On 7/19/06, Sean Schofield <[EMAIL PROTECTED]> wrote:
Ok this might be a bit drastic but if we're going to do it, lets do it before the first stable release. I propose we require JDK 1.5 for shale. My reasoning is that JDK 1.5 will be the defacto JDK standard long before JSF is the framework standard.
Much as I'd love to do this, doing so right this minute would disenfranchise a bunch of people ... *unless* it turns out that something like Retroweaver can really work to let you use annotations and generics, but still output 1.4 compatible class files. At JavaOne this was briefly discussed at the Struts BOF, and I volunteered to go look at it for Shale, but haven't gotten that far yet. I consider our work on Shale to be building for the future for the
masses - and providing cutting edge stuff for the adventurous. So if you are using Shale now you are probably OK with JDK 1.5. If we agree to this, I suggest we rename shale-tiger to shale-annotations which is more descriptive.
If we went for a JDK 1.5 base, wouldn't it make more sense to fold the tiger extensions into the core? Anyways, I'm sure there are some good reasons for keeping it JDK 1.4
but it might not hurt to revist that reasoning now that we're our own TLP and we're approaching the first release.
The reason that matters to me is I'd bet 95% of our potential target audience can't use JDK 1.5 yet. That number will change very rapidly ... but not instantaneously. Sean
Craig
