I remain unconvinced that there is currently a better format for reading novels than the book. Early in my career I worked on a system for selling books online and my personal conclusions from that experience are that general reading is not well suited to a computer. I can read a book pretty much anywhere as long as there's sufficient light, the form factor of a laptop and additional overhead and constraints can't compare. I've tried electronic books from Amazon on a PDA and that was painful. Reading a paragraph at a time is just too disjointed. I'm not sure how Amazon's Kindle will fare but it will need something compelling to survive.

I suspect that for the foreseeable future electronic publishing is going to be a supplement to physical books rather than a replacement. For example I subscribe to the Safari service which I find highly useful, mainly in bursts. I don't find it particularly compelling for actually reading the books. Yesterday I purchased a physical copy of a book (Domain Driven Design by Eric Evans, highly recommended) that I had available on Safari. This is because it's fundamentally better to read a physical book than to read it in a web browser. Even though I could read it online the additional value of a physical copy was enough to entice me to pay for it. Safari can't compete here. Where it provides value for me is that it has many books available and they're searchable. The majority of the value I've gotten from Safari comes from being able to search across multiple books for content.

Currently the content in Safari is of sufficient value to me to justify the subscription fees. However for day to day usage Google is still my first option for finding information. It remains to be seen if Safari can make sufficient content available such that it retains its value proposition. Its utility is also particular to non-fiction books. Search capabilities are significantly less useful in novels which tend to be read linearly, whereas readers of non-fiction are often (but not always) looking for specific information rather than reading the complete work.

My conclusions therefore are that electronic publishing is excels as breadth and search where physical constraints limit physical books. Where electronic publishing falls down is in accessibility and lack of ubiquity, particularly in infrastructure. DRM also remains a stumbling block, restrictions on sharing significantly impact the utility of electronic books to end users. These problems aren't insurmountable but I remain unconvinced by any current solution. I just hope one is found before my apartment completely fills up with books.