I've been looking with interest at the Google Web Authoring Statistics report, having just completed an overview for IBM developerWorks of the two contenders pushing forward on HTML, WHATWG and XHTML. A snapshot of HTML tags in use tells us a lot about the current usages of HTML and the take-up of future pathways such as XHTML.
That said, I'm quite disappointed with the ways these results have been presented.
Firstly, the report as published has no author attribution. From the writing and other sources, I am given to understand it's Ian Hickson, the prime proponent behind WHATWG, but nowhere is this made plain. A date of publication would also be useful on the document.
Secondly, the graphs are given as SVG, which is laudable, but leaves those using IE or pre-1.5 Firefox browsers out in the cold. There's nothing in the data that means they couldn't be presented as PNG images. This is simply making a statement about browsers.
Thirdly, the report mixes political viewpoints about the HTML standard in with observations about the data. The report references WHATWG's HTML5 in various places without setting it in the context of the various ways forward. Hickson's views over XHTML 1.0 and the text/html media type were presented without recognition of it at least being a contentious issue, rather than a matter of fact. If authorship were attributed, this bias could be contextualised somewhat, but as it is it can only be construed as Google's endorsement of a particular viewpoint.
Don't get me wrong, I think this survey is highly valuable and congratulate Google's staff on the work done. I enjoyed reading the commentary very much, within the context I was able to place it in.
However, I think Google needs to raise their game here, by attributing authorship and separating the editorialising from the presentation of the results. The exclusion of non-SVG capable browsers from the results is also inexcusable, and rather against the general spirit of scholarship and the web.
Fascinating and enlightening as the current report is, without some basic information such as authorship it's hard to take it seriously as an addition to the body of research on the web.