The MathML Handbook (Charles River Media Internet & Web Design) [Paperback]

I bought this book as a MathML beginner and was greatly
impressed by its easy intoduction to the topic. My XML
was a bit rusty, but I wasn't a total newbie. This book
has covered all my needs for MathML processing and I have
seen no need to buy any other reference on the subject. As
I have been able to do all my project on the subject without
getting lost, I must really have found a good book
My co-worker, who is much more advanced than I also found
this book very useful as a reference guide and even learned
from the chapters on the various MathML supporting
applications and dynamic web.

Aside from the content, this book reads very well. The
author is clear, concise, and doesn't clutter the book with
personal nonsense like so many other technical authors. As
the users of MathML continue to increase, this book appears
just in time for the early adapters. As a big fan of MathML,
I hope others read this book and use MathML themselves. There
is no longer an excuse to refrain from MathML on account of
a lack of hardcopy documentation.

The MathML Handbook by technical writer and expert Pavi Sandhu is a straightforward, "user friendly" guide to MathML, -- an XML-based markup language especially designed for mathematicians who publish on the World Wide Web. Individual chapters teach the basics of MathML, how to display MathML in web browsers, how to convert between TeX and MathML, how to utilize MathML for computations, and much, much more. With an accompanying CD featuring demo software for viewing and creating works with MathML, The MathML Handbook is highly recommended as being a genuinely first-rate resource.

This is a typical poorly written, poorly edited technical book that was obviously pumped out under the pressure of a deadline. There are some obvious mistakes that should have been caught before publication, e.g., where the text describing the example doesn't actually match up with the example. The author tends to introduce new markup simply by giving an example of the XML and its rendering, without explicitly saying what it does or why. This is particularly unfortunate since MathML tends to be extremely verbose, so it's hard to scan through the examples and see what's going on. There's not much useful information here about browser support. For example, I spent a lot of time trying to figure out why I couldn't get malignmark and maligngroup to work, before finally testing one of the example in the book and finding out that it didn't work; I then had to google to find out that this feature isn't actually implemented in Firefox yet. The book comes with a CD, which, as usual, is not very useful. They mainly loaded it up with a combination of open-source software and demo versions of proprietary software, all of which could have been obtained online. What they didn't bother including was a complete set of code examples, which is pretty much the only thing I would have liked. Many of the examples are quite lengthy, and it's a drag to have to type them in. The CD only includes a certain subset of examples from certain chapters. Of course if you're putting out a hacked-together book on a fast schedule, this all makes sense. You make sure it includes a CD, because that makes the book more marketable. You don't bother putting a lot of useful content on the CD, because that would be too much work.

At this point, the information in this book is getting obsolete fast. It's 7 years old. Xhtml was basically a failure, because it was never implemented properly in IE (only in a nonstandard way and with an optional plugin), and the w3c has given up on xhtml and instead decided to include all the mathml (and svg) tags as part of the html 5 standard. Microsoft has announced that it will support html 5, so it sounds like that's the implementation of mathml that will actually have a future.

So in summary, as is usually the case with books of this type, the book is not as good as the free online documentation on the topic, and by the time you buy it, it will already be obsolete.