From: Chris Cera
To: Ask Tim
Subject: Open Source Books
We all know how interesting Richard Stallman can be, and how he believes that all books and documentation should be free. Many of the free documentation available is extremely convenient because it is available in electronic form, so queries can be used to search the document more efficiently. I would certainly prefer a bound version of a textbook, rather than one that I had to print, like I do with most free documentation. So the biggest benefit of all the free documentation (for some) is not that you don't have to pay for it, but that you can generate queries which reduce the time it takes to retrieve what you want from it. A solution for this problem would be for book publishers to provide the textual data of book on a CD. Then you will have to deal with the problem of piracy, which I'm guessing is why no publishers have opted to offer this. I know that I'm speaking for most of your customers, being computer people anyway, when I say that this functionality is highly desirable with any book. It sounds like your new Safari Tech Books service might be reaching out to this problem, but I'm not very familiar with it.
I was just re-reading the Linux Network Administrator's Guide (The NAG), and the author mentions that this was also published as an O'Reilly book [which is now out of print]. I would have never printed out the free version, had I known that there is a bounded version of the document at a bookstore. In any case, this is a very interesting experiment for the business world, and I'm hoping you can share some of the results.
The author states in the preface: "O'Reilly & Associates do take a risk in publishing the Networking Guide, but I hope it will finally pay off for them." My first question is, Did it? And was the output lower than that of your other books? As I browse your site, I see that you've published a second edition [which is also available as a free book], so I'm guessing that sales were at least satisfactory the first time. I must say that The NAG is an excellent book for people of all skill levels, so it is definately top shelf. If your output was higher than expected, do you think this is just because the book is so often needed? Did its free nature and availability affect your sales in a drastic way? Will this be the future of publishing?
I know that you have printed books which are freely available, or eventually allowed the text of some of your books to be freely available. I understand you have a living to make, so when and why is this path chosen?
Thank you very much for your time.
Chris Cera
Hi Chris,
You've asked several questions here, which I'd like to answer in reverse order. You asked about the Linux Network Administrator's Guide and what the results had been of publishing it under the GPL. I wrote about this several years ago in a previous Ask Tim column. But let me recap briefly here.
The NAG has certainly done well enough for us to continue to sell and update it. That being said, it appears that it sells less well than it would if it were not also available in a free version. Why do I say this? Because the NAG has always been at the top of the list of everyone's favorite Linux books, and it has been available longer than any of our other Linux titles, yet it has been far outsold by the rest of our Linux list. Books such as Learning Red Hat Linux, Understanding the Linux Kernel, Linux Device Drivers (which is also available through our Open Books Project), Programming with GNU Software, and of course, our bestselling Linux in a Nutshell and Running Linux, all sell a considerably greater number of copies.
The NAG also sells fewer copies than our other books on network administration, such as Craig Hunt's TCP/IP Network Administration, or even his Windows NT TCP/IP Network Administration (even though our Linux list, on average, significantly outperformed our NT list).
This situation was particularly acute when we first published the NAG back in 1993, because the social norms around publication of free books were not yet clear. Specialized System Consultants (SSC) republished the book without any royalty payment to the author, and as a result, was able to offer the book at a lower price. Que incorporated the book "whole cloth" into another Linux networking book, also without any compensation to the author. This definitely cut into our sales. But eventually, as we kept marketing the book, and as our brand became better established, the "freeloaders" dropped out of the market. As we've released additional books under free licenses, we haven't (to our knowledge, at least) seen a repetition of this pattern.
Still, even now, we believe that there is some erosion of sales from the free availability of the online version. Other books that we've put out under free licenses, such as Using Samba, do well, but still somewhat underperform relative to our normal expectations.
Of course, this is second guessing the market, and may be incorrect. But more profoundly, it is beside the point. Our goal in publishing is not to maximize the revenue and profit from every book (many customers have noted that O'Reilly books tend to be more reasonably priced than competing books), but to strike a balance between the costs of developing and producing the books and the optimum spread of the knowledge they contain. Many of the authors of books on free software would like to tilt the balance in favor of availability rather than profit, and we are glad to oblige them, as long as we feel we can make enough money to support the book and the company in doing so. Other authors are more concerned with maximizing the return on their investment of time, and in these cases, we will tend to make the choice to put the book out under a standard copyright license.
As I've written on numerous occasions, my principal goal is the spread of knowledge, not the ideal of software freedom, in and of itself, that is Richard's cause. I admire and support his dedication to that cause, and we are fellow travelers in many respects. I wish he could be more accepting of those who share some of his goals but do not embrace them completely. However, I also know the truth of the saying that change depends on unreasonable men.
For more thoughts related to this discussion, see my articles Why O'Reilly Books Aren't Free, My Definition of Freedom Zero, and Open Source and the Obligation to Recycle.
Now, to your first question. Wouldn't it be nice if there were a searchable text version of each book available on CD-ROM in the back of the book? Yes, it would be nice. Unfortunately, it's not economically all that viable. The cost of printing and binding in a CD-ROM increases the manufacturing cost of a book by about 50%, believe it or not. And remember that the publisher receives less than 50% of the cover price of the book, with the rest going to retailers and other middlemen. As a result, most books containing CDs end up being priced at $5-$10 more than books without CDs. (Simply to recover manufacturing costs, a book with a CD needs to be priced about $3 higher; usually, the publisher tries to make a little profit as well, especially if the CD assembles some real added-value information.) But how many of those CDs would go unused? For every reader who would love the CD in the back of the book, you're taxing others who might not. This is why David Pogue's Missing Manual series waggishly includes a picture of a CD on the inside back cover of each book, with the title "The Missing CD" and a reminder that you just saved $5, and can go to www.missingmanuals.com for the utilities and other goodies that might otherwise be on a CD in the back of the book.
So if putting the information online rather than on CD is the answer, why not a searchable text version of each book online? It's not a bad question. In an alternate universe, it's probably something we could have tried. But we've had a vision of a much richer online service, and have put our energy towards that instead. As you guessed, I'm talking about Safari. As with our book publishing, we're trying to strike a balance between spreading knowledge and getting revenue. We're working very hard to build an online information service that is valuable enough to pay for, because we believe that we can provide more services and add more value that way. In fact, we believe that Safari (and other online services like the O'Reilly Network) are critical to our survival as a company. We find that more and more people are seeking knowledge online rather than in print books, and we believe that if we, as an industry, don't have a way to charge for our online information, we will eventually be out of business.
The biggest competitors to computer book publishers are not other book publishers, but vendor-supported online services such as MSDN (the Microsoft Developer Network) and java.sun.com. These online services are terrific, but they are subsidized by the sale of the vendor's products and are beholden to the vendor's goals. I'd like to see a vibrant, online information industry, and that requires business models beyond advertising, and also beyond print books.
There's a final point that's really important, and that is that the format in which information is presented online may not, ideally, be the same as the format of a print book. One of the things that bugs me about demands that we make our books free for online redistribution is that we do in fact make the information from many of them available online, just not in exactly the same form. The relationship between our Perl books and the online Perl documentation is rich and complex. When folks like Larry Wall and Tom Christiansen and Jon Orwant are updating the Camel book, they are also updating the Perl man pages and other online documents. Material flows back and forth rather porously between the two.
I've made the same offer to Richard Stallman and other free software advocates. I'd be happy to see more material from O'Reilly books incorporated into the documentation of various free software projects. This is good for both parties. The free software project gets better documentation, but the publisher has added value and a different format.
Tim
Return to: Ask Tim
