Marginal Costs vs Total Costs
Kevin - the current high price of textbooks is precisely because the cost of writing / proof-reading / etc is relatively high compared to the marginal costs of merely printing the things, which can now be done near enough on demand - if not literally so.
You can see this in the way that University textbooks cost even more than school textbooks, even though they have frequently lower production values (i.e. less artwork) - it's a factor of them having a smaller audience to spread the cost of development.
For what it's worth the same thing has applied with music for a long time - the fact that 'CDs only cost 50p to make' actually shows that the marginal cost was always the least relevant part of the total cost.
Now I'm all for reducing the costs and increasing the quality of education, but there is no reason why this has to have some link to digital technology per se.
In theory, the cheapest form of education would be a single national curriculum, a single exam board, and a single textbook for each course, with all copyright owned by the state/public. That would actually make even more sense if the marginal costs of textbook production were high than if they were low. Low marginal costs actually suggest more room for creating customised variations.
Not to mention the ideology that private companies competing should deliver a more 'competitive' solution than a monopoly (state or otherwise) - which is what paying someone to deliver a definitive 'free' work would be.
On the upside - BECTA are starting to recognise the importance of open source in educational software which may be the start of a transformation in that culture - equally the growth in sites for teachers to share lesson plans and resources (expect the government to decide to implement a teacher specific social network soon).