Re: "stable" platforms -- MIDRANGE-L

** Reply to note from "James W. Kilgore" <qappdsn@ibm.net> Tue, 24 Mar 1998
15:33:36 -0800

> I don't think that IBM is that adverse to the consumer market.  With
stated plans
> to derive an ever increasing portion of revenue from software, an
opportunity to
> engage in an OS war may prove very lucrative.  Even with the hardware at
a penny
> profit.  Look what it did for a company that wasn't even in the hardware
business
> <g>.

True, but IBM still sells more software than Microsoft does. IBM's
software development can center around creating a more solid operating
system and providing greater efficiency since they need not worry about
building drivers for every possible feature card to come down the pike. 

IBM did take a shot at the consumer market. Warp is part of that. But to
break into that market they will need for Windows based games to run on
their platform. Let's see, who is the biggest producer of home
entertainment software? What are the odds they will port to IBM's OS?

If market conditions change significantly, IBM may do what they did last
time, buy up a number of smaller titles and give them away free, build two
internal groups for supporting vendors, one to assist with device driver
development and the other to assist with porting code. They probably won't
pay anyone to port stuff to OS/2 again, because last time it simply proved
that a software company that doesn't need to produce to make a profit,
won't (umm, couldn't IBM have just looked at Microsoft to see that?).

IBM also put together demonstration CDs of software products and
distributed them free with OS/2.

But remember: to get into the home market IBM would have to be able to get
applications on the shelves, and OS/2 on PCs in major chains. There might
be money in the market, but not if you have to buy up all the software
developers other than Microsoft just to get code written. Also, MS has
enough control that they are choking off the little guys one at a time
anyway, so even buying them up doesn't necessarily help. 

If the software developer isn't allowed to develop for your platform, and
if the hardware vendor isn't allowed to bundle your OS, it is a little
difficult to gain in market share! Even a very strong DoJ settlement will
take years to make much of a change, so I think Microsoft's only consumer
competition is the Mac for a while. However, a strong DoJ decision could
move the Mac into a much better spot! Also, the new Mac OS is purported to
be a microkernal OS based on an improved Mach kernal. If so, it would be a
natural to carry multiple "flavors" include a Java OS (running at the same
level as Mac OS) and, if IBM swung a deal, OS/2. All at the same time. Is
there an advantage to that? Not right now.

> The point I was trying to make is that certain utility functions now
incorporated
> into an OS could be replaced by a utility purchased on the open market.
In my
> example, the "Device Manager" then could behave/look/feel totally
different from
> one shop to the next.
>   
> The same Pandora's box that would allow the application market an even
playing
> field would also be open to system level utilities.  My question is how
do we
> deal with as many device manager et. al. applications should they
proliferate
> like word processors?
>   
> The commercial market is a lot of boxes.  Currently the same ones being
sold on
> the consumer market. Just by looking around my client sites and seeing who
> actually "needs" a PC, I would venture that 1/2 could be NC's, maybe
more.  And
> if a browser IS your desktop, that's a big hit to Microsoft.  Unless it's
their
> browser.
>   
> Maybe our difference in view could be that I envision a JVM as more than
just a
> browser with the ability to serve up applications.  JAVA applications IMO
could

I left a lot of stuff in to show the gamut of what I am responding to. I
think this
is a matter of perspective here. 

The key to all this is "just a browser". Okay, pause, take a deep breath.
Try to imagine that Chris is talking about something new, not the browser
or Java he's been spouting about for the last year or so. 

The browser is really the delivery method for the network JVM, and for
applications which are not sensative to the OS you are running. What does
this mean? It means that the browser is not just the toy for clicking from
web page to web page. You configure your desktop both within (editing an
HTML page) and without (on the local os) the browser. Java applications do
indeed replace non Java apps. In some cases, a mix of HTML, Java,
JavaScript, etc. will comprise the part of the applications sent to your
desktop. All object a clicked on and refered to by URLs.

You fire up a 5250 session and connect to an old V3R1 machine to do a little
coding. The 5250 emulator is written in Java, the same ol' RPG is running
on the AS/400. Maybe some of it is encapsulated in a GUI front end also
Java based. 

This part I think we have both "seen". Via platform independant tools and
languages, we are now connecting to a world of different devices each of
them conforming to an external standard (HTML, Java, whatever). This is
consumer heaven. 

So the browser provides the JVM for use by internet applications, and also
the HTML interpreter and whatever universal functions. Certainly if it is
a Microsoft browser it provides services for functions that Microsoft
provides in products only released on Windows machines. This isn't
necessarily a bad thing, by the way. In an open market this is great.

But at the local level, the hardware interface level, Java is out of
place. Java memory handling methods are designed for ease of coding. They
are slow and not very efficient compared to C or C++. That's because those
languages are developed with the need to talk directly to hardware in
mind. 

