[concurrency-interest] Final specs

Doug Lea dl@cs.oswego.edu
Thu, 29 Jul 2004 08:52:59 -0400

The versions of specs etc linked from
have been submitted as the JCP "final approval ballot materials".
Almost surely, these will be exactly the ones appearing in the Sun FCS
Tiger release. (Currently planned to be late August, I think.)

JSR166 will not abrubtly terminate though:

1. We'll keep this mailing list up, and continue to welcome questions,
complaints, etc., including those about specs that we will try to get
clarified and integrated into Sun update releases.  (We already have
one that didn't make final code freeze: ConcurrentHashMap fails to
list in javadoc that it uses the same (0.75f) default load factor as
java.util Hash tables.)

2. Within a few weeks, we will release classes in a new package
"jsr166x". These are classes that we would have included had they been
ready, and that will probably make it into java.util.concurrent
sometime in the future through the RFE process. There won't be many of
these -- we'll use the same criteria of generality, performance,
reliablilty etc as in JSR166. They will include a scalable concurrent
sorted map class with a superset of functionality of
java.util.TreeMap. Later additions may include a general-purpose
concurrent linked list, and possibly a few other concurrent
Collections and custom Executors.

3. Expert group members will be writing about using the Java
concurrency utilities, possibly including a book.  (And someday, I
will get out a third edition of CPJ book, but it doen't look like it
will be very soon.)

Thanks to everyone on this list for your help in developing JSR166!