servers 

OMG – we are building railroads

So in the 90’s dozens of companies laid fibre across the globe to get in on the Internet wave.  It was obvious to all they were laying the modern railways – even if the railroad boom, whilst making some rich, really was a middling profitable industry over decades – that is around 10% profit amortised over 1850-1870.

What happened was that the enormous value was not capturable by the railways (competition mostly) and the value of goods moving easily was just spread around society.

Listening to how Square and Uner and so on are spending 10’s or hundreds of millions a month capturing their growing markets sounds to me an awful lot like laying application level railroads – and unless your moat is awesome, welcome to 10% amortised over time.

Varoufakis – In one minute, how would you reform the Eurozone?

Varoufakis blog
Greece is due to crash.  Greek people and businesses owe the state 95bn Euros.  The economy has tanked from 250bn GDP to 178GDP. And this month the Greek business year end falls due and the tax payments for 2016 are due.  No one can pay.

Greece’s problems are not being shouldered by the European State – it is still each nation’s bonds, each nations banks, not one set together.  Varoufakis argues this is the root of the problem 
His solutions

• Split member state debt into good and bad (meeting Maastricht criteria and not).  ECB funds this through Eurozone specific mutualised bonds where there will now be a single euro bond funding the “good” debt of countries, relieving each county of the debt that complies with Maastricht treaty borrowing limits and mutualising that debt.  It will reduce debt load and encourage counties to follow th rules 
• QE replaced with massive direct investment – startups green energy, financed by European investment bonds, sold by EIB and bought / supported by ECB in secondary markets.
• Step by step euro zone framework for failed banks
• Poverty relief fund