Re: "stable" platforms -- MIDRANGE-L

** Reply to note from "James W. Kilgore" <qappdsn@ibm.net> Sun, 22 Mar 1998
23:00:30 -0800

> Well I hope they have better luck at actually doing something about
opening up consumer choice
> than the judge who was merely removed.  And I lack confidence that
anything of substance will
> result.  A fine only results in Uncle Sam skimming some profits.

Well, there are a handful of things that would even the playing field. A
fine would be a good idea if it were strong enough to affect the stock
price, but it won't really control anything at all in the long run.

> But at this point in time, what other OS with the variety of apps would
the consumer actually
> choose?  Would a company like IBM make a push for OS/2 V4 bundled w/
Lotus Smart Suite (voice
> controlled) to become a "consumer" product?  Would they use this
opportunity and push their
> 1000Mhz CPU and 10Gb/sq in. disk technology to market sooner giving a
double whammy to the
> "Wintel"?  Would the Mac take off like a rocket?

I think you're losing sight of exactly what this means to you and I. Sure,
this is important as a consumer product, but that is last year's
battlefield. Don't spend your time looking back or you'll be looking back
on what happened to the AS/400 world.

Certainly many here don't feel that network computing is going to impact
us, but IBM, Microsoft, and every other major player in the computing
industry does. 

IBM won't be making a play for the consumer market. There's no money in it
unless they get into consumer software and IBM isn't likely to do that.

But let's suppose that we do evolve to a network computing world. Let's
suppose that applications are developed which access data on more than one
source machine to comprise a single display. 

For a moment, imagine that your browser has evolved into your desktop:

You don't know where the applications or files are actually stored, you
just click on icons and run the things necessary to do your job. 

When your word processor is upgraded, the object on the application server
is updated to the new object. The next time anyone in the network clicks
on that icon, they are using the new object. It doesn't matter what client
is being used, Windows, OS/2, X-Windows, etc., they all run a browser
which interprets the application. All data files might be saved to the
same server, or not, depending on the wishes of the administrator that
maintains the word processing for the company.

Okay, now you click on the payroll app icon. The department you are
responsible for opens to you (controlled by security administered at the
data server by the data base dependant on your logon). You remember you
need to update the rates of those employees whose pay is based on a ratio
to the minimum wage, so you pop up an SQL generator (Access, FoxPro,
Approach, whatever) and pop together a couple selections. You see the
results and click on "commit" and they are updated. You key the selection
parameters and launch the pay run. You don't care if the checks are
generated in your office or in Japan. They are redirected to the printer
you select and you update the check numbers once they print. 

Okay, now let's suppose that the browser you are using is a 100% Pure Java
compliant front end. Every word processor vendor has the same chance to
provide you with an app. They have to provide the best value to compete
with the others in their class. None of them can rely on the Java spec
including some special API or function that they can exploit to provide a
feature the others don't know about. They all have the same shot at
reliability. Bundling doesn't matter any more, since the application will
be loaded into the network, not on the workstation (which could just as
easily be a Java workstation as an NT PC).

Also, when looking for servers to provide the data and service for the
payroll app and functions, you can look for the most cost effective
hardware that is compliant with the industry spec. You don't have an
AS/400 app, or an NT app, or a Unix app, so you can buy whatever back end
machine you want to serve it. 

Now, let's suppose the browser is not compliant. Let's suppose it is a
proprietary browser which implements functions specific to the platform it
runs on. Now, would application vendors have the same level playing field
to deploy? Bundling would still be a factor, since competition would need
to find a way to deploy pieces which allow them to provide the same
services. Also, companies would have to standardize on a single
computer/os if they wish to achieve the promise network computing offers. 

How about your payroll app? If the industry standard front end is provided
with proprietary hooks to, say, NT, would AS/400 or Unix applications have
a level playing field? Think of it this way, what if Microsoft owned the
rights to the PROM in all the green screen terminals installed in
businesses. Let's suppose they wanted people to buy a certain class of
host. Do you suppose they would have the ability to manipulate that
without improving the quality of the host in question? 

Now let's take a wake up call. If you were asked last week what is the
number one hardware competitor to the AS/400 in the midrange market, what
would you have answered? My answer would be the HP.

So, what did it mean to me when HP agreed to build and license out a non
standard Java Virtual Machine with MicroSoft? It means to me that HP sees
that install NT on the 3000 and the 9000 and hooking them to prorietary
front ends gives them a shot to displace AS/400s. They don't give a damn
about the consumer market, this isn't about what people play games on at
home. 

