Failure demand and value demand

Today’s lesson is not about how long it takes to drive to Newquay and why it rains when I am driving but not when my partner is. That would be yesterday’s lesson (and that would explain why I didn’t post yesterday). The surf was quite impressive in Newquay. As always, I wish that I’d gone down into it rather than stayed up on Pentire Point with the wind trying and failing to get me off my feet and face-first into the gloopy mud.

Today’s lesson is from John Seddon, and it’s about what work is worth doing. He splits the work people have to do into value-demand and failure-demand. Value-demand is when your customers/clients/ citizens ask you to do something for them that you want.
Failure-demand is when your c/c/cs ask you to fix something for them that they didn’t want.
So, selling someone a piece of software that does the job they want is a value-demand. Responding to a support call is a failure-demand.
Registering their details, is that a failure-demand or a value-demand? I think it’s a failure-demand. Most customers don’t want to fill out their details unless they can see the point. Yes, they’re happy to provide a delivery address, that enables you to give them something they want. They may or may not be happy to register the software, depending whether they can see the benefit. They want to provide the minimum of information to do so. I.e, if you hang your marketing questionnaire on the registration document, they want want to complete it. And every time they make a mistake filling out a form and you ask them for more information, that’s a failure demand. You’re not asking them to fix their problem, you’re asking them to fix your problem.

Back in the days when I used to train people in using software, we used to provide a questionnaire so that people could comment on the training. At the end of it was a comment box, asking people what sort of training they would like. Some people wanted online training that they could access as and when they wanted. Some people wanted a member of staff to come and sit with them and help them through the problem. They didn’t want training, they wanted a good fairy.

There is a question. If someone wants help are they asking:
1. Please can you make this problem go away (that’s a failure-demand)
2. Please can you show me what to do in this situation (that’s a value-demand)

Many of our current problems are because people are in situations that they don’t want to be in in the first place. The usability problem is how do you design things so that they help people not to be in the problem in the first place?

Sometimes that’s done by changing the product’s interface. Sometimes it’s done by changing its function. And sometimes it’s done by changing its marketplace. The water pouring down on the West of England this December would be highly desirable in other circumstances. Unfortunately, the only desirable aspect I can think of at the moment involves a steep slope and a mud-slide (followed by a bath, a fire, and a lot of washing).

My solution to all society’s problems

“Be nice”

There you are. That was easy, wasn’t it.

I started reading the book on Systems thinking in the public sector, and my heart has already sunk into my boots and then slightly further, when it began with a quick summary of the basis of monetarism in game theory. Not because I’m not interested in the subject matter, but because of the discounting of altruism as an important motivator in human affairs.

When I think about my life, the key thing is how I spend my time. Time is something that we know is a limited resource. We all have some and we all share that knowledge. Other things, such as family, money, status, justice, friends, stuff, individuals have more or less of, depending on circumstances. How they value them can be measured in how much of their time they dedicate to them.

Each time you make a choice in how you spend your time, you are unconsciously stating your values. So, if you’re not sure what you think is important, then see how you spend your time. And if there’s a mismatch, you are probably not at your happiest. Anybody who visits me knows that I put a higher value on drawing than on housework, but they also know that a certain level of housework is done before I can start drawing, because otherwise I have fears that the evil monster who lives in the back of the fridge and sows mould on cold leftovers will come out and eat me. (Yes, I know that mould spores, don’t attempt to out-pedant me).

Well, how societies encourage their members to spend time gives us an idea of what that societies values are. So, if we’re encouraged to sit in traffic jams rather than use public transport, what does it say about society? Does this mean that we somehow value traffic jams? Is it the opportunity to meditate that they provide us with? Is it the conversion of fossil fuel directly into pollutants without any side-benefit of transportation? Is it the fact that it’s a community activity that many people can take part in.

Suggestions in the comments box please.

The smell of brandy lingers

Coming down to the kitchen this morning, there was a definite odour of brandy. This may have been from the left-over Christmas pudding (flamed in brandy) or the sundry articles involved in flaming. The new Christmas experiment this year was playing Snapdragon. We didn’t research how to play, so merely put raisins on a plate, poured flaming brandy over them and then snatched raisins out of the flame. There was a certain excitement in watching raisins re-ignite. Apparently we should have put raisins in a bowl of brandy and then set light to that. Oh well, maybe next year. We still have some brandy left, and it’s not high on my list of liquids for easy drinking.

