THINKING MACHINE

Dick Pountain/Sun 10 August 2003/1:46 pm/Idealog 109

Recently I've devoted rather more of this column than intended to arguing the case that computers are never going to be very much good at thinking, in the human sense. However if you deduced from those columns that I no longer believe computers are very important, you'd be badly mistaken. On the contrary my computer is the greatest aide to thinking I've ever had, and I cherish it to the point of perhaps unhealthy dependency - most of my other columns are devoted to measures taken to avoid the trauma of losing its services for even a couple of hours. The fear of it failing to boot (even once) now deters me from upgrading its OS; I devote inordinate amounts of time to devising backup procedures; and I catch myself drooling over pictures of new and desirable storage devices.

Actually there's no contradiction between these two positions. My computer may be absolutely useless at thinking but it does things to help *me* think that I wouldn't have believed possible 40 years ago when I used to toil amid the dusty Victorian stacks of the Science Museum library. For one thing it holds huge amounts of information, relieving me of having to keep getting up to scrabble through shelf-loads of heavy books. A favourite cliche when writers write about their own profession is that the biggest problem is applying bum to chair. That's true enough, but you seldom read its codicil, namely that the second biggest problem is *keeping* bum in contact with chair. Every interruption is a disaster, especially when you're trying to think hard, the way programmers, mathematicians, poets or philosophers have to. Having all your reference works on screen at your fingertips is perhaps the biggest boon a writer could have. And of course the computer can search those works thousands of times faster than I can scour the index of a book, which means I look up things that previously I wouldn't have bothered because it wasn't cost effective: knowing the answer wasn't important enough to cover the cost of the interruption.

You've probably jumped to the conclusion by now that I'm talking here about that cornucopia of knowledge (real and pseudo) the Internet, but in fact I'm not. Sure, I use Google as much as the next person, but it's a last rather than a first resort. This has nothing to do with connection speed, 'always-on' availability or whatever, but to do with quality. Having truly excellent reference sources for me still means picking and choosing and storing them locally. The works I keep permanently on my hard drive include the following: Collins English Dictionary; New Shorter Oxford English Dictionary; Oxford Language dictionaries for French, German, Italian and Spanish; Encyclopedia Britannica 2002 Standard Edition; the Grolier Encyclopedia; an interactive Periodic Table; the complete works of Shakespeare; my own Penguin Dictionary of Computing (in eBook format); a complete archive of PC Pro Real World columns; and many more. They still fill barely half of a modest 10GB hard drive, and I'm not worried that they don't share a common user interface - all the interface you really need is an edit box and a button marked 'search' or 'find' (which is exactly what all of them have).

There are one or two reference sources that I still run from CD-ROM - for example Microsoft Art Gallery or the collected Byte Magazine - but those I use most go on the hard disk since space is so cheap nowadays. I know there's a whole class of software out there for doing this job but I tend to do it the old, dirty way by copying the contents to a folder and using the DOS SUBST command to fool any bewildered retrieval programs.

As for those miscellaneous research materials I gather in the course of a writing project, I've stopped worrying about them too. I used to be a great devotee of free-text databases (having worked for years with both askSAM and Idealist) and stuffed information gathered from far and wide into one of those. Now I've stopped all that and just stick them in regular Windows folders, under a subtree of My Documents named after the particular project: that way it doesn't matter that they're in a random mixture of HTML, PDF, .TXT and .DOC formats. If I really need to search for something I use Windows 'Find | Files or Folders' which is plenty fast on a modern machine when pointed at the correct subtree (I don't even bother with the Windows indexing engine any more). For more complex searches there's always the 'Find in Files' command in TextPad, but I very rarely have to search at all because the titles of the files remind me where to look.

It's only taken me 20 years to come to this realization that the OS file system is the best place to store my data, so I'm in no position to sneer at Microsoft for taking around 10 to arrive at the same conclusion. I refer to the forthcoming Longhorn release of Windows which begins the process of integrating SQL Server and XML into the file system itself, so you'll be able to take multi-dimensional views of your data. For example, at the moment I store past Real World columns in two different ways - by issue number and by author - for my convenience, but under Longhorn these would just be two different views of the same set of data. It might even be enough to persuade me to upgrade.