cutives with Anixter. So I would like to formally, on behalf of I-B-M, 
thank you very much for putting on this evening's event and arranging 
for such a wonderful facility and for food and for the time here. I 
thank you for spending your own time, so let's give Anixter a round of 
applause.



My job at I-B-M has changed a bit over the years, of being in develop
ment positions and marketing positions. And I'm somewhat in a general 
management position where my job is really to spend time with our cus
tomers and with people in the industry. Whether they're business 
partners or they're I-S-V's or O-E-M's. Or they're consultants or secu
rity analysts on Wall Street or in Frankfort. Or business partners who 
are going to help us sell the products. And I spend a lot of my time 
talking, but I hope that I am spending more and more of my time listen
ing. So I'd like to have this session this evening be interactive. Now 
I'm going to cover a variety of topics and what I would suggest we do 
is, I'd like you to ask questions or make comments or yell or scream or 
flame. Don't worry I wear asbestos underwear. I've been on the net for 
enough years that I know that. It's important to wear that at certain 
occasions. But what I would like to do is I like to hear your comments, 
because what I hear on the Internet and on CompuServe and A-O-L and 
elsewhere that I visit, and when I visit customers, are a lot of com
ments of what we could do at I-B-M to improve our service and support. 
What we can do to improve our marketing and sales, and there are some 
things we can do, believe it or not. I get a lot of comments on adver
tising. And I'd like to be able to address those, because I hate to type 
out the whole thing and I want to give you an answer on that. Or product 
development, and I've been involved with product development for years. 
At heart I'm a nerd. This is probably one of the best jobs in the world 
that I could possibly have. Imagine this. My hobby is computers. Per
sonal computers. Before I joined I-B-M, I started a bu



siness with a couple of guys in college and graduate school and we were 
in electrical engineering, three electrical engineers. And Intel in the 
early nineteen seventies came out with this product called the four 
thousand four, a microprocessor. The first commercially available on. We 
had this brainstorm that we thought we could take on I-B-M and build 
computers with this thing. Because we knew I-B-M was taking seven years 
to manufacture and we said, we could assemble that quicker. Well, it 
took us actually only about two and a half to three months to design and 
manufacture the first one. And we sold it. It shocked us. We had no 
marketing background, no financial background, we funded this out of our 
beer money, basically. I did go to Michigan Tech and that was a second 
major for almost all of the students at that school, all 4400 of them at 
the time. There are over half engineering students so you can imagine. 
And there were only 400 women at the school too. So, there were 4000 
guys who didn't have any chance, so beer was the only alternative.



What we started is we saw this opportunity, to get into P-C's and we 
made some really nice P-C's. Unfortunately we knew nothing about mar
keting or finance. We ended up on a Friday night taking a look at our 
wallets and between the three of us had six dollars. And in those days, 
that was 20 some years ago, there was a brewer across the river from 
where we lived, where you could buy a keg of beer for fifteen dollars. 
So the challenge that we had was find, there were three of us guys, we 
wanted to find three women who had three dollars each. Together, we'd 
have fifteen dollars for the weekend and we could get through it. So 
that's the start of me into the computer industry. But I love P-C's and 
now, imagine this. I work for a company who gives me P-C's to deal with. 
A neat little butterfly think pad and one for my office and another one 
that I've got. I get all kinds of free software. And I get paid for 
doing what I really like to do. So it is an honor to join you and thank 



you for being a user of our particular products.



Now I want to conduct one other survey. Would each of you stand up for 
those of you who use O-S 2.Okay, hopefully, there's a couple sitting 
down, but that's fine. We'll convert you by the end of the evening, 
hopefully. If you'd stay standing, please, for a second. For those of 
you who have used O-S 2 for less than a year, if you would sit down, 
please. Okay, some people in here. If you've used it less than three 
years, would you sit down please. Thinning out, but not much. Less than 
five years, this has got to get a crew. Okay. This is going to get you, 
less than six. Less than seven. Less than eight. This will take you back 
to 1988. I am honored. Thank you all for using it. Less than ten years. 
Okay, if anyone was to do that, then we pick out the liars in the audi
ence. Because it wasn't out then. Came out in December of 1987, if you 
remember. O-S 2 standard edition, version 1.0. No graphical interface. 
Was a character interface on the product.



Hopefully you agree we've come a long way since then. Now what I'm going 
to talk about, let me outline it, so that we can talk about subjects. 
First thing I'd like to do is just talk about I-B-M in general. And some 
of the big changes that we've gone through at I-B-M since I-B-M's new 
chairman, Lou Gersner, came on board a little over three years ago. And 
I want to be able to talk about some of his primary initiatives within 
the corporation, because it establishes a context for what we're doing 
with O-S 2 warp. And I think it's important to understand that context. 
Now from that level, what we developed is a strategy, a vision. A 
strategy. Now at one point he was quoted as, we don't have a need for a 
vision. And I'll explain why he said that. But we did establish one. And 
I'd like to describe briefly what that vision is, of where we see the 
future of computing going. Because that certainly affects what we're 
doing as far as investments in O-S 2. I'd then like to talk about the 
O-S 2 warp family of products. Where we're go



ing with that in the relatively near future, within weeks and the coming 
months. I'd also like to talk about what's happening in the industry 
with I-S-V's and O-E-M's. And what I-B-M is doing with marketing, 
including merchandising, channel development, advertising. And what 
market segments we're targeting. And I'd like to talk about the service 
and support and of other things. Now hopefully that will address almost 
all the topics. And I'll talk about Win 32 and I'll talk, if you want me 
to, I'll be happy to talk about competition also. But I don't intend to 
badmouth them, there's no need to. Now, I didn't say who I was talking 
about and those you chuckling are just guessing.



When Lou Gersner came on board at I-B-M three years ago. Prior to that 
year who knows how much money we earned the I-B-M Corporation. About a 
minus eight billion dollars. Now to the best of my recollection, no 
company in any industry has ever lost that much money. Now that's not 
something you hold up with honor. Now fortunately, not all of that was a 
loss. The vast majority of that was actually a write off of assets. And 
also setting aside money because we recognized that we had to shrink 
from, at one point we had 406,000 employees. We set a goal of shrinking 
down to about 215,000 and we set aside a lot of that money to offer 
financial incentives for people to leave I-B-M. Because we needed a 
smaller number of employees and a different mix in order to be profit
able. The first thing we had to do is, we had to re-establish the I-B-M 
brand image. And one of the things you have establish is that I-B-M is 
going to be there when you need it. And if you're not a financially 
secure company, a lot of people are going to question, should I buy this 
product, because they might be out of business soon. And if you lose 
eight billion dollars a year, it's hard to stay in business very long.



