Re: JavaWorld article on aglets.

Mark Baker (
Tue, 18 Mar 1997 14:21:33 -0500 (EST)

On Tue, 18 Mar 1997, I Find Karma wrote:
> Multithreaded, persistent, mobile, secure Java agents. IBM Japan is
> living the dream!!

More than that, they're productizing it! Aglets were submitted to
the OMG in response to the Mobile Agent Facility RFP. The details are
pretty sparse, but Aglets appears to have been chosen "winner". The
process of standardizing on agent transport (Aglets had ATP, but I
believe it was rejected) continues with input from the Aglets team, and
Crystaliz (amoung others).

> > Aglet hosts use object serialization, available in JDK 1.1 or with the
> > RMI (remote method invocation) add-on to JDK 1.0.2, to export the state
> > of an aglet object to a stream of bytes.
> And they're keeping on top of Java 1.1 ...

They've been ahead of 1.1 for almost a year. They had their own
serialization mechanism before Javasoft. Necessity is the mother
of invention, so they say.

> > The aglet development and run-time environments provide a library of
> > Java classes that support the creation and running of aglets.
> I wonder if it's any good.

Tahiti. It was extremely simplistic last time I played with it (last
summer). I haven't tried it since.

> ...but not all of the state
> The state of the execution stacks and program counters of the threads
> owned by the aglet are not serialized. Object serialization touches
> only data on the heap, not the stacks or the program counters.

There was a recent discussion on the Aglets mailing list about just
this topic. It might require some major rework of JVM implementations,
but from the developer's point of view, Thread just needs to implement
Serializeable (I think that's all that would be required).

FYI, the list is run from Unfortunately, there
doesn't appear to be an archive.

More info available at;


Mark Baker, Ottawa Ontario CANADA.                  Java, CORBA, OpenDoc, Beans,   

Will distribute business objects for food.