There's an old computer joke about God being able to create the world in seven days because He didn't have to worry about the installed base.
Nobody has had the chance to do that in computing since before computing was invented - or at least, since it was mechanical rather than electronic. The "installed base" of existing hardware and software, programmers, users and even standards, provides a context that everyone has to live with whether they like it or not.
The qwerty keyboard is an obvious example of a computer standard that predates computing, but there are others. The still common 80-character screen, for example, ultimately dates back to the 80-column punch cards that Herman Hollerith used to tabulate the US census in the 1890s.
The Americans who invented the first generation of electronic computers also took for granted the Roman alphabet being read from right to left, and, of course, the dollar sign.
Even new inventions are often much older than people think. The mouse, for example, was demonstrated by Doug Engelbart in 1968, predating both the microprocessor and the first personal computer. The floppy disk was invented in 1950, and the first conference on artificial intelligence was held in 1956.
Not only does everything have a past, the present is built on it. When Apple's software engineers sat down to create the new user interface for Mac OS X, for example, they did not start with nothing: they started with NeXT's NextStep, which was launched in 1989. And when NeXT's software engineers created the NextStep operating system, they didn't start with nothing either: they started with Unix, which dates back to the 1970s.
The worldwide web is similarly built on top of the internet, which traces its roots back to Arpanet, the US government-sponsored research network, started in 1969. Many of the web's underlying ideas can be found in books written by Ted Nelson in the 1970s. Nelson also coined the term hypertext.
There's nothing wrong with any of this: everything's right with it. We build on the past because it is what we have. I would argue that if you don't understand something historically, you don't understand it at all.
Developing innovations is not the whole story, either. Computing is not an abstract or "pure" science: it is based on implementing practical ideas using commercial products. The ultimate test of many a computing PhD thesis is financial success.
Of course, the financial rewards may not arrive quickly and they are not very likely to accrue to the innovator. Innovation has to be implemented, and it has to be delivered to the market in a form that users are able to adopt. For example, it was possible for the web's popularity to explode in 1994, when there were tens of millions of people with affordable PCs and modems. It was impossible in 1974, when those things were not commonly available.
But delivering products to consumers introduces yet more problems in a wide range of areas including manufacturing, distribution, product positioning, advertising and support. In the process, countless decisions are made by large numbers of faulty human beings, and just one mistake can consign an otherwise good idea to history's wastebasket. Empires have been won or lost on anything from pure flukes to educated guesses.
In this series of columns, I've explored some of the ways these things go right or wrong, and the approaches that companies use to increase their chances of success. Again, I'd argue that if you don't understand this stuff, you really don't understand the product. IT companies are not just peddling tin, or silicon. Products are neither developed nor sold in a vacuum, and they are not designed to be put on a pedestal and worshipped. They necessarily involve compromises that enable them to fit either well or badly into technical, historical, social and commercial contexts. Those "whole product" factors are just as important - and often much more important - than the specifications.
While there is much more to be said, I'm sorry to say that this is the last of these Saturday perorations. I have been asked to write a different kind of column in a different space, in Thursday's Online section of the Guardian. But I'd hate to go without saying thanks for all the e-mail and the stimulating discussions it has prompted. I've learned a few things, and I hope you have too.