I am feeling slightly jaded this morning, and pondering the pleasures of doing things that are difficult and entertaining and challenging. Playing music and singing, well, all the arts in fact, are not improved by making them user-friendly. User-friendliness is about tools rather than skills.Many human pleasures involve setting oneself challenges and putting barricades in one’s way (otherwise called rules).

I went to the public library before Christmas to stock up on reading material for the festive period. I was very entertained when I realised that my heart had leapt in excitement – not from the new Barbara Kingsolver or Martin Amis, but when I saw a newly returned copy of John Seddon’s Systems Thinking in the Public Sector. Just what I needed. Admittedly I’ve read the new Barbara Kingsolver and am in the middle of Rupert Everett’s second volume of autobiography, but I’m sure that once the mince pies have been digested, I’ll be telling all my friends about how to set up systems in a better way – because usability applies to organisations as well as objects. Yes, my new year’s resolution – do a blog post on systems in my company and how to improve them.

I’m not bitter, honestly

I would like to know how people agree on a vocabulary for interface items.

Do they have a style guide (wonderful but under-rated and unappreciated) that lists all the features in the product and what are used to describe them.
Then, if they do, do they have a system of checking whether the term they’re using already exists in it with a totally different definition?

For example, I can now get on my hobby horse. One of my bugbears is the way that computers have hi-jacked the words data and information.

There’s a whole philosophical argument that it isn’t data unless it has meaning. It certainly isn’t information unless it has meaning. Information technology is a total con. You can waggle a wire in a box and get something. Now, how do you work out what is noise and what is signal? That’s question one. You’ve separated them out, but the signal doesn’t have any meaning, you can have a morse key being blown by the wind and it will send signal, but it won’t necessarily send data, let alone information.

Information only exists in context. For example, I cannot display a web page on the radio. The radio can pick up wireless signals. The signals representing the web page can be transmitted using wireless, so why can’t the radio show it. It can play sound transmitted using that medium, so why is light any different? Well, duh, it’s obvious isn’t it. It’s like saying that just because a computer uses heat it can boil water to make tea.

More fun to think about is “does a book contain information if it’s written in a language you can’t read”? How about, does a USB stick contain information if it’s been encoded with an uncrackable code and you’ve lost the password? Yes, you can think in terms of systems which would enable you to access the information, but if you don’t have those interfaces, does the information exist?

A lot of what people do is try to give meaning to the world around them. And one of the ways they do it is to take words and overload them with meaning in different contexts. So a sheet on a sailing ship has one meaning which is different from a sheet in a laundry or a sheet on a printer. And we normally decode these multiple meanings by context. But when you are in a known context, you really, really don’t need people using the same word twice with different meaning.

So, how do you get people to think if the new term they’re using is in fact an existing term. That already has meaning in this context? Judging from my jaundiced view of humanity, you don’t. But if you’re lucky, you can point it out to them and they realise that they have made life a little bit harder for everyone who uses the product.

How many boxes do you need to be incomprehensible?

Calculating tax isn’t easy, I agree. So why not make it as difficult as possible. Yes, we are in the wonderful world of options.

And my favourite set ever are the binary boxes. Yes have three boxes that can be checked. Hence you don’t merely have three possible states, you have 2^3 states.

You can have no box checked. That, apparently is the default behaviour. Let’s call it run payroll(monthly)
Then you can have one box checked. That means update payroll to today
Then you can have another box checked, that means calculate payroll to date x
Then you can have both of them checked, that means calculate payroll between date x and today
Then you can have other, which means calculate between whichever date it was and the next day that payroll would have run in the normal course of things (I think – or maybe it means do it but using offshore tax rules – or maybe it means do it without public holidays unless you are in the southern hemisphere) .

How am I supposed to explain how stupid that is? That just because something is efficient and delivers the maximum options in the minimum space doesn’t mean that it is useful.

Sometimes I hate my job. Sometimes I want to groom dogs or persuade the NRA that guns are not a basic necessity of life.

And I want a Ben Goldacre t-shirt saying “I think you’ll find it’s a bit more complicated than that”
(look Ben – I’m putting in a link

Like anything else, computers are easy if what you want to do is easy. What’s hard is finding the easy way to do something.

