I think it is unfair that you didn't include answers to the exercises in the book.

After spending money on a book, I would expect to have the answers (which to me make the book complete) included. Second, the fact that I can't buy the answers on your website is really frustrating. I don't imagine it would take very long for you or one of your cohorts to put the answers together.

I've gotten a couple of comments like this, and I felt it would be worth some research to address the issue.

First, you're right about the delay. It took me almost 1.5 years to get started on the solution guide (for the 2nd edition; the third also ended up being a very big job), and that's far too long. It would have been much better to have had it available when the book came out, but that would have delayed the book (background sounds of publishers gnashing teeth). Since then, I seem to have a plate full and have also been procrastinating because it looked like a big job.

I certainly apologize for the delay. I've only started to get good at coming up with exercises in the last couple of years. Inventing good exercises is hard enough, but now I'm learning about solutions. I'm sure I'll get better at this with experience.

To address the last sentence of the question, it has turned out to be a very big job, about 2.5 months of work. It's like writing another, albeit smaller, book. In print, it looks like it would end up being over 400 pages. So no, it's not a trivial project.

The number of pages explains why it isn't part of the book. As it stands, I was trying to cut the book down and the publisher was squealing about the size. Adding over 400 pages wouldn't have flown -- the book would have been enormous.

So one option would be to publish a separate solution guide, which would almost certainly cost you more than what it will cost via the web, and I suspect you wouldn't prefer that. I think the electronic version is the best approach; if there's a big enough demand for a separate printed book I'll consider that, but right now I don't think it would be justified.

There's often an illusion about how much an author gets for the book that you buy. For a US author, if the book is sold in the US, roughly half the cover price goes to the bookseller, the other half goes to the publisher. Of the amount that goes to the publisher, roughly 10% of that goes to the author -- if the book ever pays back its advance, which the great majority of books don't. If the book is published outside of the US, then the foreign rights have typically been sold for a flat fee (usually not a very impressive amount, since foreign publishing is very risky), and there's no connection with the amount you pay and any royalties paid to the author. I point out all of this so that you can be aware of the incentives (or lack thereof) involved in creating and maintaining a separate solution guide. As it turns out, it has taken a long time for the time and money I've put into the solution guide to begin to pay for itself.

I agree, everything should be available to everyone for free, like in that "Star Trek Next Generation" episode where they go back in time to San Francsico and see Guinan and Data's head and bring Mark Twain onto the ship and explain to him how they don't need money anymore and that everyone just tries to fulfill their potential. Someday that may happen, but in the meantime I feel pretty good about giving away more than the majority of computer book authors and publishers. I hear from people all over the world in countries where it's too expensive to buy books, about how they are able to learn from the free electronic books whereas otherwise they wouldn't be able to. I work very hard to produce quality work, and to go the extra mile to provide explanations, and supplements, yes, and exercises that you don't get other places, and it is wearying to hear "but you should give me this for free, too!" The fact that you don't have to buy the book at all seems to me to be a nice gift, so if you must have it in convenient print form because "you could never read it on screen," then remember that what you're primarily paying for is exactly that convenience, and the paper, printing, and distribution that is required for it.

Finally, I decided to look and see how many books (1) Even had exercises and (2) Included solutions. My Java shelf includes a lot of books, and I've gotten rid of the ones that aren't so good, so I think I have pretty good coverage of the better Java books. I found only several that even had exercises. Then I decided to look back at the C++ books. The most popular two C++ books were Bjarne Stroustrup's The C++ Programming Language and Stan Lippman's C++ Primer. Both had exercises (these are the only ones I found that did, besides my own C++ book), but neither had solutions. In fact, both had separate solution guide books written by separate authors (supporting my assertion that it's a big job), each costing 30$.

So according to my bookshelf, it doesn't appear to be so unusual that solutions are a separate item from the book. What does appear unusual is to have exercises at all. So if you still feel it's unfair to have a book without exercise solutions, then look at the alternative, which you'll find in most books: no exercises at all.