Norbert Sándor wrote: > - rethink the IOC container of t5 (use hivemind 2.0 or maybe Spring > instead of a custom "unsupported" solution) I also agree that we shouldn't have another IoC container. Spring is the de facto standard. Either take Spring and work around missing features. E.g. use naming conventions instead of namespaces or whatever until Spring adds this, or stick to Hivemind and enhance this IoC container to meet T5 needs. > - t5 should come with a compatibility layer for t4.X. Jesse "promised" > this but Howard said nothing about it. +1... At least T4 users need a migration guide like the one we used when migrating from T3 to T4. If it's a mechanical task it might be okay even if we need to trash a lot. Without a proper replacement guide however users either won't migrate to T5 or the will migrate away from Tapestry. > - the development resources should be focused first on the 4.1 branch, > because the competing development of 4.1 and 5 delays the release of > 4.1. Users of t4 are currently waiting for 4.1, not 5. > - the most important one: pay more attention to the needs of the current > users - without them tapestry would be dead... Certainly true. Don't forget that Tapestry is a Apache top-level project. That means "stability" and "maturity", too.
Tapestry should evolve to maintain its large user base. It's not yet time for another revolution! There are lot's of Tapestry applications out there - even dozends that made it from T3 release candidates to T4 final ;-) - that should be maintainable in future and we need a path to T5. No need for a revolution for T5, maybe for T6 again, but T5 should be an improvement release first. A revolution now, might lead to a community split (T4 vs. T5) or even cause Tapestry to die in the rise of JSF. The best framework won't be choosen if you can't build on it because it remains a moving target. Developing for a moving target is something difficult to explain to business people. Explaining to develop using a best-of-breed GUI framework instead of JSF & Co., because it's a, b, c, d, e, does f,g,h better is easy, if you can tell them that an even better Tx is on the way we can upgrade to, instead of waiting for the vendor-driven JSF process. > Cheers, Michael --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
