So far I have no idea if anyone’s thought it worth thinking about; but if it was put into place, the de jure file format standard would become the de facto standard overnight.

FWIW. ;)

When, as an American, I even think of a date, I mentally do it as April 1st, 2000. Month, day, year. Any date, historical, personal, whatever, is mentally stored and retrieved in that order. Same with speech. I wonder if it is different in Europe? Do people process dates mentally in a different order? Or is it just a presentation difference when writing?

Wraith, I agree with you that OOXML was approved in December, not July. I was using American date conventions (12/7/2006), not European (7/12/2006), thus the confusion. Having two standards for the same thing only leads to confusion, as you’ve demonstrated.

Wouldn’t the use of the sole ISO 8601 international date standard (2006-12-07) fix the problem once and for all?

I cannot imagine this could result from an attempt by Microsoft to create a feeling of exponential growth.

If I download any such source tree from IBM or Sun, for example, the one thing I am not going to escape, is the license terms, or the repudiation thereof.

I get the feeling that this is more of a puppet show than a real attempt to engage anybody – still, I’ve offered them the opportunity to show their commitment to interoperability by being ready to accept bug reports from me when I try compiling it with Mono. I’m not sanguine about it.

It seems to me that since every copy of Office 2007 is probably pinging Redmond every day, with all this Genuine Disadvantage, registration and auto update stuff, so if they wanted Microsoft could tell us an exact number of running Vista or Office 2007 machines. Surely they know the exact number of activated Vista and Office 2007 users. And surely if it was good news they would have told us by now.

In any case, that shows you the interest level in a format, not necessarily the deployment or use of a format.

It would be interesting to hear how Google determines document types. By content type headers in the HTTP response? Or inferring from file extensions? Certainly, depending on how this is done, a misconfigured web server could cause this to break. There may be other web sites hosting ODF documents that are similarly misconfigured.

I believe there is a flaw in your data. It seems that the Google filetype:nnn query does not report Open XML counters correctly.

Consider the following queries:

Scope, results
[filetype:docx site:microsoft.com], 75 documents
[filetype:docx site:blogs.msdn.com], 0 documents

Yet Doug Mahugh has posted DOCX documents on his blog (blogs.msdn.com/dmahugh). Example:
Original post: http://blogs.msdn.com/dmahugh/archive/2007/03/26/custom-xml-markup.aspx
DOCX hyperlink: http://blogs.msdn.com/dmahugh/attachment/1950872.ashx

Did Google filetype:nnn not see those documents?

Further analysis reveals that Google filetype:odt does recognize the ODF types, but filetype:docx is reported as “Format unrecognized”. The preview text displays the names of parts inside the ZIP file and not the body of the document.

All the analysis in the world means nothing if you’re data is flawed.

I wonder if you have any thoughts about whether any deeper meaning can be read into measuring seach volume.

Ben, thanks for the details on where the docs are coming from. I didn’t think of that.

Hmmm… anyone remember the “Internet Tidal Wave” memo? In that document Gates lamented that, “Browsing the web, you find almost no Microsoft file formats”.

I’ll take a look at a log-log chart as well, although I admit not being a fan of them. In a log-log chart, almost any data can be coaxed into straight line, thus its power and thus the opportunity for misuse. Better in my mind to fit the data to models and check for goodness-of-fit.

So what growth curve would you expect? With a constant number of users producing documents at a steady rate you would expect linear growth in document count over time. Suppose the user base is also growing with time. Then the document count would be a time integral of the user growth function, right? If user growth is linear, then document growth is quadratic, etc.

But we don’t have enough data points to really distinguish growth models at this point, so I’ll continue to track the data and post on this again when I can make more conclusions.

In the long run, the reasons why OOXML documents are not appearing may or may not be important. A great deal of the “network effect” has to do with perception. If people see and use ODF documents on the web, especially with plugins from Firefox that let them use those without any other software, the perception could arise that ODF is “important”, and OOXML is not. If I were Microsoft, I’d be working like crazy to get allies to post any documents in both binary and OOXML formats, just to fuel that perception, but they have not been very successful at that as of yet.

OOXML is not that format, and it probably won’t ever be. ODF might.

By analogy, who would now want a web page that only worked in IE7?

MS may have some hopes of maintaining their grip in the office, but the web is lost to them, IMO.