Very little / super-brief explanation in many cases, but as it's a book of algorithms, not really about, it's acceptable. And I found the brief demonstrations / explanations / visualizations to be terse, accurate, and quite explanatory - near perfect. Not a book you can casually peruse, it's largely a list of algorithms, though it appears to go roughly in order and build off itself. Where necessary, it goes into a fair bit of detail (sometimes several pages), but truly most of the code is pretty self-explanatory once you know the purpose.
Not a book to skim through, but definitely valuable. Uses C/C++ (minimal C++ functionality, classes + templates) heavily, of course, but big shocker there given that it's largely focused on efficient, low-level code. I'd rather have functional C than questionable pseudo-code anyway. A little of a language called GP and a little pseudocode, but very little from a quick skim. My main criticism here is that the variable names could be more descriptive, often just a single character or two, but most of the code is small enough that it doesn't matter, you can do it in your head easily enough.
Definitely a keeper, IMO, unless you despise C.
Disclaimer: I write as little code that looks like this as possible, and have not examined it in detail. Anyone care to comment on the code library itself? I'm definitely not qualified.
I get while you don't like pdf, but I'm not sure I understand why you want to use HTML files. HTML is for web layout, not publishing. I would much rather have an open format like DjVu[1][2] catch on -- it's like PDF, but has better compression and is an open format.
HTML is universal and can be edited with any simple notepad. It can look good if you know how to style it. Also sending them over the wire is easier than zipping/unzipping folders with tons of imgs/resources.
I've got some books and have put them in one html file, they look good enough for reading on my laptop, netbook, ipad, etc. So portability and universality are key.
Here, have a look at some of my private collection, most of Sherlock Holmes from gutenberg project:
I've had a couple multiple-meg HTML pages on my disk, largely documentation. So, no images, simple HTML & CSS, etc. Every browser I've tried works longer than Preview for opening the file vs a basic PDF. Searching also typically takes longer, and pauses the application while it searches. As much as I wholly agree that I'd love HTML resources instead of PDFs, it seems browsers aren't made for things this large, and getting proper display for all the code & examples & math notation is not easy.
Admittedly, Preview only does a subset of PDF capabilities, and it's one of the fastest I've encountered, so it's not a wholly fair comparison.
HTML5 + CSS3 makes it easier, because fonts can be embedded, and there's a LOT more control over many things, and I know of no normal way to do anything like pagination. And many PDF viewers have interfaces optimized for reading, where browsers have interfaces optimized for navigation. I hope it's the future, but it's not the present.
Your first reason is why I prefer to use LaTeX and distribute .tex files along with .pdfs. I think the syntax of TeX is nicer than HTML and since I'm not a good or fast designer I don't have to worry about designing things nicely, adjusting for different browsers, etc. I do like your idea of embedding base64-encoded data for the browser to read, though.
Not a book to skim through, but definitely valuable. Uses C/C++ (minimal C++ functionality, classes + templates) heavily, of course, but big shocker there given that it's largely focused on efficient, low-level code. I'd rather have functional C than questionable pseudo-code anyway. A little of a language called GP and a little pseudocode, but very little from a quick skim. My main criticism here is that the variable names could be more descriptive, often just a single character or two, but most of the code is small enough that it doesn't matter, you can do it in your head easily enough.
Definitely a keeper, IMO, unless you despise C.
Disclaimer: I write as little code that looks like this as possible, and have not examined it in detail. Anyone care to comment on the code library itself? I'm definitely not qualified.