On Fri, Jul 20, 2012 at 11:09 AM, Kevin Wright <[email protected]> wrote:
> That would be a claim can't be easily demonstrated by citing industry
> trends.  Then again, I still contend that there is no other proposed
> mechanism which would allow deprecated methods to truly be removed, and
> method signatures to be changed (both important aspects of keeping up to
> date), all without sacrificing backwards compatibility.


I think that is the crux I have been unable to understand.  How is it
truly possible to support some of the changes that are desired while
keeping backwards compatibility?  Even with jigsaw.  Easy example, how
would jigsaw have helped move on to a better api for dates?  I can see
how it would work in some cases.  Smaller example, we have Vector and
ArrayList because the api for Vector was synchronized, right?  With
modules like this, the idea would be that you could change the api
between versions.  But doesn't that fall completely on its face when
you have communication between two modules using this datatype, where
one relies on the old api and one the new?

Or are we essentially just talking about the ability to have pseudo
statically linked applications running in a vm?  (In a cleaner way
than multiple classloaders?)

My jerk comment.  OSGi style modularity has not helped Eclipse compete
with IDEA.  Why do we think it will help everyone else?

-- 
You received this message because you are subscribed to the Google Groups "Java 
Posse" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/javaposse?hl=en.

Reply via email to