Text books are the ideal public good. Some people get paid to make the book, and after creation, it should be placed in the hands of anyone who can benefit from it. I remember seeing Maharashtra board textbooks for Physics and weeping from how shoddy their quality was. A lot of them were filled with errors and the diagrams were totally misleading. What was worse: I bet the teachers were unaware of Feynman's lecture series or other alternatives.
In the US, textbooks suffer from the opposite problem. The quality of production is spectacular, and that makes the cost of production high. In addition to that, publishers print new volumes often, jiggling chapters around, so that a student with an older version would be lost. New chapters also change problem numbering and other trivial issues to force the user to upgrade. Hey, that sounds a lot like a convicted monopoly I know! This forces a very high price of books on students.
If a pdf copy of a book is available online, the marginal cost of producing another copy is zero. And all of us benefit when every kid has access to correct, royalty-free textbooks. Most children might not have access to a computer, but keeping the content free for publishing would ensure that different publishers can compete to publish the book at competitive prices. Old books are valuable again, and children can just download the changed pages instead of the entire book again. Almost like patch files that reflect the changes in the kernel of a lovely OS I know.
Inexpensive, accurate textbooks would reduce the cost of education, and would help to standardize the curricula of science. Newton's laws apply equally to both Israelis and Palestinians anyway.
Here is a great physics text available online. How I wish I had access to such a lovely book in my school-days: http://www.motionmountain.net/welcome.html
Leave a comment if you know of such books for: Chemistry, Math, Statistics, and other Sciences. I do know of the wikibooks project and have great hopes from them.