Code is law. The case of law 36.

Larry Lessig’s Code and other Laws of Cyberspace of remains one of the most important books written about how software impacts law.  Lessing and other academics have shown how code has fundamentally changed copyright law.  The design of technology, especially software, changes the nature of regulation and enforcement of laws.  My academic work is in part about applying the insights of Lessig and others to enterprise software.

I have a habit of seeking out weird metaphors to explain things, sometimes in a long, convoluted fashion.

Readers of this blog will know that I am a fan of cricket.  Today, I have SA v India and Australia v England on in the background.

Law 36 is the leg before wicket law (LBW) in cricket.  Here is the law, explained brilliantly by the brilliant Stephen Fry.

Over the last few years technology has come to play an ever more significant role in determining the LBW decision. I’m generally a fan of DRS, the Decision Review System.  There are a variety of technologies in use today. Slow motion review, infra red cameras, edge detection and ball tracking. In many cases, it has made decision making fairer, and it adds to the fan experience. But DRS is not without controversy.

drs-aus-v-wi

Ball tracking example. from DRS website.

hot-spot

Infra red (hot spot). (image via global cricket community)

Related image

Snickometer. (image via skysports)

In the Guardian today, a leading English cricketer, Jonny Bairstow “spoke of inconsistencies in cricket’s decision review system “messing with careers and livelihoods”

The article noted.

  1. It is not always clear with edge detection technology as to whether sound is from the ball hitting the bat, or from the pad or clothing noise. (for more on the acoustic theory, and how edge detection could be improved, see this post)  and this paper.
  2. There are different technologies in place around the world, and these work differently.
  3. the parameters of the same software are set up differently in different countries.

 Back to the metaphor, and some observations.

Code has partly replaced the umpire as the arbiter of the LBW decision. The programmer who codes the algorithm is now umpire. Code is Law 36.

Code creates an aura of infallibility that is not always earned. The algorithms are model representations of the real world, not the real world. It is a black box and unless you have an advanced degree in physics, impossible to understand or challenge. The fact that the technology is improving, by definition means that is still not infallible.

Consistent procedures around the code’s deployment are important in determining perceptions of procedural fairness. Standards, consistently applied,  are key.

People are willing to abdicate to code, provided that they perceive that the code and supporting procedures act consistently.

Making judgements with the input of technology takes practice, especially where the decision is not clear cut. For instance, interpreting sound wave curves requires training.

The technology can change faster than the law can adapt to it.

The use of the technology has changed the game. How batsmen respond to spin bowling has changed. The former England captain, Andrew Cook had this to say. 

“It has changed the way you play spin over the last couple of years,” Cook said. “You have to keep your pad out of the way and it certainly doesn’t matter how big your stride is because gone are the days when you would say, ‘Well, he’s a long way down’.”

More broadly than cricket.

As we build code to replace or augment human decision making, we need to hold those building the code and the processes around it to high standards. Jonny Bairstow has a point, when it is wrong,  you mess with people’s livelihoods. Law 36 is a fine example of the rather messy place where law and code meet.

 

 

 

 

 

 

Advertisements

Weak ties and breakfast

I was in Hamburg this week, with my family.  My Dad and I were able to go to the the New year’s eve daytime concert at the Elbphilharmonie. That experience will be the subject of another blog post, I’m still processing the majestic marvel that is the Elbphilharmonie. And I’m really impressed by Hamburg.

Just before heading up to Hamburg, I pinged Paul Jozefak to see if he could meet up for a coffee.  We met up for breakfast, and I’m very glad I did.  He is wise beyond his years, and generous with his advice and ideas.  We both enjoy cycling and as relatively eingedeutsched  ex-pat/immigrants we had similar views on living and working in Germany. Both of us have been here for the best part of 20 years, without really planning to be. 

Paul’s insight into the state of digitalization and the opportunities that it opens up in German economy is profound, and I have not met many people who understand the big company world, venture capital, and start ups as well as he does.  He has an excellent, long running, blog.  

It was Jeff Nolan who introduced us many years ago, and we stayed in touch via social media. I think this was the first time we had managed to do face to face since then.

It reminded me of the Dunbar number and Granovetter’s research on weak ties that I first learnt about from reading JP Rangaswami and Andrew McAfee, if my memory still works correctly.  

While social media can be time sink, at times it offers up a connection and relationship that makes me realise that it has real utility.

That reminds me, it is high time that I meet JP in person too. But rather than breakfast, I hope it is at tea, at the Oval, or Lords, perhaps.

 

Sandalwood and SaaS.

It was a lovely day in Geneva.  The lake and sky were competing for which had the nicest shade of blue.  The alps looked close enough to touch.  It even smelt like spring.  I’d like to say it was the blossoms on the trees, but it wasn’t.  Colleagues from SAP SuccessFactors and I were visiting L’Occitiane’s office. I’m not sure what precisely they were distilling that morning, but I do remember the lovely scents of sandalwood, lavender, and jasmine.  I’m sure there was a lot more in the air, but that is where my olfactory sophistication ends.