Normal service will be resumed as soon as possible

Mr Grumpy has been off work for the last three days. Cold, flu, job interviews? I have no idea. David has been rushing round the building in a blue silk suit pretending that he is a media medico, ready to reassure and persuade. And Gavin and Jiri have had a full head-on shouting-match. (I could possibly insert another couple of hyphens into that sentence but I’m not quite sure.)

Remember all that happy excitement when Jiri agreed that redesigning his section of the interface was a good thing? That was then, this is now. Gavin wandered through, saw that Jiri had messed with his code, and suggested that now was a good time for a review meeting.

They went into the presentation room, with the screen and the white board and the digicam and the video conferencing facilities. They shut the door, but after a few minutes noise and steam began to emerge from round the edges. And I knew that Jiri had finally cracked and started telling Gavin that he was wrong, that his approach was wrong, that his code was inefficient, and that a design path that consisted entirely of Gavin and David sitting together and feeding sharks was not effective.

I also knew that if any of the other developers had been in there, they would have crumbled by now. Beardy Nick would do anything to avoid a row, up to and including coming in at 8am every Sunday morning for six months while the system went live. And CJ doesn’t really believe that he knows enough to disagree with anyone, about code. And Ian likes working within the realms of the possible. But Jiri, whether it’s because he’s Czech and has been through National Service, or whether it’s because he is unutterably stubborn and weighs as much as a small car, will pick his corner and come out fighting.

I knew the obvious thing to do was to leave them to it. Instead, I wondered if there was anything I could do to influence the outcome, at the very least, distract David so that he couldn’t weigh in on Gavin’s side.

It was surprisingly easy to persuade Lulu, the receptionist, that it would be a good idea to put an announcement over the intercom that it was such a lovely day we were all going to head for the pub.

The joys of getting it right (and Janet)

I am so, so happy. I sat down with the lovely Jiri last week (yes, Jiri, if you are reading this, you are totally lovely and I will keep my promise to buy you a two litre bottle of Diet Coke to drink at the Christmas party) to go through a little featurette that he had been adding. It wasn’t an amazing thing: it was just an infographic tool so people could see exactly what proportion of people had taken sick days against standard benchmarks. You could compare people to other people in your practice, other people in that role in the UK and other people of your age, gender etc.  So the practice manager would see (this is a random example that has no basis in statistical truth) that nurses in their practice take more sick days than GPS, and try and track it as to whether that was because nurses in general take more sick days than GPs, and if so, is that because nurses are women of such and such an age, whereas doctors are a different demographic.

So we’re going to market it as a tool to see where you might be able to improve things: see if there are particular problems in your practice where people might be overly stretched, or to see whether there are things you could do to make things a bit more adaptable, or whether someone is consistently taking off more days and so on and so forth. It’s not full on accurate stats, but it gives overviews and places to look.

And Jiri had set it up so it would virtually do somersaults with a sugar cube balanced on its nose. But, to be able to get it to do the somersaults, you had to be incredibly precise and accurate and lay out the exact mix gradient, including opacity, xy grid lines and the exact height of the Great Pyramid in pixels.

And he demonstrated all of this stuff too me, and we discussed what the average user… as tribute to one of my best and coolest friends ever, I’m going to call her Janet. Janet is, of course, clever and funny (just like the real Janet) and a dab hand with numbers (ditto), but she’s got into practice management from a nursing background (ok, real Janet, this is veering a little bit from reality) and while she is brilliant at the HR side of it, she would prefer not to have to fiddle with computers. Her ideal is that she sets things up once a year, and they then work. And she doesn’t have to set up anything she doesn’t need to. So Janet would love the idea of a quick overview of where her nurses were doing against the UK norm (because she obviously knows who’s actually doing what on the ground), but she doesn’t have the time or interest to waste on tweaking grayscales (and if she did, she would point out that’s not what she’s supposed to be doing or what the practice can afford).

And Jiri, (yes, you, I’m so happy….) understood. And I did a nice big poster for the Janet persona, and stuck it on the wall, and there is this new mantra in the office, “Would Janet care?”. And there are basic sensible defaults for everything, and all the tailoring is tucked behind a big curtain so it’s only there if you actually want to look at it and it’s grouped by function and it’s lovely. And I am bouncing up and down (must be all the additives in the diet coke) because it’s working, it’s working and I feel I have actually done something useful.

