[Home Page][Index of Reviews][Achilles in the Quantum Universe][Beak of the Finch]


The Aesthetics of Computing
by David Gelernter

David Gelernter is one person who argues that computers should be more than just functional machines. He feels that they should also be 'beautiful' pieces of work and sets out to show it in this book.

Initially, he outlines what he means by beauty. He defines something as beautiful (whether it be a mechanism, a mathematical equation or an algorithm) as meaning something that is both simple but powerful. As examples, he talks about 'Quicksort' (an algorithm used to quickly sort through a list), object-oriented computer languages, tubular chairs, etc. I find this to be quite in line to the way I see beauty in computer programs too (they should be simple but powerful and 'elegant' to be considered beautiful).

He then goes on to look at why beauty is so often overlooked in the technological world. As with other people, he looks at the Apple Macintosh and why, despite being such a 'beautiful' and 'elegant' machine, it failed to capture the marketplace. He feels that it may partially due to the perception that beauty has no place in a technological workplace, a perception that change later (too late for the Macintosh to take its rightful place on a user's desktop).

Next, he looks at the GUI's (graphical user interface) desktop metaphor and points out its strengths and weaknesses and proposes his own ways of working with data. First, he looks at how data is stored and proposes 'Linda', a unique 'storage system' for data. In Linda, data may reside anywhere in what Gelernter calls 'tuple space' which may be spread out over many computers or concentrated on one computer. Only Linda knows where everything is.

Programs send requests to Linda to store or read data; they never know where the data actually is. To Gelernter, Linda is beautiful as it is both a simple but powerful way for programs to access data that can reside anywhere. He contrasts its simplicity to the way many current parallel programs must 'hardcode' the location of data, making them harder to work with. Having seen a demonstration of Linda many years ago during my university days, I believe Gelernter has a point here.

Next, he looks as the way we work with data and proposes a totally new way of looking at data by proposing 'Lifestream'. In Lifestream, data is not separated into folders and files with (usually) arbitrary names. Instead, all data is seen as a stream of files arranged in chronological order. By 'cruising' up (forward in time) or down (backward in time) the stream, you can obtain any data you want. Categorising is done by creating a substream consisting only of the data you want to see. This is an unusual (to me) way of looking at data and until I can see a working example, I reserve my judgment on this.

Finally, Gelernter looks as the way computers are built and finds them wanting. He proposes (with drawings) several novel ways of arranging computers and furniture to give a more 'aesthetic' work place.

This is a fascinating book that you will find much to agree and disagree with. However, despite his attempts to make the book accessible to those without computer knowledge, you do need to know something about computers to appreciate his arguments. I also found some of his analogies to be somewhat awkward.

You may not accept all his ideas on how to make machines more beautiful but it will make you think about the way computers are being used in our lives and what may be in store for us in the future.


[Home Page][Index of Reviews][Achilles in the Quantum Universe][Beak of the Finch]


Copyright (C) 1997-2002 Soh Kam Yung [Best Viewed With Any Browser]
All Rights Reserved
Comments to author: firstspeaker.geo(at)yahoo.com
Generated: Thu, Apr 11, 2002

1