Requirements

Requirements can be a little problematic. User, system and technical requirements are difficult enough to define well in a customer-driven field where there is a defined customer who is expert and knows exactly what they want, but in a field with a non-expert customer or no customer, they can be close to impossible.

In innovations, there is typically no defining customer, so you have to create requirements against a series of expert guesses about who the end user is, what that end user is likely to want and to use the system for, who the buyer is (and their own estimate of the end user, which is itself likely to differ from the eventual end use) and against market moves that you can only roughly predict from what you can see of your potential competitors, suppliers and market-makers (where they exist too).

And if you're innovating in a fast-moving field (e.g. consumer electronics or finance), the merit of carefully creating and working against requirements can itself be debatable because the field could itself move on whilst the requirements are being created. At which point we get into spiral development, rapid application development and other agile requirement-development models. Most of which, from the outside at least, look rather chaotic (in the bounded rational sort of sense of chaos). Just ask yourself: how many projects has Google delivered on time recently? And how much does that matter?

On the Importance of Being Honest

How important is honesty in business? For many of us, the image of businessmen - I mean businesspeople - is that of players in a large-scale game of poker, where deception is common and nobody ever wants to show their hands. Maybe that exists in the middle layers of some places, from and between people using image and perception to gain their next promotion, but on the whole people at the strategic layers of business appear to be essentially honest about what they do, what they want and what they believe their company to be. This is amply backed up by studies of leaders and enterpreneurs, which show that honesty is one of the most externally valued traits by venture capitalists, and the trait most likely to be rewarded with money. So honesty at the strategic level of a company is to be broadly applauded.

The flip side of strategic honesty is when it highlights a disconnect between the layers of a large company; where an honest assessment at the top of the company does not match an honest assessment from the bottom of it. A great deal of political courage may be needed to admit that this is the case, and a great deal of effort may be needed to bring the two assessments into alignment, but the cost if this is not done is almost always much higher (in trust, stress, and even in lost contracts) than if it is. I'm not suggesting that every strategic player puts all their cards face-up on the table at all times, but they really shouldn't kid themselves into thinking they're always holding 5 aces.

On Competence

Sh! Whisper it quietly, but some dinosaurs (*) encourage incompetence, i.e. people who don't understand the job. It works something like this. The company hires someone based on their own assessment of their competence (i.e. a 'good' CV and interview), or promotes then through the company to above the position where their natural skills are adequate (see the Peter Principal, Parkinson's Law etc). And then the real problems start. Not being able to do a defined job is bad, but not being able to judge if your own staff are doing a good job can be disastrous. And if you've built a position of influence because you present well (see CV and interview comment above), you can put yourself in a position to destroy large parts of the right sort of company.

* Dinosaur = large monolithic company. More on this soon.

Living organisations

I've been reading "transferring tacit knowledge in extended enterprises" by Nousala et al. Yes, yes, I know, but it's useful for work. So today's thought is "how far can we take the organisation as organism analogy?". The paper picks up on a previously popular theme, that organisations (companies, cities, any complex set of interacting people) can usefully be modelled as if they are living organisms. Which includes concepts like autopoiesis - how do we know when something is alive, and rather more Popperian theory than I've seen since I was last at university.

Now I have a soft spot for the independence of mitochondria - the idea that something as complex as a human can contain cells that are not only doing their own thing, but just happen to be hanging around in the neighbourhood. But I digress. What is perhaps more important is the question "is a company a cat or a big shaggy dog?". Does it exist on its own terms and selfishly decide its own fate, or hang around with its tongue sticking out waiting for someone to throw a stick for it? If a company like that is a dog, then how come dogs manage to survive? And can I use the caveman analogy now, that the dog-human symbiosis evolved because dogs were a really good early-warning system and hence worth being fed by the humans.

I really really want to explore these ideas, I have a dozen others hanging around them, some serious, some not-so-serious, but I can't. I've had my one thought for today. And this is hard, so terribly terribly hard. But I'll sit on my hands now and keep trying to reach for that elusive simplicity, that golden possibility of clearer communication. Aargh.

