I saw a post from O'Reilly that they were looking for reviewers for Mark Piligrims html5 book and jumped at the chance. Mark is pretty well known and the subject is fresh. I asked and was given an epub version of the book. I am using FBReader to access e-books, although I also used the book site when I needed to see some parts. There are subtle differences, but they generally agree. This is my review. I wrote it as a series of notes, as I read the book, hence it's in no way a look back at reading it.

I find the intro history lesson a bit long? Although the summary and lessons learned are revealing.

Re Detection. I'm just a little concerned at the level of such an early chapter? If the reader is an HTML novice, this level of js could be something of a deterrant? Perhaps a Warning in this chapter? Then again, it is 'dive into' Mark! I eventually came to the conclusion that this book assumes a reasonable level of competence at writing HTML and using jscript.

I'm beginning to like the writing style. Nice and relaxed yet with the detail. Suites me anyway.

Mmm. when will we 'dive into' html5? Lots of potential described, nothing for the reader to do or try yet?

OK, chapter 3 and we're off. Very pragmatic approach. I like the 'Ask professor Markup' sections.

Seems as if Mark has really done his homework on the history of HTML. Impressive.

I note that in most examples, XML closing isn't used. <link ....> seems preferred, yet isn't mentioned.

I'm appreciative of the notes re people with disabilities.

In the footer notes, the copyright date is 2001-2009, not 2010

The canvas work seems more related to js than html5, but I do see the overlap. Even the halma game has it's purpose.

IMHO the mix of 'pen' and lines needs a bit more explanation, e.g. the sequencing, when the 'reset' (beginPath) is needed etc? the change color example in the 'graph paper' isn't clear how this works.

Chapter 5? Video. Lots of background. Informative, though as Mark says, the eyes do tend to glaze over somewhat. Not sure how the firefogg tutorial helps. Ditto the other creation options. I'd be more interested in viewing video. The final 'you should be able to watch this' example didn't work on my system (no flash perhaps?).

chapter 6? geo stuff. Brief, but clear, well written and quite intriguing. No information provided on the format of the returned time string?

chapter 7. Local storage Less code/ detail than other chapters.

chapter 8. Offline Detail is becoming scarce. No links to the demo (the js or html files for the halma code).

chapter 9. Forms Very good. Shifted me towards Opera, it having the most advanced forms handling. I couldn't get email or url types to validate in Opera 10. Potential, not yet realized?

chapter 10. Use of dereferenced namespace stated as 'best practice', yet contra to W3C namespaces? why not use foaf? Very thin vaneer over RDF triples, blunted by using photo as a verb? Seems rather strained when 'affiliation' is contorted into 'employed by'.

with no browser support this is somewhat tentative, though hopefully it will gain support beyond the search engines.

I can understand why Mark uses this markup, I would have preferred it mixed with other semantics, just to show the options.

The clash of W3C geo-location and microdata geolocation sticks out like a sore thumb.

Again, google seems to have extended the microdata ideas along rdf lines with its own conventions.

I do like the list of references at the end of the chapters, although they will date in the epub and print versions.

chapter 11/Appendix A. Needs a better introduction, perhaps providing a utility listing of the code snippets used in this piece.

Nothing on using SVG or mathml? I'm disappointed.

Summary? Good intro. Should suffice as a launch pad for an author to start using html5. I guess the assumption is that he/she is familiar with jscript and is prepared to hack with the 'browser capabilities' dance.