Fortunately we've turned that around. And that's what Lou Gersner was 
focusing on first. He wanted to turn I-B-M back from losing money to 
making money. And to being



 a financially stable company that our customers could count on and see 
that we're going to be around long term financially. And he's pretty 
much achieved that objective. Our stock has basically tripled, except 
for today it dropped down. We announced increased earnings, increased 
revenue, and the stock goes down. I understand computer technology but I 
still haven't figured out Wall Street yet. But financially, we're doing 
very well.



The second thing, and that is, once we've stabilized the company and it 
is strong financially, the second focus that Lou has had is, where are 
we going to invest for growth in the future. Because what helps get a 
better return for our stockholders. Because we feel our stock is well 
underpriced of what it should be. So we've got to be able to create 
value for our stockholders. What we have to do is invest in areas that 
are growing because part of the perception, we think, that's out there 
is that a lot of people see I-B-M and said, well, they've got some big 
technology but it might be dinosaur stuff. Main frames are going to die, 
some people say. Well we wish that would happen more frequently because 
we can't manufacture enough of them to meet demand generally. And our 
main frame business is booming. The new E-S 9000, the C mas versions, 
the sys plexes, are really selling well. And the new O-S 390 operating 
system that runs with it is going well. So there's definitely a future 
there. But there is still a perception in the minds of people that 
that's not the growth industry of the future. And we agree that there is 
other growth industries that are just booming. So our strategy that we 
go into is continue to invest where we're doing well. We're not going to 
abandon that business. But invest in some growth areas. And that's what 
I want to lead in then to our strategy.



When Gersner came on board, what he did is basically said for the 150, 
180 days, I'm not going to do any work. I'm going to sit here and lis
ten. I'm going to listen to our customers. I'm going to listen to people 
who are 



not our customers and find out why. I'm going to listen to business 
partners. To competitors. To analysts. To the press. To consultants and 
hear what they have to say.



One of the clear messages that came across was that they said, I-B-M, 
you guys are building noble award winning technology. But it doesn't 
work together. And we, as your customers or potential customers, we're 
not in the I-T business. Our primary business is to manufacture goods or 
deliver health care services or financial services or insurance poli
cies. That's the business we're in. The greatest value you, I-B-M, can 
provide to us is help us build solutions. So what happened is, we 
became, as our top priority, a solution company. Where we want to align 
ourselves by industry with the companies that operate in those indus
tries, to understand their wants and needs. And then help them build 
solutions. Regardless of who the technology provider is. Whether it's 
I-B-M's product division or one of our competitors.



Well, that then placed a requirement on our product divisions. What it 
said is that, I-B-M if you're going to build technology, it better work 
together. Well how many of you in the past have tried to get various 
I-B-M products to work together. How many of you were successful. A lot 
fewer hands. There's still a couple up. Those are the ones who said they 
were using O-S 2 for ten years also. The thing is, the technology was 
hard to put together. We had inconsistencies in operating systems, the 
databases were different across the products, the communication protocol 
supported on different systems varied. So it made it very complex. Even 
amongst our own heterogeneous set of systems, then you bring in H-P and 
Sun and Microsoft and Oracle and try to get that to work together. And 
they just folded their arms like this and said, no, that's not the way 
to go. So what that precipitated for us, it said for our product and 
technology divisions, we need to take the lego block approach. This 
company in Denmark, builds little plastic blocks, if you don't h



ave kids. Snap them together, they're all built to the same standards. 
They snap together and you build solutions with it. Well our analogy to 
that is to build our new technologies on industry standards where they 
exist. So when we build object technology, build it on industry the 
standard. When we build networking technology, build it on industry the 
standard. When we build compiler technology, build it for industry 
standards. And that is a major shift for I-B-M. We still haven't over
come the perception, the perception that we rightly so deserved and 
developed over the last several decades is, I-B-M was a very proprietary 
company. Sure we build standards, it was the I-B-M standard. It's a very 
open communication protocol that everyone adopted called S and A. The 
industry standard database programming interface, the D-B two program
ming interface, etcetera. And it was all based on I-B-M... (END OF SIDE 
1, TAPE 1)



... So that's the shift. So we've directed all of our product divisions 
to build their technologies on standards. On dejour standards or defacto 
standards. Because there's something like T-C-P-I-P which it will evolve 
to become a defacto standard. We're building that into all of our prod
ucts, for example.



So with that as a principle, being a solution provider and build tech
nology components that can be snapped together quickly, based on 
standards, became the foundation for our strategy. Now as we look for
ward into the future, we say, okay, what are the technologies we better 
be providing as we start building these solutions. And by the way, you 
may have noticed I-B-M reorganized along 13 industry lines, the banking 
industry, health care, manufacturing, insurance, etcetera. Where the 
people in those divisions are to become experts in that industry to get 
to know the wants and needs of their customers and help them build 
solutions. The I-B-M product divisions then are to provide products. The 
people in the solution divisions may be choosing solutions from other 
companies, including I-B-M's competitors. S



o it's up to us to build the best technology for them to build solutions 
with.



As we look forward, there were, basically this is my summary of the 
Reader's Digest, John Soyring's Reader's Digest version of the key 
technologies. And there's five of them. And I'll explain what each five 
of those are. And I'll describe what the big shift is.



The first is, we think that the industry is going to be dominated by 
commercially available microprocessors. Now the shift is, before we were 
all building proprietary ones. We had this industry standard processor 
called the Series One. Industry standard processor for the System 38. 
Another industry standard processor for the System 34. Another one for 
the 8100. For the 9370. You recognize all these. Commercially available 
to everyone, right. Wrong. They were available to I-B-M product units 
and that was it.