On the one side, one of the fun things about Java is you don't have to pay
attention to pointers and addresses and pointers to addresses and
addresses of pointers (to addresses). 

On the other side, one of the bad things about Java is you can't use
pointers and addresses and pointers to addresses and addresses of pointers
(to addresses).

The average programmer (oh, yeah, I'd like to see that!) is not really
interested in playing with pointers. But you aren't going to animate much
video without them! 

Okay, maybe this will help. Java makes life simple by adding a level of
abstraction between the programmer and the data. This is how data/code
becomes objects (like the characters on your screen appear to be
characters, but they are really just points of light). Programmers use
this abstraction to form their mindset for creating programs (like you
keep thinking of these as letters instead of little glowing dots grouped
together). But there's a penalty in performance (the abstraction costs
when the things the programmer said about objects needs to be converted
into a series of calls to different routines) and in control (the
programmer doesn't know where the data is or what it really looks like
because he has always been told it is an object). When what a programmer
really wants is to spew zillions of glowing dots (maybe an animated
graphic off the web!), this data to character converter that was so handy
a minute ago can be kind of a hindrence. 

So, there is a place where Java doesn't go. This place is up to the
hardware provider to fill. The hardware provider can buy Windows, OS/2,
Unix, whatever as long as it has drivers for his hardware. The JVM is
written for the specific hardware platform. It is not written in Java
(although some functions might be implemented that way). 

Plus, as you may know, the local JVM and the network JVM have to play by
very different rules or the security of the system is in trouble. 

So, the "device manager" is the current OS. Imagine that! An OS being
relegated to the lowly task of operating the system! Barbaric. 

In effect, the OS becomes the MI layer, with Java being the application
interface layer. Of course, the OS could also still provide services to
locally running programs. 


> grow to the point that, like drivers provided by the component
manufacturer, your
> CPU/MB provider could be the issuer of a kernel and at this point the
questions
> would be "Who's Microsoft?"  I feel that this is the scenario that has
Microsoft
> doing what they can to win the browser/JAVA war.

Yup. Microsoft would still have a market, probably bigger than before. But
what scares Microsoft is they would have CONTROL of that market. They
would have to compete with anyone who came along with a JVM and provided
hardware services. 

> This question also reinforces my view that the desktop war was not "over" in
> 1996.  If the browser is the desktop, the battle is still raging.

The desktop war was and is over. The network war exists. Now, IBM would
indeed love to ship OS/2 to everyone and make billions in sales. Who
wouldn't? However, they are tired of spending hundreds of millions to
advertise, promote, port software, and otherwise push the best technology
at the market only to discover the channel cannot distribute it and the
trade mags continually give IBM a black eye by flaunting the "failure". 

IBM isn't going to enter that fray again without good cause. IBM instead
is pursuing the network computer market. They feel this has much greater
income potential and that they are better equipped to deliver to it than
anyone else. However, control of the desktop has a great influence in this
area.

The other side of this coin is that IBM has nearly as much influence with
host/centralized sites as Microsoft has on the desktop. Since the Fortune
500 would have to spend trillions to port their data to anything else, IBM
has the opportunity to wedge in some of their tools. If IBM is successful,
the JVM will be a requirement for any machine that wants to have access to
corporate data. To this end IBM spends $200,000,000 a year in Java
development including the afore mentioned 100% Pure Java port to ActiveX.

IBM is fighting an impressive fight. They are an apparently invisible 500
pound ape. Take a day or two to browse through IBM's Java sites and get an
idea of what their initiatives look like. 

Keep in mind that if Java is successful, the guys who will win market
share will be those who can deliver the best technology solutions. Do you
know who gets the most patents every year? IBM has raised the bar in every
area of technology (drives, processors, memory, etc.) and they provide the
very best solution on the planet (that would be the 9406 creatures I am so
fond of). If IBM can only connect with their Java sledge hammer, they will
no doubt squish a lot of competition like bugs.

Then, after 15 or 20 years the DoJ will wonder why having a cowlick and
wire rimmed glasses should be a capital offense and Lou Gerstner will have
some questions to answer!

> I was just kidding out porting OS/400 to the desktop.  It does fine right
where
> it's at.  Except for when my PC locks up and I'm giving it the three finger
> salute, I can't help but wish a little...

I know, but what I was trying to get at is that Java really IS porting
OS/400 to the desktop. See, if all you need is a Java workstation, then
all your code can run on the AS/400! Voila, you have a full featured
graphic AS/400!

Sorry about the length of the post. I know I went on far to long about
some stuff. I am very frustrated with this stuff because although it's
been going on for a couple of years, it does seem that very, very few of
our peers have any awareness at all of what Java, the JVM, etc. mean to us
and to the market where we work.

 

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 MIDRANGE-L@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---