Why read and write tech books?
johndcook.com
johndcook.com
Technical books are generally divided into chapters somewhere around 30 pages in length. This is also in the ballpark of the length of most standalone technical articles. So the book can be thought of as a series of shorter articles that form an overarching narrative, even if it's by a single author.
I wonder if things can go even more concise from the 30 page article form, and still maintain the property of being pleasant to read, easy to learn from and imparting a deeper understanding through a narrative going through multiple chapters. We are still bad at using proper hypertext for long-lasting useful and serious writing compared to the conventions in writing for print.
You're equating books with printed material.
This is most likely going to change, but we don't seem to be there yet. It might even be that the word "book" just can't shake off the print connotation, and will end up going the way of "folio" and "vellum" as an obsolete format-specific word when printed books start becoming curiosities.
(Also re: rsaarelm's comment, the two books I mention above were available for free online long before anyone gave a thought to publishing them. Their length is in no way dictated by printing technology or market pressure.)
By the way, I find it ironic and disappointing when O'Reilly's "nutshell" books run to 700 pages.
Petzold's Code is another example that I'd recommend to anyone interested in how computers work.
There is always new tech coming out. Tech workers have a perception that they must be up on the newest tech in order to remain valuable. Halfway decent tech books therefore sell like candy in a continuously renewed market. You also establish yourself as an authority on the newest tech, increasing your own perceived value.
The challenge nowadays for tech writers is people become impatient, they can't sit there and wait for you to tell the whole story. They simply jump. So it would be better that the book is narrative and also jump-friendly.