That day, we met Xavier, who was pushing hard for a major HR transformation at L’Occitiane.  We talked a lot about the business, and how it was this mix of retail, marketing, design, fashion and R&D. We heard from him about the history of the business and how it has grown rapidly, but that it still has its roots in Provence. He was passionate about building better processes for the employees across the organization.

L’Occitiane were kind enough to give me a bunch of product samples that afternoon. These were quickly annexed by my girls when I returned home.

It made me smile today when I saw their SuccessFactors go live press release.  Congratulations to Xavier and the team. The scent of sandalwood somehow made its way back through my memory. It also reminds me that I still have Christmas shopping to do. Their after shave balm is rather nice too.

Asking the awkward question about employee engagement surveys.

For all of my working career, I’ve argued that HR needs to get more analytical, so I’ve welcomed HR’s growing interest in data based analysis and benchmarking. I have been something of a flag-waver for data driven HR for many years.  I’m thrilled to see the increasing sophistication in workforce analytics. The posse of machine learning, predictive, big data, and artificial intelligence is going to transform HR in ways we can’t imagine today.

But for analytics to work, to add business value, and to be fair and just, it needs to be based on robust methodologies. The analytics are only as good as the frameworks and models they are built with.  No matter how compelling the charts and graphics, if the models are weak, we risk doing the business and its employees a massive disservice.

Employee engagement is a hot topic in HR circles. Start ups, HR tech vendors and consultancies of all sizes and shapes are building solutions and practices focusing on improving and measuring employee engagement.  Gallup’s survey though, is the granddaddy of them all.  It is used for as the justification for all sorts of HR interventions.

Armin Trost, a German HR professor, asks some awkward questions of the Gallup employee engagement. Have a read here.  Here is a quote:

So, why is this bothering me? I am a social scientist and expect clarity as all good social scientists do. Published scientific results must be replicable so that others could repeat studies in the same manner. Empirical methods and outcomes not being communicated in a clear way are useless in the eyes of the scientific community.

When algorithms are hidden behind the firewall of intellectual property, it makes it hard to really know how robust they are. If we are to base major decisions on poll data, it is beholden on us to make sure that we know how the results are derived.

After the failure of polls to effectively predict politics, it is time for a lot more scrutiny and indeed scepticism. Not just for this survey, but for the whole industry.

I am always nervous when mentioning physics and psychology in the same blog post, but the words of Richard Feynman popped up in my twitter feed as I was writing this, so I couldn’t help myself.

 

 

Cycling, Crisp Research, Dimension Data, and a touch of Neuromorphic computing.

Stefan Reid of Crisp Research invited me up to Frankfurt today, to attend their conference.  I’ve known Stefan for a while,  while he was at Forrester and I was at Gartner, we frequented the same taxi queues and analyst dinners at vendor conferences.

I learnt a good deal about the practical state of ML, IoT, cloud etc in Germany.  Strong presentations from the analysts,  customer panels, and case studies. Osram’s massive transformation into an IoT platform player. Continental’s data lake and mobility services strategy, including live demo. neat.  Continental is a lot more than tyres.

I also learnt a new word during Carlo Velten’s keynote, Neuromorphic computing. Apparently  lots going on at Heidelberg University on this, funding in part by the Klaus Tshira Stiftung.

Beate Spiegel, Managing Director of the Klaus Tschira Foundation
“Klaus Tschira was very interested in the investigation and development of new computer architectures that are modelled on the human brain. Beyond his personal interest, he was keen to support the ongoing development of information science for the benefit of humankind. That is why he agreed as early as three years ago to become a sponsor of the European Institute for Neuromorphic Computing through his foundation. We are very happy that with construction of the new facility now under way, the University is taking the first visible step toward new and exciting research findings.”

I glanced at the agenda yesterday, and I was thrilled to see Rob Webster, who runs the sports practice at Dimension Data, on the agenda.

I’ve been very impressed how the South African/global tech company, Dimension Data, has developed its brand recognition through its sponsorship with  Tour de France / ASO, and its pro-cycling team.

The philanthropic dimension of their engagement is particularly compelling, enabling kids in Africa to receive bicycles of their own. Check out Qhubeka. While for some of us cycling is the new golf,  and we argue about SRAM v Shimano, at a more existential  level, owning a bicycle might be the difference between getting to school or not.

Often the link between sport sponsorship and the core business is a tenuous one, but in the case of Dimension Data, there is a technology play with both the TdF and the DD team.  Anyone who rides a bit will know that the last decade has seen an explosion in measurement and data in cycling, even for back of the field weekend riders like myself.  With powermeters, GPS, Heart rate monitors, go pros,  Strava, Zwift, cycling is a rather interesting coalescence of IoT, Social, Big Data, and even Virtual Reality.  Fertile ground then.