Well we think the future's going to be driven by the X 86 which is Intel 
and its clones. As well as risk processors such as the Power P-C. And 
there may be other risk processors that will continue to be successful 
as well. Part of our strategy from a software standpoint is, if the 
hardware is going to be built that way, we need to design our software 
once and make it be able to run on Intel and risk processors. Because 
what is the I-B-M hardware company going to do, it's going to build 
P-C's with Power P-C chips in it. Risk workstations with Power P-C chips 
in it. Servers with Power P-C chips in it. A-S 400's with 64 bit Power 
P-C chips in it this year. And massively parallel processing units that 
attach to the main frame to off load processing from the 390's. So 
across I-B-M's product line we'll have Power P-C chips in it.



I-B-M is absolutely committed also to the X 86 family. And continuing to 
build personal computers using Intel or Intel clone products, for 
microprocessors. But again, the bottom line, all of our products would 
have industry standard microprocessors and we see the rest of the 
industry going that way. So from a software standpoint, we're building 
across



 these platforms. It also means cross platform means to put our middle 
ware technology, products like D-B 2, C-I-C-S, the messaging product 
called M-Q Series and compiler technology and other technology that's 
used to build applications, that's why I refer to it as middle ware, 
across not only I-B-M's four major operating systems, which are O-S 2, 
A-I-X, O-S 400 and O-S 390. But also across our leading competitors' 
systems. And I'll describe a little bit of that more in detail after
wards.



So the first part is, we're going to build it cross technology now. 
Microprocessors are showing up in many other places then P-C's. In the 
last report I saw I think that in 1995 there are almost as many micro
processors used in other then P-C's then in P-C's. They're used in 
teller machines, automated teller machines. By the way, 90% of those 
being produced today happen to be running my favorite operating system 
also. As well as having an Intel microprocessor in it. A lot of manu
facturing plant floor equipment is running with microprocessors in it. 
Automobiles are running in it, one major automobile manufacturer is 
going to be putting Power P-C's in their automobile. Now they're going 
to run it on a local area network with multiple Power P-C's controlling 
the environment within the car. Controlling the various engine func
tions, etcetera. So just imagine a very cold and very windy day in 
Chicago, and you walk out and it's 20 below zero and you got your brand 
new automobile with Power P-C's in it, latest state of the art technol
ogy. Turn the engine on, 15 minutes later the system boots up and the 
engine comes up. So we are advancing the state of the art of the auto
motive industry with some great user value. But it will have 
microprocessors in it so it's a good reason to buy the product. Now in 
reality, it will start up. Hopefully everyone caught that that was a 
joke. And when this automotive manufacturer comes out with the car, 
please buy it.



But there's a lot of other equipment and we see that market actually 
growing faster tha



n the P-C markets were. Entertainment equipment, office equipment, 
manufacturing industrial equipment, will have microprocessors in it. 
Guess what. They're going to be hooked into a network also.



In fact, that's the second part of our strategy is that we think that 
almost all of the computers that are being purchased in the future will 
be tied into a network somehow. A public network like the Internet or 
CompuServe or A-O-L or a private network within an enterprise. Or, how 
many of you heard Lou Gersner's speech at Comdex. Just raise your hand. 
One person. Let me tell you about it. I love this story.



This is reality, this isn't just a joke. I-B-M's research laboratories 
in Yorktown, New York and Hawthorne, New York are working with the Mas
sachusetts Institute of Technology's media labs. It's a very famous 
technology lab at M-I-T in Cambridge, Massachusetts. They're working on 
something that's very interesting. It's called the personal area net
worker, pan for short, if you pronounce acronyms. Analogous, the local 
area network, or a wide area network. What they do is they put a micro
processor in the heel of a shoe. And when you walk the energy that's 
transferred from walking is converted into electrical energy and it's 
used to power the microprocessor. And there's software that runs down 
there, there's some memory. The way it communicates is through the 
electrical currents in your body. Now the story that Lou talked about 
is, imagine two people walking up to one another. When you grab each 
other's hands to shake hands, you don't have to talk, you just exchange 
information through your personal area network. Now you might just 
exchange your name. Now if you're in a single's bar and you hold on a 
little bit longer, your telephone number and address might exchange on 
there too. And if you hold on too long, I'm not going to go into that 
what you might exchange.



It was used, actually to be humorous. It's actually a real research 
project. But it was just to try to get across the point, we really do 
think that these m



icroprocessors are going to be used in networks because what we find is, 
when we do studies of people using P-C's and beautiful graphical oper
ating user interfaces, we've got one. Our competitors have one. Most of 
our competitors have one. That we often find that there's an adverse 
impact on productivity. Now how many of you, 20 years ago with a char
acter oriented interface on a dumb terminal, could have created a 
document with as much content as you do today, but spend less time on 
the cosmetics. Of how the information is arranged, nice pie charts and 
everything else. Or if you go up onto the web and you're downloading web 
pages, how much of that is content versus cosmetics. Your bitmaps come 
down, and a lot of images, so we're getting a lot of cosmetics and less 
content. So in some cases there is an adverse impact on productivity 
because you've got some very well paid people using a P-C with a word 
processor and it's so fun, they're spending time laying out the document 
very nicely, making pretty pie charts and presentation charts, that they 
would have done equally as well 20 years ago with a character interface 
and still convey the same information. It just looks a lot better.



What we also found, though, is that when people were connected into a 
network, and they collaborated. They got big returns on productivity. 
When you're creating a document or processing a document that has to 
flow through the enterprise and multiple people have to sign off on it 
or they have to edit it. That's where the big changes came. And that's 
what fueling the drive is that people want to be connected into network 
to access information and be much more productive in their work. And we 
think that's going to happen in homes and in business.



What is says, though, when we look at the technology, is that today's 
local and wide area network technology is pretty good. But far from 
adequate. Far, far from being adequate to connect millions or billions 
of people in a network where they have access to a wide range of infor
mation. It's a ni



ghtmare if you try to manage that environment today. It's an absolute 
nightmare for most of the people charged with that responsibility. So 
what we are doing, at least at I-B-M, is we're trying to address that 
problem by building a set of technologies. One of I-B-M's world wide web 
pages, we have a page that describes and documents the I-B-M open blue
print, which is our specification for some key technologies that we 
think would be able to create this network environment.



