I'd like to throw this up again. What are peoples thoughts on the following
1. Make Ant 1.6.x the last JDK 1.1 release. This would be clearly documented
+1
2. Make the subsequent release require JDK 1.2+ (what about leap frogging to later versions?)
I don't know the stats on which JDKs are still in use, so it is a little hard to talk sensibly about this. We don't know how much pain we are inflicting on people by insisting they upgrade in order to use the latest Ant. But here are my suspicions:
JDK 1.4 - This is wildly different from 1.3, so people are being cautious about upgrading to it. We would leave many people in the dust if we adopted this one as the standard.
JDK 1.3 - Since this is so similar to 1.2, I think most people have moved to it already. But there will still be a significant number of people who haven't bothered upgrading yet.
JDK 1.2 - I would guess that there are still a number of people, particularly those running older software that don't want to pay for upgrades, who are still using this JDK.
JDK 1.1 - apart from people running applets in IE, I doubt that this is much in use any more. We lose the large installed base from Windows, but that has never really been our "market" anyway.
Those are the costs of each. Then there are the benefits. Clearly going to JDK 1.2 is such a win over JDK 1.1 that doing so is a no brainer. JDK 1.4 would also give us much more functionality. JDK 1.3 over JDK 1.2, however, doesn't seem to do much for us.
So in considering moving up from each previous JDK, we have:
Cost Benefit JDK 1.2 Low High JDK 1.3 Medium Low JDK 1.4 High High
So going to JDK 1.2 is a good tradeoff, JDK 1.3 not such a good one, and JDK 1.4 is too expensive. I'd say stick with JDK 1.2.
3. Name this subsequent release Ant 2.0 (due to its change in system requirements)
Absolutely.
4. Drop all the Ant2 cruft from the website.
Just the website? What about the cruft in the code?
I had thought that there was general agreement that once we went to Ant 2, we could finally drop some of the code kept in there for hysterical raisins. We could break some of the assumptions of custom task writers, provided the impact was minimal and well justified. And we could get rid of deprecated methods, classes, and tasks.