IoT, Social media, predictive analytics, machine learning all got a mention,  each with a cycling proof point.  He discussed the impact (pun intended) that the real time data about a major crash had on the TDF’s social engagement levels, and being able to actually prove how fast the pros actually descend. Apparently Cavendish isn’t especially speedy up the hills, but he is pretty nifty on the way down.

“The purpose of IoT  in cycling is not for technology’s sake, but it is to deepen our and the fans’ understanding of the sport.”  Rob, you nailed it.

Thanks to the folks at Crisp Research for having me along.

Disclosure:  We are Dimension Data fans IMG_4062

 

 

 

Of Cobblestones, Solomon, Paula, Gunter, Joseph and the GDPR.

I’ve been reading a fair bit of software vendor marketing and press from around the world about the GDPR. It seems to me that a lot of it misses the point. GDPR is seen as a compliance burden, an unwelcome dose of EU bureaucracy or at best a useful opportunity sell security software.  It is perhaps useful to reflect on why the GDPR and its predecessors in data protection legislation came into being.

I was walking to the train station in the rain this morning, and I paused for a moment by the pair of  Stolpersteine (tripping stones) on the corner of the street where we live. I’m not sure why I took the picture today,  perhaps they glistened from the drizzle.  I  wondered what Salomon and Paula were like, what were their hobbies and their foibles, did they watch football or play tennis together, what jobs did they do, was she left handed, who were their friends, what colour was his favourite tie,  did he make puns that made her smile, did she play Chopin on piano so that the notes drifted down the street on the breeze, did they hold hands as they walked beside the Neckar on that summer’s evening for the last time?

stolperstein image. two next to each other. Deutsch family.

Gunter Demnig began this art project in 1992. The first stone was laid in Salzburg, Austria, and now there are over 27,000  plaques across  22 countries, and growing.  Think of it as a distributed  museum.   They all follow the same format, size and font.  In situ, on the doorsteps of houses, for me they are more powerful and poignant than any centralised memorial or museum. They bring an uncomfortable intimacy and they force me to think about  how easily such an evil could come into being.  (check out more about the stones  here).

The GDPR exists to protect our data (and our person)  from abuse.

This Regulation protects fundamental rights and freedoms of natural persons and in particular their right to the protection of personal data. (Article 1 (2) GPDR) 

Software has the potential for enabling goodness, yet it can also empower evil. Software can encourage democracy, but it can undermine it too. Software can level the playing field, or it can entrench privilege.   The power of software to find, sort and group people is both awesome and awful.  It is a mighty thing that we wield.

As an industry we need to see people’s data as something to treat with care and respect. The GPDR is a long overdue firm nudge for us to remember that.

One of the pioneers of artificial intelligence,  Joseph Weizenbaum, fled Berlin for the US as a child in the 1930’s.  I suspect there is a stoplerstein for his family on a street in Berlin. His book, Computer Power and Human Reason, should be required reading for all those building software.

““The computer programmer is a creator of universes for which he alone is the lawgiver. No playwright, no stage director, no emperor, however powerful, has ever exercised such absolute authority to arrange a stage or field of battle and to command such unswervingly dutiful actors or troops.”

We proclaim gleefully that software is eating the world, and data is more valuable than oil, so it is high time the software industry took its human rights responsibilities more seriously.

I, for one, welcome the GDPR.

Customers, colleagues, Facebook, Twitter, LinkedIn and my writing day.

I have been spending the last few months deep in accessibility law, testing, standards, universal design, GDPR, the early history of business computing and of course my old friend, Sarbanes Oxley. I am an expert on Heidelberg and Sandton coffee shops, and I have spent far too long debating the value of one font over the other, and merits of footnote or in-text citations.  This week I have finally felt the adrenaline kick that comes from writing several competent pages and seeing a couple of pieces start to fit together. Long way to go though.

By the way, if you do any kind of research work, get hold of the tool called Mendeley. It is genius.

While I’m no longer in hurly burly of product management at SuccessFactors – Facebook, Twitter and LinkedIn keep me abreast of what’s up back at the farm.

It was cool to read that Employee Central had hit the 2000 customer mark. Whenever I see those milestones I think back to the early customers without whom EC would not have 100 customers, never mind 2000.  Earlier this week, Liezl’s Facebook gave me a detailed account of  her visit with a South African customer to Timken, one of those early adopters.

But what prompted me to write this was a new connection today on LinkedIn.  Tim Gregory, the Director of HR Ops at Corning sent me a connection invite. While I had been involved in the early stages of the Corning project, I’d not actually met Tim.  We had a nice online chat, and he was cool with me quoting him about their go live.

I’m the Director of HR Operations here at Corning Inc – we went live with SF in July (23 countries, 12 languages, 70 integrations, all modules – except learning).

Not to over state it… but yet we’re pretty euphoric over here.

While 2000 is a cool number. Corning as happy campers is even better. Thanks Tim, you made my day.  I’m going to be following up with you on the blockchain thing.

Now, time to get back to this pile.

image1