Re: What makes Java so special?

** Reply to note from Patrick Townsend <townsend@patownsend.com> Sun, 16 Nov 
1997 11:09:20 -0800
>   
> Hi Chris,

Hi Patrick, and thanks for writing some very good and insightful questions
about the AS/400 Java topic.

[snip]

> The JVM in the Java Toolkit lacks support for AWT. Without AWT, exactly
> how will developers write native Java applets for the AS/400 that can
> port to other platforms?

Good point. While IBM has not yet disclosed their mechanism for using the
AWT, they have confirmed that it will be implemented. The obvious method is
to use a mechanism similar to X-Windows. This would mean that Java
applications would still not run on green screen terminals, but would allow
IBM to state that the AS/400 is fully compliant with the Sun license. The
application would run at the AS/400, passing all screen and windowing
information to the workstation. 

Maybe IBM intends to have some form of Workstation controller that will
help to implement this interface, but I don't think that would be
necessary. 

The X-Window style AWT interpreter would sit on the Java workstation,
accepting whatever screen calls were sent to it from the JVM on the AS/400.
It would return to the JVM whatever screen (mouse, keyboard) events occured
at the workstation. 

> The Redbook on writing Java applications spends a lot of time describing
> the Visual Age development environment and how to write applications
> that use Data Queues, etc. on the AS/400. Since data queues are only
> supported on the AS/400 how will these Java applications really be
> platform independent? A Java application that is 100 percent Java on the
> client but not on the server is of questionable value, I would think.

Well, yes and no. 

First, there are a couple of reasons for the Redbook to spend a lot of time
on Data Queues, one being that it is a topic you are not likely to find
covered elsewhere. Another reason that Rochester would like you to pay
close attention to those areas is ... so that you'll need an AS/400!

But this is not quite the same as platform dependance (it's just close).
The application you are writing will not need to run on an AS/400, it will
just need one around to talk to. 

What you are given is a way to talk to an AS/400 in a native fashion. If
you have an AS/400 and you want to write to it's data queues, you can do
that with Java. But you don't have to. What would be a big mistake is for
IBM to NOT have a way to talk to AS/400s via data queues, as this is the
fastest way to converse with your AS/400. 

If you like data queues but are not happy with the idea that this might
someday mean that you would be tied to an AS/400, switch to MQseries, since
this is supported by IBM across many platforms and has a native Java
interface. But, it is not as fast as data queues.

> There are probably Billions of dollars invested by customers in green
> screen applications. How will IBM help customers retain the value of
> this investment via the Java implementation on the AS/400?

Java doesn't disable green screen apps. 5250 display emulation runs just
fine on Java workstations. Those apps and that investment doesn't go away.
It is possible to encapsulate older applications into beans that can be
used by newer applications. This is not a pure java implementation because
the RPG app is still AS/400 native, but it does let you use new development
tools without throwing away working code. 

Over time, though, it may become more costly to maintain those RPG pieces
than to discard and rewrite them. 

> I'm glad the AS/400 division is working with Java but I'm worried that
> it will be a half-hearted effort that forgets the incredible investment
> already made by IBM's midrange customers. If the effort results in a
> poor implementation it will turn out to be just another reason why
> developers will leave the AS/400 platform.

Patrick, from everything I have seen and heard, "half-hearted" is way off
base. You have nothing to be concerned with there. The AS/400 division, and
IBM as a whole, are taking Java _VERY_ serious. This is a big opportunity
for them. There are many, many, many billions at stake over the next decade
or so. 

If the Java effort fails, IBM will come up with a new strategy and take
another run at it. They have far too many resources to consider that any
single initiative is necessary for their long term survival. 

But for Rochester to play an important part in that future, OS/400 has to
retain some market value. Within a few years, that market value will be
greatly reduced, as NT will be maturing into an operating system with
similar reliability (hey, it could happen!). At that point, there needs to
be some reason why buying the AS/400 brand is valuable. If there isn't,
then IBM will just happily sell NT server hardware instead and Rochester
will make high end SMP servers for NT. 

Rochester has seen this. They are implementing Java at the same level as
the operating system. This isn't just a small thing. It shows how they
consider Java. This will give the JVM the highest possible speed, but still
with hardware independence. It also means a lot more work for Rochester.
When they made this choice, it showed how important they felt this was. 

> Patrick
 

Chris Rehm
Mr.AS400@ibm.net

How often can you afford to be unexpectedly out of business?
Get an AS/400.
+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to "JAVA400-L@midrange.com".
| To unsubscribe from this list send email to MAJORDOMO@midrange.com
|    and specify 'unsubscribe JAVA400-L' in the body of your message.
| Questions should be directed to the list owner/operator: david@midrange.com
+---