>   
> It is my understanding that the big threat in the browser war is that the
browser is just the
> tip of the Java iceberg.  With a mature Java bundled with a SQL compliant
DBMS and 100% Java
> applications, the OS becomes more of a bundling of utilites similar to
the arrangement with
> Linux.  The up side is that it would beg to question where M$ would play
a role, the down side
> is that at every location you went to the "Device Manager" may behave
totally different.

I am not exactly sure what picture you are painting. Java allows
applications to be written independant of what platform they are running
on. The Java workstation would have some very simple built in
configuration to handle the display and keyboard. PCs running various OSs
would use the same configurations they use now, but one of the
applications they are running would be a Java Virtual Machine that would
interpret the application shipped to it from the network and make requests
of the local OS. Nothing new here.

SQL compliant applications? You mean SQL compliant data bases? They
wouldn't necessarily want the data base on the same machine running the
application. But you might.

> How would this effect us trying to tie desktop machines to the AS/400?
Would the beauty of a
> multi layer OS approach that the AS/400 uses to separate application from
low level services
> from hardware services become the "standard" and who would decide on such
a standard?  How
> long will it take them to make up their mind?  Would anarchy set in? For
example, Netscape got
> tired of waiting for the standards organizations to approve a function so
they just put it out
> there and essentially created the "standard".  Good free market ability.
But now a published
> web page must be frame/text selectable.  How many of these "selections"
do you want to have to
> embed into your applications?

Again, this is a very important part of the Java implementation. Java
implements graphics via the AWT (and other tools but I am just trying to
indicate how this works). When an application wants to draw a window and
stick stuff in it, it sends a set of AWT defined codes out. 

Okay, so if the application is running on the AS/400 and it realizes it
needs to draw a pop up window to display a list of files it does NOT try
and talk to a graphics card. Instead, it sends a stream to the JVM which
then forwards the data. Where? to the Java workstation you are sitting at
which interprets the stream and writes the window. The same way your
current app sends a data stream to a workstation controller and it is
passed to a 5250 display which interprets and draws it. 

Effectively, Java is the display. This ties the workstation to the AS/400
the same way 5250s are tied to it now. The difference is that the Java
workstation can be running apps from a number of different sources.

Now, let's suppose Microsoft controls the display. What OS will the
display require? Since MS has done such things before, isn't it likely to
assume that they will implement controls which give NT servers an
advantage over the AS/400?

> In my worst case scenerio vision, the technologies would fragment for a
period of time which
> would not help the AS/400 or ourselves progessionally in the short term.
TCP/IP and Telnet
> are lacking services so we would have to take a step backwards in order
to take two steps
> forward. Timing is everything and with the world facing Y2K issues I'm
not sure I want ot see
> the desktop going through a major shift right now.  The last time M$ was
in court would have
> been a better time.  A lot of dust would have settled by now.
>   
> I'm afraid that we'll be living with Win9x for some time to come.  And
changes will start to
> occur just about the time CA/400 makes peace with Windows :-)

I think that you are showing a little tunnel vision in thinking that the
battle is over the desktop. That battle was over in the beginning of '96.
The current battle is for the enterprise. 

Now, maybe CA is buggy and I think there is every chance that is the
problem. I think it's useful to recall that when Microsoft wanted to take
over the office productivity application market, Lotus, WordPerfect,
Borland, etc. had a problem with the fact that the published APIs they
were using were buggy while Word and Excel made used of unpublished,
undocumented APIs which weren't. I am not a conspiracy theorist. I do
think that Microsoft has learned that it doesn't matter what they do,
their customers will love them. When it became public they had doctored
APIs, even when they stole compression technology from Stac, nobody cared.

Then, when the DoJ took them back to court this time and disclosed that
Microsoft had not even kept the weak agreement made last time, I could not
see any reason to believe that Microsoft would pass on the chance to erode
the IBM's market share. After all, what would MS lose from this? The worst
outcome they could expect to see is that after years of users feeling that
using green screens is unreliable (because you can't tell when 5250
sessions will crash your PC), MS _might_ be told to stop it. Probably,
they would not be told that, since their competitor is IBM.

> Now if we could port OS/400 to a desktop.....

Please try and understand how frustrated I get when I see this statement.
It isn't just you, it is the vast majority of the AS/400 community which
has failed to pay attention to what Java means to us. IBM has been trying
to give you a graphic workstation which makes the applications served by
the AS/400 (like Notes, web pages, Word Pro, etc.) a part of your desktop.
But instead of identifying with the direction IBM is taking and why, we
see comments about "language of the month" and "Gee, just like SAA." 

This myopia is what allows MS to evolve the market. If people looked two
moves deep they might see where they are losing.

 

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
+---