My next open source "want"

In the mid-1990's I ran engineering at ON Technology. One of our most profitable products was MeetingMaker, an enterprise-grade calendaring system. It was a much loved - and equally despised - product among companies that (for whatever reason) weren't running Microsoft Exchange but needed a "real" calendaring system to run their company. Faithful customers included Cisco, Apple, Kaiser-Permanente, Boeing, and many other large companies. From what I hear, a number of these companies still use (and hate) MeetingMaker to this day.

Today, calendaring is in a sorry state. The effective choices are:

• Microsoft Exchange
• Google Calendar, in the form of Google Apps for business
• Zimbra
• A few also-rans, including MeetingMaker, Oracle Calendar, OS X iCal server, and Chandler

The key problem is that calendaring is simply hard. Unexpectedly hard. Every engineer I know will swear that calendaring is a simple problem. Every engineer I've met that has had to work on one will knowingly tell you that calendaring may be one of the most painful things to engineer. The problems are subtle, there's a huge number of them, and there's not a lot of "glory" in building a great calendaring system. So while lots of little false-starts exist, few truly good systems exist.

The resulting state of the calendaring world is that unless you're using Exchange, the pickings are pretty slim:

• Google calendar lacks a ton of features in a "real" business calendaring system (e.g. delegation, auto-pick, granular access control, time zones, etc.; don't get me started....) While Google has a reasonble (not great) web-based UI, and made it easy to adopt, and was able to generate a nice buzz that led to some nice "sync" products around that will sync Google calendar with an iCal client, it's really insufficient for real corporate use.
• Zimbra has put the most serious effort into building a good calendaring system. They built a nice Ajax interface, and used the IETF iCalendar protocols. But deploying it requires you to become facile with the entire Zimbra server beast, which is big and complex. We've not been able to tame the beast despite many, many man hours of effort. And frankly, we may not want the rest of Zimbra.
• The world is moving _away_ from the also-rans. This company is, in fact, making a living helping people migrate off of the also-rans onto either Exchange or Zimbra.

So, my next open source "want" is for somebody to extract the calendaring portion of Zimbra, and make it into a standalone project, and tackle the remaining hard problems, and build an open source enterprise calendaring business around it. I think there's a pony in there somewhere.