Yes, it's a good gesture. I wonder about the choice of syntax however. I guess we have to assume that they checked their data and picked a class name that is not already used by many webmasters. I think I would have preferred a comment tag syntax rather than a span-plus-class.

I have no problem with the meta tag specific to the company:

<meta name="google" value="notranslate">

But appropriating class names sets a very bad precedent.

I wonder what other classes Google, Yahoo, Microsoft etc will reserve for themselves.

...

I just wonder whether there's a reason they picked class. Perhaps they're looking at class values anyway in the course of returning a translation, so this setup would involve less computation.

As I remember, they preserve text and page formatting on translated pages.

>>>Perhaps they're looking at class values anyway...

Makes you wonder what else they're doing with CSS files. Time to rename class="bold" to class="xyz"? ;)

Maybe using comment tags to mark content blocks (as with the AdSense program) would be more appropriate so as to stay closer within web standards.

Here's a case where invalid markup could affect the outcome of what you wanted to do.

There will be an end tag for the

notranslate

block but are you sure which one in your code it actually is?