Ask HN: Are older programming books useful?

Tasks specific books age very quickly, and are easily replaced by googling.

Books about learning programming for beginners or just about a language have a longer shelf life. Some people like to work through them chapter by chapter as a course, and they can be particularly useful to kids in smartphone only houses who don't have day to day access to a PC.

There are, as others mentioned, lots of "theory" books that are ageless. How To Solve It: Modern Heuristics for example. But these are most useful to professionals and university students honestly. There are also pop-science type books like Algorithms To Live By or Invisible Women that are interesting to a wider audience but don't directly teach skills.

The problem with throwing all the task specific books in the trash is that they do provide some kind of a bridge between the very basics and actually doing something useful. Users don't want to learn python, they want to Automate The Boring Stuff or do machine learning or whatever. I think pointing them at the latest ebooks for that stuff might be best. If my local library let me e-borrow the latest packtpub and oriely books that would be awesome.

One thing I find wierd as I get into harder, more architectural problems, is that as an industry we spent decades thinking about stuff, creating standards, and writing books about stuff, and no one reads them. We prefer to invent wheels again because it feels more like progress than reading books.

I think it depends a lot on the language. Some languages are very stable and backward-compatible, like Perl, tcl or common lisp. Old books would be very helpful for them. Other languages, like Python, change a lot with a couple releases, so old books would not be very helpful. As others reminded, some books are timeless, because they deal with the fundamentals of programming, not specific languages.

Books on my shelf that got /more/ generally useful with a bunch of years passing since I bought them.

Simply Scheme, Structure and Interpretation of Computer Programs, How to Design Programs.

They're a hundred times easier to get going with the exercises using racket than the incredible faff it used to be when purchased, especially with SICP.

Maybe if the book is good enough the wider community supports it?

Then there's K&R C, K&Pike, Richard Stevens, of which any edition from the last 20 years is still pretty useful. Although there must be something better than K&R for C by now - the example programming style can be improved upon while still appreciating its considerable strengths. In that general direction Van Der Linden's "Expert C" and "The C Puzzle Book" remain fantastic books to grok C and low level programming.

The Wrox "Professional Assembly Language" is both out of date (32 bit x86) and still about your best bet, update it and play with the x86-64 abi docs, a download of the intel optimisation manual, "Hackers Delight" and http://aggregate.org/MAGIC/

Huge, huge amounts of kudos to you to ask here about this.

My advice to a librarian buying programming books: just stop!

There is a deluge of crap that is simply full of wrong information. So even when the topic is still fresh and relevant, those books are still wrong. An expert in the area could have a spot-the-error field day with most of it.

To be able to know what the nuggets of gold are in all the dirt, you'd have to have a background in it, or have access to a consultant who does. I would say, multiple consultants due to the amount of stuff that is out there. You can easily find heavily experienced developers who have no clue at all about a vast amount of stuff that they don't work with.

The valuable books that will stand the tests of time might not even be of interest to the public library audience; they may be more at home in a university library.

Basically, the public library is not a good venue for conveying up-to-date programming information that focuses on doing specific tasks with specific versions of tools on specific platforms. It's not a good use of the library's budget, staff time and space to be on that treadmill at all. Making effectively purchases in this area (skimming the cream of what's out there) requires knowledge that librarians typically don't have, and even if you manage to pick out the momentarily good stuff, a lot of it will have to be prematurely discarded to make space for new material.

The old-school method of learning from books has many advantages over articles on the Internet. But, unfortunately, we must admit that books are no longer relevant. In programming, for example, frameworks are updated almost every day, and yesterday's experienced professionals are forced to retrain again.

I see a compromise solution is to start with books if you are a total zero in programming. And, with understanding, switch to modern sources of information as soon as possible.