One a day

Hokay, I've committed to one a day. One post every day; one paragraph on just one idea. It's going to be difficult: the fear of not capturing every idea immediately is always with me, but I'm going to try. Today, I have no ideas. Tomorrow, I'll think of something.

Women Engineers Show New Confidence About the C Word

I've been spending quite a few evenings of late at the Girl Geek Dinners, the Women in Technology events and various women leaders in technology things. The generation gap between the older girl geeks who fought for equal treatment and the girl geeks coming through now who thankfully don't have to see or do the things we oldies saw or did back in the day is fascinating to me, and worth a post in its own right, but today I'm excited about something else: the 'C' word.

Until relatively recently, the women engineers that I've worked with have been very careful about mentioning their children (the 'c' word in question, in case any of you thought it might abbreviate something else). As a general rule, I've seen secretaries and PAs put pictures of their children on their desks, chat openly about their children's development, take time out to look after them as a right. But female engineers. Photos were rare, talking about kids restricted only to close friends, time off always accompanied by embarassment. And never, and I mean *never*, did anyone have that conversation about relative priorities, about the real work-life balance. And even considering it in front of the men? Noooo.

But this week I've seen a senior woman engineer do just that: tell everyone that her priorities shifted for a while when her kids were young. And she did this in front of her (male) boss. And nothing exploded. This past year, I've seen several women presenters list their children amongst their significant achievements (and we are talking some very senior women indeed), and professional women out and proud of their side-by-side roles as both professionals and mothers. And this is so wonderfully refreshing. I never became a mother myself, but I am proud that we have grown up so much as a profession that women don't have to choose between images any more. I couldn't be part of this particular movement myself, but to those women who did have the courage to say "I'm a mother too", I salute you.

Meanwhile, on a slightly less historical note, I've found and signed up with an interesting volunteer effort - IT 4 Communities. It's a place that puts geek and charities that need geek skills together, and as such it's much to be encouraged.

The Internet doesn't know everything any more...

...not that it ever did, but for a long time there were lots of people like me around the world, spending time getting as much information into the Internet as possible (with the added aim of making as much of it accessible as possible as well). But now there are gaps. There always have been gaps, but where they existed and were important to someone or anyone, they were filled with information by enthusiasts of model trains, language quirks (big nod to Tylor Jones and his June29 language site), research papers (citeseer) and a whole host of things that most people didn't know existed til then. But that doesn't happen so often anymore. I was an early adopter of the Internet (see under Geek Code) but I think I blinked, and in the time that I was blinking, the Internet went from being an inderground collective dedicated to freedom of information and became a series of newsletters, with editors and rules and regulations. I think I've been aware of this change happening, but it really hit home last week. I tried to look up my road race times over the last few years. I run - slowly but steadily - and I needed proof of that for a race entry. And so I looked myself up, in the comfortable expectation of finding at least 10 years of staggering round courses all over the country at just under 10kph. And nothing. Or nearly nothing. I've run some very big races: the Great South, Great North, London; and some very small races: little local things that one man and his dog turned out to watch (and the dog was more interested in the local rabbits). And I thought, naively, that the times and field for each of these races would still be held somewhere on the 'net, easily accessible for me to check whether I came in at 6704th or 6705th. Nope. Not there. Now back in the day, runners were an enthusiastic bunch: you ran, the results got posted and stayed on the net for you to look at later. Not any more.

I did find a couple of little local results back from -ahem- 1999, which was nicely parish-magazine. But I was quite shocked at the lack of big-race data. F'instance: the Great X series. One year's data only. Now given how much it costs to store data for a 10000 person race (name and time) compared to even a small video... heck, let's do the thought experiment on that one here... a race field of 10000 people: give them 30 digits each for their name and two digits for the time they finish in: I make that 320,000 bytes. Now compare that with a single 480*640 pixel digital camera image: 307,200 bytes in black-and-white, and three times that for colour (each pixel in a typical colour image has one byte each for red, green, blue). So if storage isn't the issue, what is? Is it access costs? Publication rights? If the race organisers put up the results for each year, they must actively remove the old ones every year, and I'm puzzled about why. I mean, it's bad enough with people selling rights to view public information without that information being completely removed as well. What's going on, folks?