I presented a half day training class and a paper at CMG09 in Dallas last week. I won't be at CMG10. The conference has shrunk to a few hundred attendees, mostly mainframe oriented, a few vendors, and a bunch of independent consultants. Its main value for me was to maintain and extend my social network of capacity planning and performance people. There were some useful and interesting papers, but not enough to justify a full week long conference in a very expensive hotel. CMG10 is in Orlando FL, and I can't justify traveling to the east coast for more of the same.
A few years ago I discussed with the CMG board members what was needed to keep CMG from shrinking into irrelevancy, at the time I was at eBay, and early December is the peak business level for retail industry, so very few capacity/performance experts from retail can get away at that time of year, and I suggested they move the date. The dominant industries attending CMG have been banking, insurance and finance, which have been hit hard in the recession. The other change I advocated was that CMG should be held in the Bay Area, so that it could attract a lot more people from the major Web companies and computer hardware and software companies that are based here. Unfortunately CMG is locked into a long series of Hotel commitments for several years, and can't change its plans.
So my position now, is that I will attend CMG again when it comes to me. In the meantime, I will encourage the people I met every year at CMG to attend the Velocity Conference in San Jose next June. http://en.oreilly.com/velocity2010 - the call for papers closes in January, so we have a few weeks to come up with abstracts.