The problems of legacy code

Did I tell you that Gavin had written most of the original code for the system? A solid mass of C++ with unlimited overloading. For any of you out there who don’t know the full joys of C==, it works on the idea of creating objects with associated behaviours. So if you say “Go!” to a car object, it drives, and if you say “Go!” to a horse object, it gallops, and if you say “Go|” to a bird object if flies. That is, they can either use a default “Go!” behaviour (say, run) or you can write a special “Go!” behaviour for them  if running isn’t the right thing to do.

But this can mean that you don’t know what they’re actually going to do. And you might not dare change it, because perhaps someone’s “Go!” behaviour is skate, and when you take out all that pointless skating rink code that someone has put in there, the whole program breaks because at some point, something needs there to be a skating rink, but you don’t know which one or why or what.

Anyway, this is the reason why they never bloody change anything. They daren’t. Because if you try and change it the whole damn thing might fall down arse over tip on that damn skating rink.

Anyway, there was a team meeting today, with me and Ian and all of the other lovely lovely lovely developers. And we were reading the customer comments to try and decide what are actual priorities. And it was blatantly obvious that they just wanted something that was dead simple to use and connected to all the other dead simple things they wanted to use. And Gavin said “But they need to be able to do X”. And I lost it. And I siad “No, they don’t” Maybe one person does and makes a big fuss about it. But they are not our customer base. They are a sad geeky person who you like going drinking with.”

Our customer base has too much else to do to spend any time being geeky.

OK, I admit it, that was a very very poor exhibition on my part. I should be endlessly diplomatic. It did make Mr Grumpy laugh. I could hear him sniggering happily in the corner. And bearded Nick tried to tie himself in a knot because he always gets terribly uncomfortable if people disagree with the god of impenetrable code who is none the less the big boss. And Jiri just wants to re-write the whole thing from scratch because he despises messy code. He doesn’t want to make it any less impenetrable as far as I know, he just wants to make it superbly efficient. Christian Jack (maybe I should call him CJ) goes into a bit of a “let me dribble sort of bouncy bouncy puppy who wants to fetch the ball” look. And Gavin lost his temper.

Organisational accidents. Who needs them?

Boring post alert. In fact extremely boring post alert. I’ve just read this through and realise this is probably the most boring post I have written so far. I’ll leave it up so that I have a standard to measure my ennui against.

I’ve been diminished on the humour front and distracting myself by thinking about organisational issues (the subject of this post) because David has just required another feature in the software. Even though the code base was supposedly frozen. So the product release is going to be delayed a bit longer. The interface is being changed as well. Instead of creating pictures, they’re providing ambiguous text comments – because it’s so quick to dash off an email saying move that button over to the left a bit. Create a picture, fellows. Yes, it takes a little while (not long, honest) but it’s one person’s time. And it will save the other person’s time trying to decipher your comments, and your own time while you explain them in greater detail and then come over and waste their time while you show them exactly what you meant. Draw the picture – or even just print a screen shot and scribble on the damn thing.

OK. That’s the end of the software process rant. Now there is a bit of  uninspired writing about the individual/system model of error.

I’ve just received a book that I ordered:”Managing the Risks of Organisational Accidents” by James Reason. James Reason is famous in the world of ergonomics and safety management, especially for the Swiss Cheese model. (Swiss cheese, in this case, referring to the sort with holes in it rather than any other type of cheese from Switzerland. If you fancy imagining it as a series of bubbles in a fondue, I won’t stop you.)

I have no specific reason to buy this book, but I wanted to re-read it, and that seemed the easiest way of doing it. (Given that for some reason it’s not available next to Ian Rankin in the local library.) I was looking for a diagram (and I can’t remember if it is in this book or a book by Sidney Dekker) about the requirement to have lots of little accidents to maintain safety. If you have a spotless record, you are then more likely to have a catastrophic accident, as you tend to more and more risky behaviour, cutting corners, because it has all been fine so far. Reason (or Dekker) describes it as those moments when you veer away from driving in the centre of your line, and are brought fully back to attention by a near miss, or the juddering as you go over cats eyes. (I’m paraphrasing here, because I haven’t found the specific paragraph yet). So it is that grab at your attention that ensures that you don’t drift off to sleep and find yourself meeting a wedding party coming the other way, with extremely unfortunate consequences for the bridal wear.

