[jsr294-modularity-eg] Simple Module System Proposal

[Stephen J. McConnell]

On Wed, 2009-09-09 at 11:06 -0400, BJ Hargrave wrote:
> > What happened to "warp speed" ? 
> 
> Sorry, had to get more dilithium crystals. :-) 
> 
> Have we (the EG) moved to consider the Simple Module System proposal
> as the way forward for JSR 294? Perhaps we can schedule an EG call
> soon?


BJ:

Back almost a month ago, August 15 to be precise, you posted a message
to this list in which you and a couple of other signatories proposed the
creation of a proposal - specifically, a proposal to propose something
to this forum. Within a few hours Paul Benedict posted a generic 'I
support the proposal to create a proposal'.  Back on August 19 I asked
for some clarification on your references to 'black holes'.  You
provided an answer the following day noting that the current proposal
had potential shortcomings in that solution delivery was implementation
dependent (and that back-hole was in fact Peter Kriens perceived pain).

Paul Benedict got closer to the issue when he the the risk of depending
on a 'reference implementation' as opposed to delivering a good quality
functional specification (a point I strongly agree with).  On the same
day I asked you for some specifics, in particular, I asked you to
reference specific documents and to construct your arguments relative to
those documents.  So far you have not responded to that request.  This
lack of response is a concern in that it could suggests that the current
documentation does not support your claim, or, that you do not have the
available bandwidth to substantiate the claim.  Either way, if you are
ready to present a proposal, then post your proposal - you don't need
permission. What you do need is a position grounded on facts (i.e.
referencing existing documents and stating issues and optionally,
potential solutions).  What you also need is commitment to follow
through. Without commitment to the process an RI becomes a whole lot
more acceptable in terms of deliverable outcomes. 

I would like to add that 294 was chartered with the problem of dealing
with the security model by coming up with low level language constructs
to support modularisation. This does not dictate how the the OpenJDK
Java compiler deals with modularisation or even how the JDK runtime
handles the same question.  In fact, I would suggest that the subject of
a 'simple module system' would be more appropriate for the Jigsaw list
(or, if you want to get all formal and stuff, a process equivalent to
JSR 277).

In the meantime I'm looking forward to seeing your grounded proposal
scoped relative to the aims and objectives of the 294 JSR.

Cheers, Steve.

 
-- 
Stephen J. McConnell
mailto:mcconnell at dpml.net
http://www.dpml.net