From carson@siggraph.org  Wed Feb 12 18:19:34 1997
Received: from siggraph.cgrg.ohio-state.edu (siggraph.cgrg.ohio-state.edu [128.146.18.100]) by dkuug.dk (8.6.12/8.6.12) with ESMTP id SAA26941 for <SC24@dkuug.dk>; Wed, 12 Feb 1997 18:19:29 +0100
Received: from default (carson@siggraph.org) by siggraph.cgrg.ohio-state.edu (8.8.5/941010.52) with SMTP id MAA12147 for <SC24@dkuug.dk>; Wed, 12 Feb 1997 12:19:20 -0500 (EST)
Message-Id: <3.0.32.19970212102033.00726984@siggraph.cgrg.ohio-state.edu>
X-Sender: carson@siggraph.cgrg.ohio-state.edu
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Wed, 12 Feb 1997 10:20:46 -0700
To: SC24@dkuug.dk
From: Steve Carson <carson@siggraph.org>
Subject: SC22 Java Study Group Draft Recommendations
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"

I am continuing to monitor and participate in the activities of SC22's
Study Group on Java Technologies. Below you will find their preliminary
recommendations. Please look them over and give me your feedback. I am
particularly concerned with Point 1 which says that all Java related
activities should be done in one place in SC22. The obvious corollary to
this is that SC22 is that place, since the language work (if it goes into
JTC1 at all) will be done there. I believe that SC24 has a strong interest
in seeing any possible ISO JTC1 work in areas of Java technology such as 2D
and 3D graphics, windowing and media -- be done within SC24.

Lets have some discussion on this reflector of whether you agree of
disagree with this so I can know if I should jump in and fight with SC22
for this; whether SC24 should fight at the JTC1 level; or whether SC24 NBs
are happy to see all Java Technology standardized in SC22.

- - - 
Point 1 - All work in one group
All ISO work on the Java programming language, the Java virtual machine,
related Java class libraries and add-ons (whether part of the core language
or not), all work on JavaScript (or ECMA Script), and all work on other
programming technologies for enhanced networked applications (particularly
applications intended for the Internet or World Wide Web) should be
coordinated through a single group below the level of JTC1.

[Even though I just said it, I don't think this would work in the current
JTC1 structure. I'm not sure that a new SubCommittee would be appropriate.
I don't know what JTC1 is thinking about in their reengineering and
reorganization discussions. Such a group might be too big. I'm pretty sure
that a Working Group under SC22 is not appropriate either. I don't know of
any cross SubCommittee Working Groups.

[If all this group had initially were Java and JVM (as primary work) and
being the placeholder for ECMA's JavaScript standard, then one group in
SC22 would be appropriate and probably generally acceptable. Adding AWT,
JDBC, Beans, etc. is what would make the scope questionable within SC22.

[The other side of the problem is illustrated by a technology like ANDF
which might be a candidate for international standardization, but which has
not generated much (positive) interest in this study group. A single
committee should probably have those topics as well.

[Even though I think a single group is appropriate for coordination, I
don't know how we could make it work. But without such coordination
however, I'm afraid the standards process will become bogged down in
bureaucracy and liaisons that follow all the rules, consider alternatives,
and miss satisfying the needs of the industry and users.

[I think we need an answer along these lines, but I talk myself out of
every suggestion I come up with.

Point 2 - We're waiting on Sun to make a decision
The technology everybody wants is currently proprietary to Sun
Microsystems, Inc. and their JavaSoft subsidiary. There is no point in
trying to standardize an alternative approach. There is no point in ISO
beginning any work beyond this study group until Sun releases the
technology for the purpose of standardization. Sun may decide to work
directly with ISO or may bring the technology into another standards group
first. In either case, ISO should be prepared to work on this technology.

Point 3 - The first standard version should closely match existing
technology
There is already an extremely large base of users of Java, JavaScript, and
related technologies. There are also slight differences in implementation
and use. The initial Java-related standards should closely match existing
practice. No new features or extensions should be considered. The text
should be kept as close as possible to the existing Sun documentation. The
only variances allowed should be to resolve ambiguities and differences
between implementations (and even here this should be done in the most
conservative way possible).

This should be explicit in the Work Item description and in the management
of the working group. At our meeting in Cupertino, there was a strong
sentiment that in the first phase of standardization, the chair of the
working group should be encouraged to rule suggestions for change out of
order and the group should require a significant majority of countries
represented to override such a ruling. 

Point 4 - Java Virtual Machine and Java Programming Language first
Standardization of the Java Virtual Machine (JVM) and Java Programming
Language (usually what is meant by Java alone) is the first priority. They
should be done as separate standards, but processed together. Some classes
must be made a part of the language definition and others done in separate
standard(s). This is a matter for the standardization working group to
decide.

Java and the JVM are closely linked. Other languages might be compiled to
the JVM and other targets might be used with Java language processors, but
the two are linked in common purposes and use. Both these seemed to be
sufficiently "cooked" to be moved forward rapidly at this time.



Steve Carson
Chair, ISO/IEC JTC1/SC24
Computer Graphics and Image Processing
---------------------------------------------------------
Steve Carson                 phone:   +1-505-521-7399
GSC Associates Inc.          fax:     +1-505-521-9321
5272 Redman Road             e-mail:  carson@siggraph.org
Las Cruces, NM 88011 USA
---------------------------------------------------------
