Archive for September, 2011


My views on estimating generally turn into a rant that lasts for about as long as it takes to boil an egg. Recently though, I was talking to someone at work about it and I summed it up a bit more subtly:

“IT people will never be good at estimating, so the issue isn’t so much around embarking on some (ultimately fruitless) attempt at ‘getting better’. It is about doing what one can to prepare the organization for the likelihood of things not going to plan”.

That might seem a bit defeatist. But it seems to me that we are not getting any better or nearer to solving the estimation problem. The Mythical Man Month came out in the early 70s. The only thing that has changed since, it seems to me, is the technology. What it says, pretty much all still holds today. I still get into the same sorts of conversations I did 15 years ago.

IT will never get to the point where estimates become ‘routine’ like a car mechanic’s or a kitchen installer’s, because mending cars and installing kitchens is mostly routine, ‘same again’ style work. IT isn’t about ‘same’. It is about doing specialized non-routine work with huge variation. Therefore the notion of ‘same’ or ‘routine’ never arises: You never get to that point because everything moves on again, and what knowledge you have accumulated may have no further relevance.

As an aside, this is why I have always had some unease about Kanban. Kanban assumes, it seems to me, that the work going through your board is all similar. This is what you would expect, given it’s origins in manufacturing: You don’t have variation of work on a single production line. If you do, it ceases to be a ‘line’. IT work just isn’t similar in the way a production line is. Not in my experience at any rate.

Interestingly, the answer I got after saying all this was that actually, our estimates are improving. Or that’s what it looks like. Very few of our projects are overrunning. The reason for this seems to be nothing more intellectual than forcing everyone to adhere to the same process (which I have to say has more than a whiff of Waterfall about it) and having better reporting.

This took the wind out my sails somewhat. But who am I to argue?

With the almost biblical/religious fervor with which ‘agile’ people hold ‘tests’, it is important to remind ourselves what ‘tests’ actually are.

Tests.

That is it.

Not a design document
Not training materials
Not a means of knowledge transfer
Not a way of gaining insight into the business problem
Not a way of learning about what we are doing and why
Not a way of helping to spot opportunities, a ‘better state’ or improvement

Just tests.  Written by humans.  Subject to the same flaws that any other human created product is subject to.  Though they can certainly help with some or all of the above. Perhaps.

Tests are simply a way of attempting to prove we have achieved something. And that’s fine. But they don’t give you the basic knowledge and insight that gives you context and meaning to what you are doing.  More to the point, you need this knowledge and insight in order to make sense of the tests themselves.  

People seem to overstate the significance of tests.  The mindset is like people trying to learn geography by looking at the exam papers.  You don’t learn geography that way, and IT is no different. 

You don’t learn from tests. Something else needs to exist for that.

About 4 months ago it was reported in the newspapers that radio listening in the UK was at record levels.  To lapse into the jargon, this means the overall ‘reach’ of radio is going up. In other words, instead of stations just fighting – as they always do – to get as big a slice of the audience ‘cake’ as they can, the size of the cake itself is increasing.

These reports didn’t seem to attract much debate or analysis and slipped by somewhat unnoticed.

To me though, it is important news for two reasons.

Firstly, it might – just might – mean the radio industry ‘s controversial (and expensive) investment in digital radio (DAB) is paying off. perhaps the DAB radios that people have bought as presents because they didn’t know what else to buy (I’ve done this 4 times myself) might actually be having an effect and attracting people that hadn’t previously considered listening to the radio. Who knows?

Secondly, it demonstrates that traditional broadcasting is as strong as ever. This is important to state because there has been a kind of subtext to the technology industry in recent years that assumes that the growing availability of Internet based services would in some way undermine broadcasting, perhaps even causing its total demise.  It wasn’t that long ago that many ‘experts’ were proclaiming the death of linear broadcasting in favour of demand services, downloads, podcasts, YouTube etc etc etc.

There is clearly no evidence for this whatsoever, as the above research would indicate. 

What about television then?  Sorry YouTube, but according to all the research I can lay my hands on, TV viewing has gone up in recent years.

Of course, TV does not attract anywhere near the audiences it used to: In the UK there was a time when entertainment shows could regularly command audiences of 20 million or more.   Nowadays they are lucky to get 7 (Olympics excepted I suppose).  It is also unlikely we will see audiences of 20 million again in our lifetime - those days are over.  But the point is that those days were over long before the Internet.  And certainly long before it was possible to watch video in any significant way.  The idea that the Internet is in some way responsible for a massive media ‘replacement’ is a nonsense.

So no. Television is not on its way out.  Neither is radio.

There is a deeply engrained mindset in much of the technology industry that constantly thinks in terms of ‘migration’ and replacement.  the notion of technologies co-existing – like having both a gas cooker AND a microwave, seems alien.  Yet this is how much of the world works, and people seem quite happy with it.  

So what are we to make of it?

I haven’t the faintest idea. And neither do the so-called ‘visionaries’, ‘futurists’ and other self-appointed ‘experts’ that proclaim to know the future.  And I include you Eric Schmidt.