[Isolate-interest] JSR-121 aggregates and links (Strawman 9)

[Godmar Back]

ExitReason and Type don't have to exist as inner classes either and
could be replaced with integers if our goal is to accommodate J2ME by
bringing the number of classes to a minimum.  Going a step further,
CompositeMessage and DataMessage could be one class with a flag to
distinguish the two.  An IsolateStartException could be an exception
with a magic msg string.

Is the footprint issue intrinsic to the definition of J2ME, or is it
an artifact of a particular implementation?  It seems to be a weird
situation where all JSRs that include J2ME would have to minimize the
number of classes to accommodate J2ME.  I'd rather reduce the
compliance requirements for J2ME implementors so as to allow them to
find and use more efficient forms of representation that don't impose
high per-class overhead (?).

Has anybody done (or is doing) a JSR-121 prototype for J2ME?

 - Godmar

(*) See also http://www.spellingsociety.org/media/spoofs.html
----- Original Message -----
From: Laurent Daynes <laurent.daynes@sun.com>
Date: Wed, 29 Sep 2004 18:56:10 +0200
Subject: Re: [Isolate-interest] JSR-121 aggregates and links (Strawman 9)
To: gback@cs.vt.edu, isolate-interest@altair.cs.oswego.edu



 
StatusMessage.getExitStatus() returns the integer corresponding to an
isolate exiting. Ah, I overlooked that. I would use different Java
types for different StatusMessages (STARTING, etc.) because it allows
better OO programming, such as the use of visitors. It also gets rid
of your need to define what getExitStatus() should return if getType()
returns STARTING.However nice & OO  this is, it is problematic to the
J2ME footprint-concious implementor.
When talking to some of them last year, they got really nervous about
adding even just a single class, especially when
this one has little content (e.g., one or two methods). That really
bloats the footprint for not  much
in terms of functionalities. 

Laurent


 - Godmar 
The EG is split between various forms of isolate death causing or not
causing death of other isolates or complete destruction of the
aggregate, so we decided not to dictate this. -Pete
_______________________________________________ Isolate-interest
mailing list Isolate-interest@altair.cs.oswego.edu
http://altair.cs.oswego.edu/mailman/listinfo/isolate-interest

-- Laurent Daynès laurent.daynes@sun.com Sun Labs,180 av. de
l'Europe,ZIRST de Montbonnot,38334 Saint Ismier,FRANCE Phone : +33 4
76 18 88 18 Fax : +33 4 76 18 88 88