[Isolate-interest] deafening silence about refactored APIs

[Pete Soper]

"Curt Cox" <ccox@tripos.com> said:
> > Yes, absolutely for CLDC, assuming we're already way too large and at
> > least one J2ME authority has judged our base package way too large.
> 
> If you're already way too large, one more class won't hurt. ;)

Yes, since it may not be accepted into J2ME as it is, no worries. :-)

> Serious though, how much space does one class take and how big

It varies

> is the CLDC javax.isolate.* space budget?  Is the size of the

A few kb total. I haven't looked at the numbers for many months and
can't bring them to mind at the moment, sorry.

> class file, or the increased memory footprint the issue?

Yes. :-) 

>  
> I'm a fairly heavy user of "enums" under 1.4.  The class files tend
> to be less than 2K.  How much is too much--1000 bytes?  100 bytes?
> 
> Of course you can usually hand hack a class to be smaller than the
> one that the compiler produces, but that is pointless under J2SE.
> 
> Hello World! in 70 Bytes 
> http://www.sys-con.com/story/?storyid=37060&DE=1
> 
> Given things like JSR-200 and the research it builds on, I would be
> surprised if a single class adds much space when properly compressed.
> 
> JSR 200: Network Transfer Format for JavaTM Archives
> http://www.jcp.org/en/jsr/detail?id=200
> 
> The priority should be making the Isolate API as simple and clear
> as possible.  If we have a working implementation that is just a
> little to big to suit CLDC, then that would be a good time to
> figure out how to save some bytes.

We might be off by a factor of several X, not just a little bit, it's hard 
to judge. This is what my babbling some months back was about: the big
tension of a clear/typesafe design vs small enough for the base package
on a highly constrained device. 

Regards,
Pete