Time is very, very precious. One piece of advice to learners: A math book takes a very long time to read. Optimize for "best", and best fit only. If it's free, good, if not, it's still ultimately the best return on investment! You can borrow it from a local library, much better to spend one hour doing that, than hundreds of hours on a sub par work. Good books are a good investment.
There is a best fit algorithm where you have to balance the tradeoffs of switching books for one topic.
For example lets say you're learning calc and you have a perfectly good book but the author's explanation of Green's theorem simply doesn't resonate with you. Yes, there's always wikipedia or mathworld or one of the online lecture series... but, if its free and fast and searchable, may want to try another book's perspective.
I took a EE type microwave class in the 90s where the instructor provided a lot of value by not specifying one textbook but gave us the best one or two photocopied pages from perhaps a hundred widely varying books. Sometimes the ideal textbook for a rare enough niche simply doesn't exist, or something like it existed but its 30 years out of date, or...
Free doesn't always mean good though. I recommend reading things that are rated well and reviewed properly by a community. I think that would be the most optimized way so you won't spend time reading books that don't explain things very well.
You can always try other books if the one you're reading isn't explaining some parts very well though.
FYI: Most of these are not free books. These are PDF copies of books are most likely illegal in U.S. I agree, books are a good investment, so support the authors and buy.
Whenever I see a collection of (e)Books, I have this urge to build something that functions the following way: You pick a topic out of the collection, the page tells you the prerequisites to understand the topic along with books covering that topic. For each book, it shows a tree of books, beginning with one that requires no prior knowledge up to a book that covers the topic. Something that shows you a "path to mastery". Do you think something like that is practical?
You may not like this, but I suggest another complicated dimension be applied like learning style. A concrete example of what I'm taking about is some people learn the quicksort best in prose / analogy, some learn it best in psuedocode, some learn the best in whatever language they are most comfortable in, and some would learn the best from graphical animation flowcharty type things. So its entirely possible the path to mastering the quicksort would be four different books for four different people. Of course you can "fix" that by simply having different paths to the same goal for different learning styles.
Even worse is people playing their own multidimensional games when mushed up against a multidimensional book. So is the little schemer series really a book about learning scheme or a book about learning programming or purely educational learning how to think? One book could quite easily end up in three totally different linked list / tree diagrams that you're proposing.
Purely from a startup business strategy idea, I suppose you have a guaranteed sale to some bookseller or another.
Huh. Everything from Blast into Math! to Holonomy Groups in Riemannian Geometry and well beyond, all jumbled up together in one big unsorted uncategorized list.
I was homeschooled in math thru high school, my Dad's approach was to give me 3-4 books on any subject and his own explanation, which usu involved partial diff eqs, which meant i had no idea what he was talking about. Now you can go that way with voluminous free materials on teh web
