looking for a little light reading material, i pulled my barely worn copy of cweb by donald knuth off my shelf last night. cweb was an attempt to combine the TeX markup language with the c programming language. the intention was that a good program would read like a good book: the author could take you along on the journey to discovering what worked best.
the book is both a quick introduction to the idea (there is actually a name for the programming technique, but it escapes me now) and the aforementioned journey through the process of writing cweb. which is to say that the book is largly a print out of the source code to cweb.
as both a TeX and a c junkie, i really wanted cweb to work for me. for a variety of reasons it never did, and i'm hardly alone in that experience.
reading the book with new eyes last night, i thought of a book idea of my own: "evolutionary dead ends in software development." the idea of showcasing technology that went nowhere is nothing new. i don't know of any books that focus specifically on software, but i haven't bothered to look yet.
i like the idea because it would be easy to draw the wrong conclusion as to why cweb didn't work. but if you learned the right lesson, you would minimally have a better understanding of technology that does work. maximally you could get much better at predicting future innovations.
i'm not quite in a position to write a book of this sort yet. i love the history of computers and software, and my library is full of books on programming in archaic languages and architectures. but i've never undertaken to tie all the pieces together. some things that seem thoroughly dead just evolved, while others were crowded out by more successful competitors. it is not always obvious which one is which.
i'll throw this one in the article pile instead. if i can't write the whole book, i should take a swing at a couple chapters.Current Mood:  awake, surprisingly Current Music: random techno |