[848] in java-interest

home help back first fref pref prev next nref lref last post

Summary of Java/HotJava BOF at SIGGRAPH'95

daemon@ATHENA.MIT.EDU (Tientien Li)
Fri Aug 11 02:19:04 1995

From: Tientien Li <li@jumpgate.tientien.com>
To: java-interest@java.sun.com, hotjava-interest@java.sun.com,
        java-porting@java.sun.com
Date: Thu, 10 Aug 1995 16:11:42 -0700 (PDT)
Cc: li@deming.Jpl.Nasa.Gov

(This is a repost. The original message run into an Internet road block at
 pilot-network-service.Bloomington.mci.net (204.70.48.62) and has been
 waiting in my /var/spool/mqueue for a day. I've to send this message from
 a none JPL host.)
 
Hi,

There were about 50 people attended SIGGRAPH'95 Java/HotJava BOF
session. Several Java team member also attended BOF including:
Kim Polese, James Gosling, Mark Scott Johnson, and several others.
Kim, the product marketing manager for Java/HotJava, not only ordered
food and drinks for us but also brought HotJava T-shirt to give away!
The BOF was basically an open floor discussion on various issues
regarding Java/HotJava. The following is a brief summary of questions
that were asked and answered. (My notes are not complete and may be
inaccurate. Please correct me, if you find errors.)

1) Better support for 2D/3D graphics in awt classes.

The awt classes have been redesigned and improved significantly in the
new beta. It will be relatively easy for anyone to develop and add GUI
objects in beta. But the API for 3D graphics is not finalized yet. Sun
has not yet decided whether to provide a set of hi-level 3D API, or
low level 3D API, or both.

2) Support for inter-applet communication.

Not sure there will be a clean design for this in beta. There are
several possible ways to get around this problem, e.g., set up shared
pipe or extend the classes to allow share variables. However, there
many different types of inter-applet communication that need to be
support. Most people believe it will much easy to design and add into
the system now than later.

3) Support for Model/View separation.

This has been added in the beta. There will be toolkit classes
allowing one to provide implementations that generates different look
and feel views for an object.
 
4) Better development/authoring tools.

There will be a command line debugger included in the beta. Also, in
the beta, there will be a WYSWYG HTML editor that support drag and drop
of HTML objects including images, form, and APPs to Web pages. SunSoft
is working on a Workshop-like integrated development environment for
Java. It will probably be available by the mid-96. There are also several
3rd parties working on various tools that will become available soon.

5) The new beta will be available in Fall. It will support HTML 2.0
and "table". There will be some minor changes made on the Java language
and a reorganization of the class tree. Language changes are mostly
minor clean up of obscure features. However, the change of "NULL" from
"TRUE" to "FALSE" may break some error checking code. Also, because of
the class tree reorganization, Alpha applets will have to check their
import statements.

6) Someone asked about the future of Java/HotJava as a standard. Both
OSF, W3C, and several other organizations have expressed interests in
Java/HotJava. However, the standardization of language and VM and the
coordination of 3rd party development issues will not be resolved until
Sun complete its beta release.

7) Win/NT, Mac, and Win/95 versions of the Alpha Java/HotJava will be
released before the beta. These versions have been working and used
internally for some time. However, there are still a few minor bugs
need to be fixed, e.g., the major hold up for the Win/95 version is
a mysterious deadlock bug that needed to be fixed.

8) There are work done in Sun to support VRML in Java/HotJava. The
parser has been ported to Java, but the basic problem is still the 
design of Java 3D API (see note #1). Once 3D graphics support are
completed, it will be easy to make HotJava VRML-capable.

9) Someone remarked that Java networking classes doesn't look like
traditional Unix socket. It was explained that the net class design
is the result of various design trade off so that it can support 
not only the Unix socket but also Winsock and MacTCP's.

10) Multi-threading support in native Solaris thread system has a
few issues that need to be workout. Whether or not to implement
time slicing in Java to support the execution of equal priority 
threads is not yet decided.

11) Sun has redo the Makefiles and help to debug compiler bugs to
facilitate various porting efforts. There are several ports including
Linux, BSDI, and OS/2 ports in progress.

12) There are also 3rd party development to target other languages
to the Java VM. Several languages were mentioned including Smalltalk,
Scheme, and Ada.

There were other issues discussed such as porting Java VM to 64-bit
architecture, # of mouse buttons, ... etc., but my notes are incomplete
on these topics. I hope others can help to provide details on these 
discussions.

The BOF stopped around 7:30. Many people stayed afterward to talk with
Kim, James, Mark, and other Sun developers. Even though I had to leave 
early to pick up my kids, I enjoyed the BOF very much. I'm especially
appreciate that many Java team members came and shared their knowledge 
with us. 

Yes, it was Fun!

--
Tientien Li
li@deming.jpl.nasa.gov
-
Note to Sun employees: this is an EXTERNAL mailing list!
Info: send 'help' to java-interest-request@java.sun.com

home help back first fref pref prev next nref lref last post