I had to pinch myself several times this week and I have found that the frequency of pinching over the years has increased. The problem is the logic of this industry seems to be based on another planet, one on which rationality has not yet landed. I would like to invite you to consider the following two news items issued contemporaneously a couple of weeks ago.

News item 1: “Microsoft was heavily criticised in November for releasing the Visual Studio 2005 development suite too early.” Its victims, er users, claimed this product still had serious bugs.

News item 2: “Following Microsoft’s recent launch of Visual Studio 2005 … the company said it now intends to upgrade its products more frequently to respond to customers’ requirements.” The firm went on to say that agile development methodologies such as Extreme Programming and Scrum are just one way of doing this – anything to avoid thinking about the systems you have to build. One Steve Ballmer was quoted as saying the product upgrade cycle will be dropping from 24 months to 18 months, though if I had a quid for every other-worldly thing SB has said, I would be even richer than him.

Does anybody really believe this nonsense? Let me turn it into English. Microsoft wants to earn more money from you. It intends to do this by releasing products even earlier with loads more defects because the faith it puts in methodologies like Extreme Programming (ie “Extreme” as in taking risks), and Scrum, whatever that is, is touching but entirely unsupported by any experiment. The only thing that is supported by experiment is that if you rush complex applications to the marketplace, they will be full of bugs. The only reason Microsoft lengthened its development schedules in the first place was because of the extraordinary reliability of Linux.

So how do you produce complex applications with few bugs to marketing deadlines in order to keep the income rolling in?

Well, you must realise that I have nothing but 30 years of large systems computing research experience to bring to the table on this, but the short answer is, you can’t. Not a snowflake’s chance in a microwave. The belief that unlikely technologies will somehow deliver is akin to the belief in fairies, or telepathy, or alien abduction or that Accrington will one day win the European Cup or whatever it’s called this season.

Now I would like to introduce you to a brand new methodology which I am inventing as I write. I’ve called it “Budgie on a Swing”. The first law of new software methodologies is that they must have a really silly name.

My methodology is guaranteed to generate systems which have no bugs to any marketing deadline you care to name. This is how it works. You hammer on a keyboard for a while, dump the whole lot to CD and then ship to your hapless customers. The operating instructions are very simple: “Do not use.” Bah, humbug.