I was the main author on a For Dummies book and the whole point is to provide step by step actions to guide a complete newbie through the procedure. It's intentionally done that way because that's what the audience wants. They don't want to read it cover to cover, they want to pick up the book, open it to the section called "Installing a wireless network card for Windows Vista" and follow the instructions. When they want to read sections, they want to read something that speaks to them, not something that assumes they've read it all up to that point.
Technical books may have failed a certain segment of the population (I agree with many of his points if I look at it from my own perspective), but I know a lot of technical authors who make a fine living writing books, which means that people are buying them.
There are three kinds of tech books in my mind - Reference, Tutorial, and Fundamentals.
I can sort of understand wanting paper reference books, since it can be nicer to just flip around instead of googling, but it seems silly to complain about recency, since that's the _obvious_ tradeoff in getting a paper book instead of just googling for the wc3schools page on the <section> tag.
Tutorial books are a tougher one, because it's really hard to nail the audience. Is it really for beginner beginners who don't even know syntax, or just for programmers who are new to the language? Dummies books are for one group, the pickaxe ruby book is for another. I do feel like Oreilly has been nailing this recently, releasing a number of small but info-heavy tutorial books on things like stats and machine learning, while also pushing out their bread & butter 800 page language/MS Office books, and the Head First series for people new to the programming world.
With new versions of CRLS and Knuth in the past year or two, the fundamentals world is doing ok, although since they're primarily textbooks it may be a while before we see them in ebook format =(.
The tutorial type technical books are very successful with people going after vendor certifications. In this case the audience has a fair degree of technical knowledge but is looking for a path through a particular aspect of the domain. With the main vendors like Cisco and Microsoft, I don't know many people that don't get the book even if they are doing the bulk of their studying with online resources.
There is a big difference between the good and the bad in this realm. A good book is worth keeping even after you've passed the exam.
The Fundamentals category seems to be the only one really making sense. These are more about why than how, and age generally much better. Blog posts appear as ideal, with easy copy-and-paste of code, updateability and linking to more in-depth references and other tutorials.
I wonder if something like the pull requests of GitHub could be meshed with a literate programming tool (for easy extraction of source code from document) to produce the ultimate, sustainable tutorial infrastructure
Other way around - I can see zero point in paper reference books. Having all my reference books sitting on my hard drive and using Spotlight works far better than actually grabbing the right book and finding the right page.
E-book has far less advantage over paper when the material is being consumed sequentially.
I get more out of paper reference books than electronic reference materials. When I need a reference book it's usually because I can't remember something exactly, but do know the idea I'm trying to find. For me, it's much easier to skim through a paper reference manual, than to try to do the same in electronic documentation.
Linking content in the books to outside real-time resources.
Alternatively, what would you think of a subscription based purchase model for a book where each year you pay $3-10/year for a book which is continuously being updated?
There was an interesting footer that was present on all wiki pages inside of Google which read "Obsolete if printed" which was to remind you that if you printed something out and were referring back to it, it had probably changed and you should get the new version.
That being said, technical references have two roles, one to provide a reference 'now' and one to provide a reference 'then'. I could imagine that there is a market for an 'ebook' type thing which is a combination 'git repo / document' where you could read it like a reference book on your ebook reader, but you could slide a 'timeline' type slider which would assemble it from the document as it appeared at that time. Would be a fairly complex document, but really really useful.
I felt a sharp pang of nostalgia seeing the cover of "Write Your Own Adventure Programs". That book, and similar titles, were already old by the time I came across ragged copies of them in my public library, but I spent countless hours of my middle school years painstakingly typing out their listings and seeing the results on my second-hand 386's amber monochrome CRT.
Yes they are. Every time I tap a link within the book it doesn't go anywhere. Pagination is really bad, old school. Search sucks. Not to mention code samples never work.
I was the main author on a For Dummies book and the whole point is to provide step by step actions to guide a complete newbie through the procedure. It's intentionally done that way because that's what the audience wants. They don't want to read it cover to cover, they want to pick up the book, open it to the section called "Installing a wireless network card for Windows Vista" and follow the instructions. When they want to read sections, they want to read something that speaks to them, not something that assumes they've read it all up to that point.
Technical books may have failed a certain segment of the population (I agree with many of his points if I look at it from my own perspective), but I know a lot of technical authors who make a fine living writing books, which means that people are buying them.