Feedback and comments on ARM proposal (original) (raw)
Mark Mahieu markmahieu at googlemail.com
Sat Mar 21 10:29:57 PDT 2009
- Previous message: Feedback and comments on ARM proposal
- Next message: Feedback and comments on ARM proposal
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
2009/3/21 Tim Peierls <tim at peierls.net>
I think the hard question is how much extensibility to allow up front. Magic Marker would make language libertarians happy, possibly at the risk of increased abuse of ARM. Magic Package is the conservative response; it focuses on the known problems, while preventing all but the very determined from creating their own auto types (with bootclasspath trickery).
Is there any experience from the introduction of java.lang.Iterable that might be useful to review in this context?
Iterable gave an opening for people to misuse the foreach construct, but I've seen very few examples of that happening, personally (and none that I can recall in real, live code-bases).
I suppose Magic Package doesn't rule out later on granting new magic powers
to types in java.lang.auto, including, say, the magic marker power.
--tim
Yes... I think that would be unlikely in practice though. If the set of disposable interfaces provided with Java 7 is restricted, future APIs would naturally use those even if the method names aren't ideal for their specific case.
Mark
- Previous message: Feedback and comments on ARM proposal
- Next message: Feedback and comments on ARM proposal
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]