How many of you have ever heard of S-A-A, would you raise your hand. 
Okay. Someone chuckled. Who was that. Will you raise your hand.



There's two big differences, in my opinion, between S-A-A and open 
blueprint. One. S-A-A was based on vaporware. It was an architecture to 
be defined in the future, eventually implemented some time in the 
future. And I think we announced it in 1984. And we're still waiting to 
implement some of these. So that gives you kind of an idea how far off 
it was. It was very visionary, but not a lot of content on it. Secondly, 
if you go back and analyze that S-A-A spec, it was based on industry 
standards. I-B-M's industry standards at that time. None of the rest of 
the industry had indicated it was standard. Whereas open blueprint was 
based on technology that is available today. And it's based on either 
defacto standards or dejour standards. And that's the two big differ
ences that drive it.



Because what our customers have clearly told us is that, this is going 
to be a heterogeneous world. They're going to buy hardware and they're 
going to buy software, from many different vendors, not just I-B-M. And 
by the way, they said if we bought it just from you, it would still be a 
heterogeneous environment. So it's got to work together.



So what we're building is, a lot of what we're doing for connectivity, 
we think the Internet is where it's at. We think the Internet is where 
it's at. We're putting a big investment in building Internet technology.



Now, one of the questions you ought to ask, just jump up there and say 
well, does



n't that kill Notes. Because some people have written about that and we 
don't think so. But you know what, we're not going to make Notes to be 
complementary to the Internet. We decided not to. We're not going to do 
that. We're doing something very different. Notes is going to be part 
and parcel of the Internet. Notes, version 4.0 is the current product in 
the marketplace. We ship with it, for no additional charge, a product 
that used to cost several thousand dollars called Lotus Notes. Lotus 
Internotes. Which when snapped together with Notes turns Notes into a 
world wide web server. It has H-T-M-L and H-T-T-P support built into it. 
Now what you get, is you get the power of creating all these documents 
and forms using Notes with workflow management, with security, very 
powerful programming interface, things you don't have on Internet serv
ers. And then can publish it just by Internotes converting that 
automatically to the proper H-T-T-P and H-T-M-L formats. So it's a very 
powerful tool.



There's some very good examples of companies using it. And using it for 
example, how many people have gone into stores, I guess it's mostly for 
women. And you can go and get Levy's designed right to your own mea
surements. They take your measurements in and enter the information into 
a kiosk. That's a web server. Well you know what happens when you go in 
there, if you order your jeans specially cut. That information is cap
tured. Your name and address plus your measurements. It's sent off over 
the Internet to that Internet server called Lotus Notes back at a Levy's 
factory. Where it comes into a document, where in the first factory they 
cut the jeans, the material to your specifications. They bundle all that 
up and along with the workflow in Lotus Notes, ship it off to another 
factory and the electronic paperwork goes with it via Notes. The second 
factory sews it together. They then bundle it up and ship it to a third 
factory that stone washes it. They then bundle up that pair of jeans and 
ship it to a distribution center which ship



s it back to the customer. That whole cycle takes three days. That all 
works on Lotus Notes. It's a very powerful application but it's used 
over the Internet. And that's just one good example.



How many of you read P-C Week or Spencer Katz' Rumor column in P-C Week. 
About 38.5% of you raised your hand. It's my personal area computer 
calculating that, with the electrical currents in my body. If you read 
it a couple weeks ago, and said, oh, I-B-M has really screwed up again. 
They made two big mistakes. They're running a web server in our Research 
Triangle Park North Carolina laboratory and guess what, it runs on N-T. 
And that's two big mistakes for I-B-M. It's a marketing problem and 
secondly they're not even using their own product. Well, it's true, it 
runs on N-T. And as I said, I-B-M is building its technology across 
platforms and we support N-T with our web servers, we our database and 
with Notes and with our communication protocols.



Is anyone here from the press. Raise your hand. Okay, you guys can't 
tell them this. Because we've got to be able to send this into P-C Week. 
That is being converted quickly to A-I-X.



Now it's not because we want to just respond to P-C Week, we had planned 
to convert that for some time now. But when the application was built, 
we wanted to build it with Notes. And that was built last year. Well if 
any of you are Notes experts, was Internotes available on O-S 2 last 
year. The answer is no, it wasn't there. Or the other products that were 
used. Lotus, prior to being bought by I-B-M was building all this tech
nology on guess what, N-T. Well they still are but now they've made it 
available on O-S 2 and A-I-X. And we're shifting this over to an A-I-X 
server. So if you do see Spencer Katz, you can whisper it in his ear, 
that that is going over. The bottom line message is, when you see Notes 
you should think Internet. We think it's a very powerful technology.



The other is, the industry has so many networking technologies that we 
felt it was important to establish a standard. We had many



 at I-B-M but the industry has many more. So what we've declared is, 
we're going to put D-C-E technology, the open software foundations, 
distributed in computer environment technology, and make it available 
for all of our platforms. And by the way, if another vendor is not 
implementing it, and to the best of our knowledge, there is one who is 
not. A company in Redmond, Washington. They've got an operating system 
that some of you may have heard of and some of you might use in here. 
It's called Windows. And I'll explain what it is if you're not familiar 
with it. But to the best of our knowledge, they're not building D-C-E. 
So we are. And making it available on it. Just so that all the major 
operating system platforms in the industry would have the D-C-E tech
nologies, what I think of as the plumbing. So you can connect all these 
systems together.



A lot of you, I saw your name badges as you were walking in, are from 
very big companies. And so I imagine some of you are involved with the 
networks and some very big networks. And what we heard from our custom
ers is that, you know, this is great technology, guess what I-B-M, I'm 
not going to buy it. Because I can't manage it. It's tough enough man
aging these little local area networks and the systems that I've got 
now. Putting in tens of thousands or hundreds of thousands and managing 
them at a central location, is what I require before I put this in.



So. We said you're right. And we started along investing in our own 
technology and evolving something called Netview and Systemview. One 
hundred and fifty some products from I-B-M that address all kinds of 
management disciplines. And we put them under one umbrella called Sys
temview. And it's really good for managing from a centralized main frame 
environment. And there's some pretty good stuff that's out there in a 
network environment on an A-I-X and an O-S 2 system today. But it wasn't 
the state of the art. When we looked at all of our competitors and all 
the companies in the industry we found very few who had the technol



ogy, but we found one way out in front and that's called Tivoli. So a 
few months ago we spent $742 million and we bought the company.





