Richard P. Gabriel

This book is a colection of essays on four topics. The first concerns applying the ideas of architect Christopher Alexander, author of 'A Pattern Language' to software engineering. Alexander's book also influenced other software designers such as Kent Beck, but Richard Gabriel focuses on 'habitability' of code rather than 'elegance' or abstraction as its measure of goodness.

The second section is autobiographical and it was interesting to read about the author's life and motivations.

The third section was about running the company Lucid in the 1990s. So much effort was poured into LISP at that time and it all went nowhere. Gabriel writes as a bemused ex-CEO; he doesn't give the impression that he knew what he was doing most of the time and he certainly had no prescience about the ultimate fate of his company.

Finally there is a sadly dated section about successful technologies of the 1990s ('Worse is Better').

I found the first sections of the book to be the most interesting, especially the first, which gropes for a better way of making 'good' software.… (mehr)

I'll be honest. I'm completely prejudiced, and you should take this review with the proverbial salt. I was (and am) a huge fan of Gabriel, and I used those benchmarks. Having them published like this was just icing on the cake.

This book (and the software in it, which was also used by EVERYONE back in those heady days when Artificial Intelligence was the next hot thing) was a ground shaker. It took all the market speak out of descriptions of various hardware and software systems, and put them (finally) on equal footing.

Benchmarks of various systems started showing up here and there, and things like BOYER (a theorem-proving benchmark) and FFT (a thing of quiet beauty) made life very difficult for a while, as poor implementations were exposed, and sudden honesty enveloped the LISP world. It didn't last long, of course, but what does?

It's a great book, and deserves a place on your bookshelf, both for the historical value, and for the fine example of how to write benchmarks, and what they mean.… (mehr)