EclipseCon 2007 wrap-up (updated)

EclipseZone editor Alex Blewitt calls this year's EclipseCon "the most successful yet". He writes that there were 1,353 attendees from 447 different companies who attended 368 sessions:

"It's been a fantastically busy EclipseCon, and I've only just managed to keep up with the blog entries. I hope you've found them useful, and hopefully others will share their experiences on-line too."

More EclipseCon 2007 coverage can be found on the blogs syndicated at PlanetEclipse with some pictures at flickr.

While attendance was down a little from 2006 (1,353 vs. 1,407 (* see update below)), the Ambassador's wrap party was much larger. Last year the tab for food and drinks was about $500 (out of a budget of $300 (sic)), but this year they managed to spend over $2,800 (with a budget of $3,000)!

Looking through the presentations and blogs, one of many sessions that stood out was the one on SAFARI by Bob Fuhrer et. al.. SAFARI (not to be confused with Apple's web browser) is built on some older work by Chris Laffra (of Eclipse FAQs fame), among others. This is an over-simplification but essentially you give it some information about your new language including a grammar, and it generates IDE support for it, including a syntax highlighting editor, support for refactoring, and execution and debugging. Read more about the presentation on Doug Schaefer's blog or directly from the EclipseCon web site.

Highlight of the show: The Eclipse Ladies (not to be confused with the NetBeans Girls). 

* Update 3/12/2007: The Eclipse Foundation says the number they reported for 2006 attendance was not accurate because the original 2006 number included people that registered but then decided not to attend. The actual number, they say, is 1,353. This means that more people attended EclipseCon 2007, not less. Here's the full explanation from Bjorn Freeman-Benson, Director of Open Source Process at eclipse.org:

• The registration system allows people to register without paying right away. This allows people to register and pay by check (we wait for the check before activating the registration) or by purchase order (again, we wait for the PO before activating the registration).
• However, these "pending" registration also include registrations of people who enter data but then decide not to attend. We have no way to tell whether these registrations are really "pending a check in the mail" or "bogus".
  
• In previous years (2005 and 2006) our registration policies did not define a length of time the pending registrations could remain pending, so we kept these pending registrations in the system until after the conference.
• When we asked our registration staff for the final number for the 2006 slide, they gave us the "paid plus pending" number rather than the "paid" number. Thus Mike's final slide in 2006 had the incorrect 1,407 number.
• When I discovered the error during my post-conference reporting, we corrected the number for the auditors and the Board. I didn't blog about it because either (i) it just didn't seem like a relevant detail or (ii) I forget that we had said 1,407. I'm guessing (ii).
• These are the correct numbers as reported to and verified by the Foundation's corporate auditors:
  • 2005: 1,031
    
  • 2006: 1,291
    
  • 2007: 1,353 (well, ok, this is the number that will be reported because the 2007 audit has happened yet)
• For 2007, we changed the registration policy so that pending registrations were only allowed to stay pending for two weeks (a generous allowance for a check to arrive) - after that they were canceled. Additionally, no registrations were allowed to remain pending across price change points (very early -> early; early -> normal; normal -> on-site). As a result, the final registration query for the 2007 final slide contains only "paid" registrations and no "pending" registrations.