And what Tivoli tech brings is the state of the art technology for man
aging distributed client server environments. Some very, very powerful 
technology. Frank Moss is the C-E-O of that company. And what we've done 
is, now Tivoli is a wholly owned subsidiary of I-B-M. And we moved the 
mission for all of our Netview Systemview functions to Tivoli under 
Frank Moss and he's now merged that, and a couple weeks ago described 
our direction for the future and it's a merged set of technology, the 
Tivoli T-M-E technology and the I-B-M Systemview technology. Managed 
taking the best from the main frame environment and the best from the 
client server environment and melding that together. And it's called 
T-M-E 10. So that's the direction that we're heading in. And this chart 
is a little bit out of date, I just haven't had a chance to update it 
since we did make that announcement.



The other thing is, we have a project in I-B-M that's been code named 
Eagle. And what Eagle is, is to put a common set of middle ware across 
all of the I-B-M platforms and across our chief competitors. And I'll 
describe what Eagle is later on.

But what this builds is the infrastructure then for a networking envi
ronment to build a huge network of systems that share a common set of 
technology that allow you to much more effectively connect and manage 
that environment.



Any questions just on the distributed technology. I'm going to try and 
get out of the strategy session and down to the O-S 2 stuff.



So the first one is commercial microprocessors will rule. The second one 
is, those microprocessors are going to be connected into a network. Now 
we would hope that not just millions but billions of people are con
nected into a network. You go to China, you get a small market share 
there and that's pretty good. There's 1.2 billion people in China. We'd 
like to see a lot of the Chinese working on it. And fortunately, the 
mini



stry who makes the technology decisions in China and Beijing has 
declared that O-S 2 warp and Lotus Notes are the standards for the min
istries in China. Well that's good because the ministries are first of 
all, are very big. But being a Communist country, the ministries also, I 
think, drive 90% of the businesses in China. So having that declared for 
that standard is very good. So when the railroad systems in China now 
are all being automated with O-S 2, lan, lan server. All of the food 
delivery systems to their grain elevators, is all being fully automated 
with O-S 2 lan server and soon warp server and O-S 2 warp. The biggest 
banks in China are now, I was over there and I'm going to put this in 
perspective. How big do you think in terms of the number of branch 
offices, is the largest branch banking network in North America. And 
who's the bank. Citibank is the wrong answer. Bank of America is wrong. 
First Union is almost right. Go across the street in Charlotte. It's 
Nations Bank. Nations Bank, the second largest branch banking network is 
a tie between Royal Bank of Canada and C-I-B-C, with about 1700 branch 
offices scattered across Canada. Nations Bank right now is about 2200. 
By the way going all to O-S 2 warp, if you didn't know about that. As is 
C-I-B-C and Royal Bank of Canada, in their branch networks. 2200 branch 
offices is big business. In China there's one branch network with 
100,000 branch offices.



When I went over to Beijing I said, look, this is going to be an easy 
proposition. All I want is one copy of warp in each branch office. You 
don't even have to have multiple copies and that's a big enough sale all 
by itself. But China is a big opportunity and we're doing a lot to 
invest there. And if any of you speak Mandarin Chinese and know O-S 2, 
please come and see me afterwards because our growth rate is only con
strained by getting enough people over there to know that.



But the bottom line is, we want to get a lot of people. And what our 
studies say is that O-S 2 warp today, Windows 95, Mac O-S, and ot



her competitive alternatives are still too difficult for many people to 
use. That a keyboard and a mouse is foreign to most human beings. And 
even people who use a computer tend to suffer with a keyboard and mouse 
in some cases. Many like it, but there's still a lot who don't like it.

Some of you may have heard this, but I just love this story. We've got a 
human factors laboratory where we bring people in off the street, sit 
them down and just tell them to use the computer. Many of these people 
have never used a computer before. And this one gentleman, there's 
actually quite a few of them, and they almost always were male for some 
reason, took the mouse and picked it up as if it were a remote control 
for a television and started clicking on it. And it didn't work. And 
eventually we had to go back there and tell him, said, no, what you have 
to do is you move the mouse around and that little pointer on the screen 
moves. They guy says, I got it. So he takes the mouse, sticks right up 
on the screen. Right over the pointer and moves it around. Said, I-B-M, 
it moves now but I got to tell you this is not a very good design 
because I can't see the mouse pointer moving. Now, I don't think he's 
representative of everyone we've got to deal with but it just gives you 
an idea there are some people who are, what does David Barnes say, men
tally challenged.



But we feel we can put technology or build technology for our systems 
that make it much easier to use. And one of the most natural ways of 
communicating is with speech. Is to be able to speak and have it recog
nized. And this is another trivia point so in case you're ever in 
Trivial Pursuit or something else, if they ever ask this question. 
Here's one for you. Speech, I'll tell you has been the longest funded 
research project in the I-B-M Corporation's history. Who wants to take a 
guess at how many years. Twenty two is wrong. Ninety is closer. No, it's 
still further off. Someone say 35, 35 is very close. It's actually now 
37. It was first started in 1959. And if you remember



, I think it was in the early 1970's, if some of you are old enough for 
that to remember T-V. There was an I-B-M television commercial where 
this woman pronounced a few words including in the sentence the words 
to, too and two. The three different forms. And the computer recognized 
it. Well at that time the vocabulary was several dozen words. That a 
computer about the size of this room would recognize. In the late 80's 
and early 90's we shrunk that down to a high end risk work station with 
special processors in it. And in the early 1990's got a special card 
that plugged into a P-C with a high performance digital signal proces
sor. And in 1995, a P-C M-C-I-A card that would plug into this think pad 
that would let you speak up to 90 words a minute with greater than 95% 
accuracy.



Now I can't tell you about the future and there's been some speculation 
about what we might do in a product called Merlin. But I won't tell you 
because we officially have not said that there's going to be speech 
technology in it. But there's a new algorithm coming out and we've 
advanced the state of the art. Well beyond what was in 1995. And I'll 
leave it at that.



There's some really exciting things that you can do with this technol
ogy, whether it's separately available or in the product. Yes sir.



