The "ownership' of software is based on copyright law, and the General Public Licence (GPL) formulated by Richard Stallman is based on a new application of this law. In consequence, one might think this Open Source thinking could be applied to other works that fall under copyright law. And of course that has already been done. First, there was literature.

The invention of the printing press is what prompted the evolution that led to the copyright laws we have today, which makes it a natural starting point. Financing book publishing also resembles the software industry in a number of other ways. The actual creative work is in writing the book, after which any number of copies can be printed - not for free, perhaps, but at a reasonable unit cost. In today's Internet world, it's naturally also possible to distribute a book in digital form, which truly does make copying costs nonexistent.

So, why aren't books published according to the principles of Open Source? Perhaps the reason is simple: people don't think of doing it.

On the other hand, it might also be that Open Source doesn't give quite the same edge in writing a book as it does in computer programming. After all, writing software is usually a very complex process which requires the input, often of hundreds, of people for several years. In this complex process the benefits of openness compared to closed systems is obvious (as is proved by Linux, Apache, etc.).

Writing a book, then, is not so complicated. Actually, your average book can be written by one author, and it doesn't necessarily have to take even a whole year (provided you don't write only at the weekends, in addition to having a day job, which is how I wrote this book). But to collaborate with others to write a book, one would have to be able to divide the workload in some sensible way, and for many books this would be very difficult. If it were fiction, for instance, the writers would have to discuss their plot very carefully, agree on characters and so on, in order for the story to be told with any coherence. In the end, it would be far easier for one person to write the book on their own.

Even so, Open Source books do exist. But rather than being plot-thick detective stories they are more likely to be, for instance, Linux manuals.1

At a very early stage, when the first GNU books were beginning to take off, it was clear that a free program should naturally have a free manual. And behold, writing manuals works fine within the framework of the Open Source process. It's easy enough to divide the parts of a manual for a number of writers to work on. Another benefit is that there's usually a need to update the manual whenever a program changes and evolves. This too makes an Open Source manual better than something not just anybody can change, as most of the old manual can usually be employed in the revised one. It's only necessary to change the parts that relate to what has changed in the new version of the software, to add new chapters for the new features, without having to rewrite the whole book. Such a process would be more complicated with conventional books under current copyright law, which would not allow an out-of-date manual to be updated without permission of its original author/s. Which is why members of the Open Source community always include a text specifically granting this permission to all and sundry every time they release a manual or the code for a program.

And because that is likely to discomfit a lot of publishers, this is a good place to reiterate what has been said about software in earlier chapters. There are bound to be people who feel that publishing a book on the Internet under Open Source terms is a bad idea and will be the ruin of the publishing industry. But this is roughly the same thought that struck traditional software businesses when they first heard of Linux. In actual fact, the situation is the reverse. An Open Source book which is available on the Internet is like free money for a publisher. It's a book that is already written and only needs to be printed for the money to start rolling in. And lots of smart publishers have actually realized that. Apparently, the publisher of most Open Source books is O'Reilly, the same publishing house that once upon a time put Larry Wall, the creator of the programming language Perl, on its payroll and thereby gained its role as purveyor of books to the Open Source movement. But O'Reilly is not the only publisher to discover the Open Source market. Prentice Hall actively seeks Open Source writers for a series which has as its Series Editor the spokesperson of the Open Source community, Bruce Perens.

So there are some Open Source books. But we've slid back into the field of Linux. So is there any other Open Source literature outside Linux manuals?

• 1Linus Torvalds' grandfather, Ole Torvalds, was actually involved in a detective story collaboration with a number of other Finnish authors writing in Swedish. Each author wrote one chapter, but there was no discussion of plot or characters; the manuscript was simply handed on to the next writer, who had to come up with a way of moving the so-called plot forward. The collaborative book was published in 1950, but it was not a bestseller. The title Den rödgröna skorpionen, which translates as The Red-Green Scorpion, may have had something to do with that, as it's a title that sounds almost as classy in Swedish as it does in English!