What is most intriguing about Reason’s books (I’m going from Human Error to The Human Contribution) is the movement of the balance point between individual responsibility and organisational/system responsibility.  This is a non-trivial question, and goes to the heart of many people’s beliefs, politics et cetera. How much do you control your actions, and how much does your environment control you? It can go back to the nature/nurture debate, or even the predestination/free will question.

One of the things you get taught to do when considering UI design is to make it easy for someone to take a specific path. You provide them with clear signposts. Some systems (notably IKEA store design) will make it significantly more difficult to take a route that is not the one intended by the designers. Others (such as some road signage) will merely make it more obvious to take one route rather than another. If you are following the road signs to Cambridge station, and they direct you round the ring road rather than through the town, are you making an active choice to follow that route? If the signposted route means that you avoid a notorious accident blackspot, are the accidents that don’t occur a consequence of the signage? Is anyone ever aware that their life may have been saved by though an accident that didn’t happen?

What about if the signs take you through that accident blackspot?

Who is responsible for the accident that occurs. The council that accepted the poor road design, the drivers who weren’t concentrating, the person who set up the signage that sent people there straight off the motorway? People tend to get stuck on the last part of the chain of factors – like a game of “Touch last”. In fact, we are all playing Jenga. The same action that was safe when there was a complete row of blocks beneath your block is no longer safe. The action itself hasn’t changed but the circumstances round it have.

So to go back to my original position, I haven’t changed, but the circumstances surrounding me have. For example, the cat is now within strike distance.

All the happy penguins

The penguins, in this case, are those made from two olives, cream cheese and a carrot slice held together with a cocktail stick. They look amazing. They don’t taste particularly nice, but that’s not really the point.

And that’s an interesting question? Why is it not the point? It’s party food, made out of edible ingredients, so should the taste not be at least as important as the look of the thing? Well, should it?

What is the function of food? I’m going to wamble a bit here, because I think it out as I go along, and I see no reason why I should refine my woolly thinking for the purposes of a blog post. If you want coherent incisive remarks, you’ll have to buy the book, or the report ,or a dinner, followed by a subtitled film and tequila cocktails on every bridge in Bristol.

The obvious function of food is nourishment. You can’t survive without it, but equally obviously that’s not why we eat. Imagine the simplicity of life if one only ate for nourishment, the same food every day, containing the same necessary ingredients and as easy to prepare as possible so we can spend our lives doing more interesting stuff (I appear to have just described Iceland mini pizzas).

Obviously food is surrounded by a whole set of anthropological rituals and pleasures. Food is a point where human beings interface with the world and each other. It is loaded with sensory stimulation, touch and taste and smell, the way it looks, in some cases the way it sounds (think of the sizzle as something hits a hot frying-pan or the satisfying squelch as your spoon digs deep into the heart of a really rich chocolate mousse). Food manufacturers are attempting to trigger all these responses before you buy the stuff, loading their packaging with pictures and even odours to get you to desire that food, to stimulate memories or associate it with happy experiences.

And there are all the myths about food, whether it is the revivifying effects of chicken soup or the dogs eaten by other tribes. Food defines our bodies, because it is used to build them, but it also defines our identities. It even defines our countries. For example, we have cities that are built close to rivers because people needed water for drinking and transport, and in fertile areas because people needed to farm. In different periods of history, countries have tried to develop food self-reliance (by encouraging people to farm or garden) or use food as a way of balancing payments. There is food as status symbol and class divider. Food is a deep part of our personal narratives. where possible we choose what we eat and we think that this choice is integral to our identity.

So what is the function of decorative penguins? Well, they give pleasure, they make people smile, they are a talking point. There are equivalent to the butterfly garnish (also constructed from a carrot) that I saw in Amsterdam. They delight the eye. And they are satisfying in their transience. You can’t keep an olive penguin for very long. They make few demands on your life. They make few demands on your cooking skills. They show respect for your guests (some human being has spent the time constructing these items).

I agree, it would probably be better if they tasted superb, but you’re not going to be disappointed in the taste, because you would not have had many expectations. They’re so obviously an amusement, that you know that you have been satisfied by the sight of them.

Of course, if they were all you had to eat, you might want a little more – perhaps some seasoning. And a better quality of olive. But until then, we can enjoy food that is not food, it is a ritual object symbolising party values. So long as there is something that tastes nice as well. Pass along those Iceland mini-pizzas please.