(question from audience. A-T and T just did a big splash about voice 
recognition. Is it anything that's going to leapfrog I-B-M.)



To the best of our knowledge we've not seen anything in the industry 
that still comes close to this particular technology. The algorithms are 
very, very powerful. Yes.



(question from audience. Phillips announced their speech ... system in 
the United States. They claim that it's going to be continuous speech 
... Has I-B-M any plans to make their voice recognition system continu
ous speech.)



The question for those who didn't hear it was about Phillips offering 
continuous speech. I-B-M has actually been offering for sometime now, 
continuous speech. Where we are right now is, our voice type recognition 
does offer conti



nuous speech mode. It requires no user training, so you could come in 
and you could have a Brooklyn accent, you could have a Minnesota accent, 
you could have a southern accent, in whatever language you choose and it 
will recognize it without any training. Now, the algorithms, as you 
know, in that case, have to be much richer, then if you had to train the 
computer. And as a result of that richness and the amount of processing 
part that goes on, the vocabulary is relatively limited. And we support 
continuous speech today with a vocabulary of 1000 words. Now that sells 
really well in Texas where that's about twice the size of vocabulary. 
You didn't let me finish. Of a Texas politician. Having just gone 
through the elections, we understand that. But it is very important 
technology and there's certain groups and we build special vocabulary 
for people like radiologists because when you take an X-ray, these 
radiologists who get paid a lot of money, have disposable income so we 
charge $899 for it. Put it up there and they look at the X-ray then they 
write a long report. Or type it in. Wouldn't it be great if they could 
just speak it in. So it sells well to radiologists. The other one that, 
is there a lawyer in here, yeah, there's at least one, I talked to you 
earlier. How many lawyers. You don't have to raise your hand. Don't be 
embarrassed. Lawyers are another one. They've got a unique vocabulary. 
And when you think about all the time they spend writing or typing in 
long reports that they charge a fair amount per page or word for, they 
could improve their return on investment dramatically with this. So 
we've got a special vocabulary with lawyers because lawyers came to us 
and said, we can't afford to keep our fingers on a keyboard. We've got 
to put them in the pockets of our clients, that's where they belong. 
Sorry about that, I couldn't resist. Yes.



(question from audience. Speaking about vocabularies, is there any other 
specialties that you're developing at this time ... interest in my 
field, psychology.)



In psycho



logy. We're very willing to work to develop other fields and psychology 
certainly falls into that category because I can't understand what most 
psychologists say. And it's not that I've been to a lot. So it's a small 
sample only. Right now I'm not aware of any special vocabulary being 
developed for psychologists. But we'd be very willing to pursue that. So 
if you wouldn't mind contacting me directly or through someone here 
locally at I-B-M, we'll follow up with that. I know the guy who runs our 
speech unit at I-B-M is a good friend of mine. He's constantly looking 
for new opportunities. One of the things we did is we started a research 
project in Beijing with Ching Yi University, which is considered the 
premier educational institution in all of greater China. And if you 
think about the Chinese language, I went over there and said, okay, type 
in the word Beijing. Which stands for northern capital if you translate 
that into English. And it was four major keystrokes with 16 shift levers 
going. Took a long time to get that in there. Imagine what could be done 
if you could recognize speech for double bite character set languages, 
whether it's any of the kongi based character sets or others as well.



(question from audience. You have available I believe in Japanese ... 
dictation. Is that available in Chinese also.)



It's not yet available in Chinese. The Chinese language is a fairly 
challenging language, to say the least, to be able to get it across. But 
not yet, but we are working on it. And we're working on Mandarin Chinese 
which is what the Chinese Beijing government pretty much dictates. Other 
dialects would have to come later. Yes, sir.

(question from audience. Are you doing anything with your voice recog
nition technology to eliminate background noise interference for use in 
a noisy environment.)



We are doing more to eliminate background interference in a noisy envi
ronment. The best way right now is having a directional microphone, 
which is what we encourage people to get with the product and we ship, I 
think, w



ith most of the products right now. But yes, algorithmically, we're 
constantly enhancing the product and their enhancements in the next 
generation that I can't talk about yet. That really help.



(question from audience. When is the next generation expected.)



Approximately when Merlin comes out. I think. Sometime soon after mid 
year, 1996. I always have to qualify that because when software vendors 
say mid year, you should immediately say, what year though. Or when they 
say summer, did you mean northern hemisphere or southern hemisphere. So 
be careful along that line. We've got a lot of lawyers working with us 
who help us word phrase this properly. You're going to get a lot of 
these tonight, don't worry. Yes sir.



(question from audience. Did your P-C hardware set up to recognize this, 
assuming it's coming on the market.)

Yes. In fact, we are working closely with the I-B-M P-C company on these 
particularly. Yes.



(question from audience. ... translation software then. ... )



Actually one of the market segments that this has sold best into are 
professional translators. Just absolutely love this. I think the number 
of languages we support today are about six. It's the traditional Euro
peans mostly with German and French and Spanish. (END OF SIDE 2 - TAPE 
1)





Tape 2 - John Soyring



...problem. We've actually had people now report that their vocal chords 
are developing nodules on it, so I'm sure that's going to be the next 
big class action suite against us is that this caused them to talk too 
much. Yes sir.



(question from audience. I think I read somewhere, you changed, it's 
become more open as well, even in hardware. And you're addressing some 
of these algorithms to work on like sound blaster cards as well.)



Yes we are. In fact it's very important to work on sound blaster cards 
that had just worked standard on other cards that are out there, not 
having to buy a proprietary card. It's also why you see an I-B-M, we'll 
still support micro channel but we're really moving heavily to the P-C-I 
bus from a hardware standp



oint. Yes sir.



(question from audience. Some of us are more technologically challenged 
than others here. What is the role between speech recognition and algo
rithms.)



Oh, I'm sorry. The algorithms are the mathematical routines or the 
instructions that have been, in the computer that represent the math
ematical algorithms or routines. That allow someone to take the basic 
input, split it into manageable chunks, and then go through some very 
complex routines to determine what did that word mean and what did that 
word mean in relationship to other words that were just previously spo
ken or will be spoken very soon. So there's some very advanced 
algorithms or mathematical routines that are used to determine what the 
word is in the context of the spoken speech, not just by understanding 
the phonemes, the basic sounds that are being produced. Yes.



