aaronsw: Point taken. I hadn't given a lot of thought to what your statement might mean; my complaint was with the wording of your comment, not necessarily with your opinion. It's true that TAOUP doesn't deal with "art" as much as the title might suggest.

mbp: Your post is exactly the sort of pointful critique that I intended to draw out. I do wonder why ESR appeared to be in such a hurry to finish the book. Of course, learning enough about other people's programs to use them as effective examples would take some time -- but I do hope that ESR, or someone else, will eventually take the time to do just that. You're right: it needs to be edited, and by other people.

Still, why do I like TAOUP? Again, I greatly appreciate the "case study" oriented format (even if it has some holes). Frankly, vague generalities bore me. I want the specifics, because it is in these that true wisdom shines through, "the open window through which you can see not concepts, not ideas, not beliefs, but the very goods. But if you say what it is that you see, you erect an image and an idol, and you misdirect people" (as Alan Watts puts it).

TAOUP isn't perfect. But I think ESR does a better job of exhibiting concrete examples than most people could, even if he does seem a bit enamored of his own projects at times. It's quite refreshing in a world of books discussing useless toy programs at length. What's important, I think, is that he went and did it: and now we can all read it and think about what needs to be improved, rather than just say "someone ought to write a book about the Unix philosophy".

One of my favorite sections, by the way, is Data File Metaformats. It's a piece that sorely needed writing, because so many people have trouble designing good textual formats.