First and foremost, Thunderbird is not the end all of e-mail clients. Focusing on a single client is specious at best. The fact is that many people use many different clients.(Read Outlook) If we try to force the user base to run only 'our' client they will be very unhappy and market share will erode. Instead the focus should be on making a more modular interface that accepts 'shims' to various e-mail packages. Thunderbird may be the first to be completed, but others would should be able to be easily grafted on.(at install time)
Second is the belief that a cross platform language will work. Wake up and smell the garbage. Java is a nice language, for very narrow purpose applications. Anyone here run the Java version of Word Perfect? Open Office isn't even fully cross platform. Instead, focus on modularity in the application so that the modules can be made as generic as possible. That way if you want to support another environment the changes are manageable and maintainable. You may still use Java, but do not forget to be prepared to do some OS dependant design.
Additionally, has anyone considered that the database back end should be standards driven, not open source driven? Postgresql is a great DB engine, MySql? is a great data serving engine, but focusing on only supporting one of them would be a bad idea. Instead focus on SQL89 ODBC communication. If you do that, in a cross platform way, you could use any database engine out there.
I guess what I am trying to say is that we should stop focusing on implementation and worry more about the project plan. The plan is the foundation, if it is cracked, the building will fall. Development Language, E-mail program to interface with, database back end are all implementation issues, not necessarily part of the project plan. If you specify the plan via back porting your implementation, you can expect ECCO to die. None of us want that.
So, now, where do you want ECCO to go today? (It probibly can already do some of this)
1. Handheld support 2. Outlook/Thunderbird?/etc. Support 3. Database independant 4. Full Modern Linking and Embedding support 5. Updated Interface with easier to discover controls. 6. Support for truly large files and row sets. 7. Support for Linux? and Windows? 8. Support for Terminal Server/Citrix?. 9. Some reporting capibility. 10. Standards driven interfaces. 11. Published extension interfaces for third-party developers. 12. Tablet PC support
Michael Harmer
I would even go so far as to say that Firefox is probably a fine platform to try building an Ecco-like product on top of.. Thunderbird and Sunbird (http://www.mozilla.org/projects/calendar/sunbird.html) are a prime examples of a decent cross-platform application built on top of Mozilla.
I'll also shamelessly plug Chandler as a foundation for such a product and would gladly sit down with someone and fill them in on how they could leverage the work at OSAF and contribute to making Chandler do some of the more important things Ecco does. (or if you're in the DC area next week you could come see us at PyCon?)
In fact, now that I think about it... OSAF is in San Francisco, and I if someone in SF is *seriously* interested in developing open-source Ecco-like software on either platform, I would be willing to sit down and offer lots lots of technical information on both. Just drop me a private e-mail. If not, shut your mouth!! Alec
![[Home]](http://www.compusol.org/ads/eccowiki_sm.gif)