(question from audience. Getting back to a question that was asked ear
lier where someone asked, another speaker asked that someone ask the 
question. How much money is I-B-M making, made on O-S 2. And what's the 
future prognosis. And by the way, you won't make money selling to 
patients of psychiatrists because those people are already warped.)



Well, they may be warped but they've got disposable income that we're 
interested in. So we do want to sell to them. The bottom line is, as a 
general policy, Lou Gersner's asked us not to break down our financial 
results other than what we report in the annual report. The I-B-M soft
ware company, all our software within I-B-M we've recently organized as 
a company, is a $12.4 billion company. That's larger than any other 
software company in the world by a factor of 22 with Microsoft at $5 
billion in the number two position. That software is growing at a com
pound growth rate of a little over 12%. Now, normally we don't break 
that down any further. We did recently and we've come out and said that 
of that software, most of that revenue, of the $12.4 billion, comes from 
main frame software or mini computer software. And the growth rate there 
i



s in the single digits. Where our high growth rate is, is on P-C's, risk 
work stations and networks. And that's where we're putting a lot of the 
investment in. What we described publicly for the first time a couple 
weeks ago is that the O-S 2 portfolio products, which falls in that 
latter category, for us is greater than a $1 billion business in 1995. 
So when you think about a $1 billion business when you rank the biggest 
software companies, there's Microsoft at about $5 billion. Computer 
Associates in the $3-4 billion range, I think they're in right now. 
Oracle in the $2-3 billion range. And then there was Novell, at one time 
above a million, I think after the sale of the products to Corel, that 
they're now in the sub billion dollar range and Adobe at approximately a 
billion. So we're right up there with the top four, slightly ahead, we 
think of where Novell or Adobe is in terms of revenue. If we had the O-S 
2 family of products as a separate company. And that's growing rapidly 
too. So that's where we are. We build all of our financial models, where 
you invest money up front in development, just like publishing a book. 
And you recoup it on sales. And we built that model, right now we're 
making money on all the sales. Have we recovered all the investment of 
the past, probably not yet. But we're ahead of where we had projected to 
be. And right now we're, in 1994 we sold about four million copies of 
O-S 2. The 32 bit O-S 2. That grew by 50% in 1995 to about six million. 
So that's a high growth rate. If you compare that against any of those 
other top five software companies, it's a higher compound growth rate 
than any one of them. So we're a big business growing faster than them. 
So it's a very significant part of the I-B-M corporate business. And by 
the way, our first quarter this year has been very good as well.



So we've got a very big business and when you talk about what's impor
tant to I-B-M, these are somewhat a course term terminology and it's 
probably not what we would use with the financial analysts or gen



eral but it, for me, they're kind of simple words that describe the two 
human characteristics of why O-S 2 is so important to I-B-M. One of them 
is greed. And that's a basic human characteristic. But what the greed 
is, we're committed to providing return to our shareholders. And we see 
the growth as being in the network environment where P-C's, risk work 
stations and other computers are linked in a network environment. That's 
a growth area. And O-S 2 is right at the heart of that and it's a big 
business.



The other basic human behavioral attribute. Maybe I should ask you to 
describe it. Is fear. Lot of humans have that. Well what's our fear. 
Well if you think about a billion dollar business, that's tacked on to 
I-B-M as a corporation with $72 billion of business. Some of the ana
lysts said, well, Lou Gersner could take a look at that. $71 billion 
we'll keep, we'll take this O-S 2, some people suggest, we'll toss it. 
Well, think about it this way. For the last five years we've gone to the 
world's 10,000 largest companies and institutions and government agen
cies. And we said, run your business on O-S 2, put your most mission 
critical applications. Your plant floor automation applications, your 
retail point of sales applications, your teller applications, your 
policy and casualty claims applications, your health care administration 
applications, on O-S 2 and run your business on it. Guess what, these 
people invested billions and billions of dollars and did that.



One company, their executive vice president asked me to come and visit 
him. I said, sure. Because he's buying a lot of products from us. He 
said, sit down, I need to tell you something. He said, I've spent $250 
million over the last several years on O-S 2. And I'm going to spend 
probably another $500 million in the next several years. And I report to 
the chairman of a very big company and he expects me to get a positive 
return on investment. I've got to know for sure from you, I-B-M, that 
you're going to be here a long time. Because these O-S 2 applications



 have yielded a tremendous gain in productivity within this company. But 
he said I can't recoup that quarter of a billion or three quarters of a 
billion investment in one year. I've got to be here many years with this 
product running in my organization. So we told him about our commitment 
and that we are committed to being here a long time. In fact, Lee 
Ricewig, at the time, said we're still going to be enhancing O-S 2 in 
the year 2005. This was a meeting in 1995, so he talked about a 10 year 
cycle. So we are committed to it. It's very important. So it's a big 
business. It's growing rapidly. And our best customers are committed to 
it.



Now imagine, which company are you with. You're not I-B-M. Imagine 
you're an I-B-M sales rep, and you walk into your biggest and best cus
tomer and say, you know all I've been telling you for the last five 
years that O-S 2 is strategic. And you invested many millions of dol
lars, I was just kidding. Why don't you switch over to this other 
operating system. Oh and by the way, I've got this brand new strategic 
main frame. Or networking system to install. We would just lose cred
ibility and that goes back to the fear factor.



We could lose the credibility with our customers. We committed to them, 
our biggest and best customers that they could bet their business on O-S 
2. And they have. Yes sir.





(question from audience. ...)



Yeah, I'm going to get to that later. Yes.

(question from audience. ...)



I'm going to get to the preloads, I've got a whole chart on that. If I 
can wait. What I'm going to try to do is get us to follow the outline I 
covered so if you've got questions later on. I will hit them. Yes sir.



(question from audience. Are you going fire the advertising staff of 
I-B-M ...)



