[Isolate-interest] Infoworld "news article" and more brain storming

[Pete Soper]

"Filip Selendic" <fselendic@rest-art.hr> said:
> 
> Pete, does this guy know something you don't?
> 
> 
> "Java to gain multitasking improvements"
> Virtual machine to be fitted with application sharing
> http://www.infoworld.com/article/04/04/06/HNjava16_1.html
> 
> "J2SE 1.6, which is scheduled for a beta release this fall with general
> availability eyed for 2005, is slated to feature an "isolates" function to
> boost application-sharing in the JVM, according to Murali Kaundinya, senior
> Java architect for Sun Software Services.  The isolates capability would
> enable isolated entities to perform localized computation without requiring
> a second JVM, thus enabling multitasking within the JVM, according to
> Kaundinya."
> 
> Fall 2005? Am I missing something?

Well, this is just my take on this article, as I had not seen it until I got
to work this morning and it started making the rounds.

Traditionally (I think universally) there has been one or more dot release
between each major release of J2SE (i.e. if Sun follows tradition there will 
be at least 1.5.1 coming before 1.6). Many of you have some idea of the
magnitude of effort that goes into the dot releases, let alone the major
releases like Tiger. And it runs counter to common sense to expect Sun to go
beta with one major J2SE release during the deployment phase of the previous
one! 

I don't know Mr. Kaundinya but can share the fact that he isn't part of the 
organization that creates the J2SE releases. It's the technical and management 
leadership of that organization that I'm paying attention to at the moment and
I'm not aware of anything at all to report that relates to the above "news."

If/when something official about JSR-121 comes along it will be from the EG 
via me or another member and it will be easy to understand and relate to the
real world.  If Sun has some announcement about it's future product plans that 
relates to JSR-121 then Sun will share that when it's closer to a promise than
a speculation.

Some of you will of course be thinking "Right, you've made a bunch of 
official pronouncements: all worthless!" The message I'm trying to send
is that if/when JSR-121 is able to enjoy more focus and we can open out the
process of adding real usage and the feedback that brings with it, there will
be a difference that will be impossible for the highly intelligent readers of
this list to miss. 

---
The more I think through next steps that I hope might happen to
JSR-121, the more a very simple RI that can serve as a prototype until the 
right API design can be hammered out for SE and the two ME flavors would seem 
to be of great value. We would need to run these within the constraints
of the various bootclass path API sets and I think that makes the JanosVM
code problematic (but maybe not, I don't really know). 

The "almost Tiger RI" code mentioned last week is heavily dependent on 
java.nio, which is also problematic for a "CDC only" or "CLDC only" 
environment (and when I say "CDC" and "CLDC" I'm corrupting and
generalizing these labels: CDC/FP and CLDC/MIDP seem to be the proper
candidates). On the other hand it would seem reasonable to start hammering on
the base javax.isolate package first and the current RI could may be quite
easily adapted to this.  But apart from this being a continuation of my
pure brainstorming from last Friday, my point is that It is the case that 
more "enthusiastic" implementations of isolates, by virtue of being very 
intimately connected to JVM plumbing, may simply not be agile enough to 
support the kind of experimentation we need to do to add real usage to the 
wealth of "thought investment" made by the EG to date. Again, to be crystal
clear, this is me dreaming through my keyboard, whether this compatible with
various practical constraints is a complete unknown.

Regards,
Pete