I own a few. I also own some from Addison Wesley, Prentice Hall, Morgan Kaufmann, etc. For example, "XML Schema" by van der Vlist is just incomprehensible (I don't understand how it got passed the editor). Plus the author is by no means an expert in XML Schema. That's actually the gist of the problem with O'Reilly books: they are written by non-experts and the editorial work is poor.
Your sample size sounds small, given the collection of books O'Reilly has. Regardless, even idiots can say wise things (perhaps accidentally). So, your initial complaint amounts to an ad hominem attack. Do you have a problem with the post, itself?
The O'Reilly brand is the "good enough" brand. I don't expect to be up at night reading an O'Reilly book in my bed, but I do expect to be up at 7pm referencing an O'Reilly book's index at my desk.
You mean, it is ok for a publisher to release poor quality books now and again as long as the average stays good enough? I feel that people with this kind of approach are not worthy of giving this kind of advice. So I didn't read the post itself.
You mean, it is ok for a publisher to release poor quality books now and again as long as the average stays good enough?
Yes. I doubt there is another way.
Do you know how tech book publishing works? It's not fiction publishing -- publishers aren't deluged with a large pool of manuscripts from which they can pick out the best ones. The financial incentives are very poor: Unless they target a very broad audience (see: David Pogue) tech book authors don't make enough royalty money to reimburse them for their time. So publishers must approach potential authors in advance, woo them, and sign them to book-publishing contracts before the books are written. Since they can't offer enough money, the publishers must approach people who have other incentives -- e.g. existing experts or inventors who want to promote themselves or their tools. Unfortunately, the skills required to become an expert in something like XML are not necessarily correlated with the skills required to write a good book about it. And, of course, it can be hard to identify an expert in advance. And it's often better to hire a lesser writer, or a lesser expert, than to have no book on a particular topic in your product line.
Once a book is delivered (if the book is delivered -- a lot of authors burn out in the process), the publisher can work with the author to edit it, but the option of rejecting the manuscript is probably difficult and expensive and politically nasty. So, once written, I suspect that a tech book tends to be published. Might as well let the reviewers and the public do the dirty work of deciding that it's bad. Particularly since there are many subspecialties where a badly-written book is far better than no book at all.
No. He means that the quality of the books published by O'Reilly is essentially unrelated to the value/truthfulness of the advice that he is giving in the linked article.
I've read some good ones and some not so good ones. The reason why Tim O'Reilly has some credibility (with me anyway) is that he has taken interesting steps beyond the traditional role of publishers.
The O'Reilly websites are actually useful (and free), contrary to what other publishers have to offer. He has an interesting blog. He is able to pay other bloggers for their work. He's experimenting with different formats and revenue models (Safari bookshelf) to make interesting content available.
Trying new things and making a living off them is awesome. That's why I like to hear his opinion on these matters.
I have mixed opinions about O'Reilly books. I don't think their quality is low but as a rule I do think they don't hold their value as long compared to books from the other publishers you cite partly because they publish many books whose subjects are less mature and often likely to change significantly in just a few years.
These days whenever I'm tempted to buy an O'Reilly book that I don't immediately need, I think about the stack I already have that I never read anymore and then check the discount. In the past two years I haven't bought any book of theirs not discounted by at least 50%.
I think that's true of all technical books. Before I part with my hard earned money on some technical tome, I ask myself whether it's a classic (Knuth springs to mind, K&R C), or at least I'll get some value out of it for the next few years. Otherwise, there's tons of good online resources that don't require you to kill a tree.