No. I'll get to that too. What I'll tell you is if you looked at all 
those television ads this last year. None of those were ever intended to 
be a warp ad. Now it happened that some of them featured warp as part of 
the ad like the nuns. Or the rain forest ad where warp played a role in 
it. I have all k



inds of people come up to me and talk about, well I watched these two 
French guys walking along the river. And I couldn't figure out why 
you're trying to sell me warp in that warp ad. Had nothing to do with 
warp, never was intended to be. Nor that whole series. The series of 
television ads are called a solution for a small planet. And what 
they're targeted for is one thing. I-B-M as a brand image. And it goes 
back to what I mentioned in the beginning. The first thing that Lou 
Gersner wanted to do, is shore up the I-B-M image. To establish confi
dence and an image of I-B-M that I-B-M's approachable with innovative 
solutions.



So we shored up the financial part but we needed advertising to let 
people know that I-B-M is approachable and they've got very innovative 
solutions. As we go back and measure the effectiveness of those series 
of television ads and print ads, we find that it's a very positive 
effect. Every year the advertising companies of the world measure brand 
equity of the name brands of companies. The I-B-M corporate logo, as a 
brand, was always up in the very top rated of all the 5000 largest com
panies of the world.



Several years ago it dropped right off the charts. And people were los
ing confidence in it, they thought I-B-M was had ancient technology that 
was going nowhere. And we had to shift that around. So that was the 
focus of those ads. If I can hold off on the warp ads until later, I do 
cover that. And the advertising agency. But we are keeping O and M, 
which is the advertising agency that produced those ads but they're 
heading in a different direction. I'll explain that. Yes sir.



(question from audience. ... )





What I'm trying to do is gather together, I know the North American 
sales. Pretty precisely. And I have hints at what our European and Asian 
sales are and Latin America. But as you gather up results from 180 some 
countries and aggregate it, then do a consistency check on it to make 
sure it's all accurate, it takes a while. So I'm trying to get that 
pulled together because it's important



 to get that information out that we had a very good quarter. And I want 
to be able to get the numbers out. So I'm hoping in the next few weeks. 
Remember the quarter closed on March 31st. Even with our computer tech
nology, it takes a while. I mean, those abacuses. Yes sir.



(question from audience. ...)



Right. Yeah. No, it's a good question. I was going to address that a 
little bit later on. I do want to cover that specifically. All I want 
you to know is that for the last five years, that's why I mention, our 
primary target market segment. And we stated this publicly, Lou Gersner 
stated it on July 31, 1995. Our primary market segment is our large 
corporate customer. That's what we're targeting with O-S 2 warp. A sec
ondary ones are some of these other market segments which I'll address 
later. But we wanted to make sure people clearly understood it was sec
ondary. But we will address them and I'll get to that a little bit 
later. And exactly the rationale.



Your other comment was on the subtitles of wanting to have English with 
English subtitles. Well we did that if you caught one, there was one 
that featured the A-S 400. A bunch of surfers out in California. They 
were speaking Californian English and there were subtitles translating 
it into American English. So we had that. I'm really not an expert in 
advertising. All I can say, all I want to do is get across the message 
that the series of television and print ads for a solution for a small 
planet, were targeted at one thing and that's the I-B-M corporate brand 
image. And from all the measurements we see, it had a very positivePead 
projector and project up in there. He's a very innovative guy. And he 
was interviewed. Because his primary area of research is on human 
interfaces with technology. And what he describes his major interest is 
this agent technology.



He defines there's two types of agent technology. One is human interface 
agent technology. And the second one is investigative agent technology. 
Now what the human interface one is, you can think about a rel



ative primitive form of it would be the wizards that Microsoft has in 
its products. Where you can click on something and it will guide you 
through step by step. But how many have used the wizard several times 
and the second time you go through, you got to go through the same num
ber of steps. There's 40 steps to do it the first time, you learn 30 of 
them. The second time you do it, you got to read through all 40 of them 
to get it out. Well what the technology he's working on is, what he 
calls adaptive agents for the user interface. Where you read through it, 
and the computer system learns that, hey, the first time this guy got it 
mostly right. I'm going to shorten it the second time. And the third 
time and the fourth time. As long as he keeps getting it right, they 
keep shortening it.



So the key thing is on interfaces is his belief is, that technology like 
wizards are not worth it. Because they make you dependent on it and 
you're constantly just going to the wizard and you read through every
thing, you don't learn. Whereas this adaptive technology, it forces you 
to learn the interface and when you learn the interface, you're much 
more productive with a computer human interface. So that's one basic 
theory and I-B-M's working on that type of technology. This is kind of 
long distance vision technology. Could take weeks to develop. Many, many 
of them. I won't say how many.



The other one is this investigative technology and his analogy is 
investigative technology is analogous to a virus. That can travel 
through a network and do things. Now our hope is that this is going to 
do good things on your behalf. But you could imagine some bad things 
that an agent could do. It could travel through a network and search all 
kinds of databases and gather information that it shouldn't be doing or 
it could damage systems. So we have to be very careful in the second 
area as to what's introduced. But there is agent technology that will be 
able to travel through the network and do work on your behalf. Because 
we firmly believe that in



formation at your fingertips is good. But far from sufficient. Which you 
really want to do is you want a network to provide information and ser
vices for you. Not just information. You don't want to have to go and 
search 20 databases to find out where's a good restaurant in Chicago. Or 
the best pizza place in Chicago. You'd rather ask your agent, what's the 
best pizza place in Chicago and use some algorithm to figure that out 
based on what I like. Domino's. So that's some very impressive technol
ogy that I think you're going to start seeing in the relatively near 
future coming out in future products from I-B-M. And for other vendor's 
products as well.



Now that's three. You got industry standard commercial microprocessors, 
connect them on the network and make them much easier to use.



The fourth one is really targeted to programmers and that's object ori
ented technology. And someone ought to just stand up and say, okay, 
John, object oriented technology was developed, I think, in about 1966 
with a simula language. That's 30 years ago and programmer productivity 
in 1996 is about the same as 1966. So why do you say this is a panacea. 
Well how many of you travel internationally. And carry with you an 
electric razor or a hair dryer or anything else. What else do you carry 
with you. Adaptive plugs to plug in. Because the electrical plugs, the 
current, the voltage and the physical size of that adapter is different. 
In fact last time I was in Moscow, in one room they had three different 
plug types. And my think pad had a fourth. And I was about to give a 
speech last year at a big computer conference there.



The problem in the software industry for the last 30 years is the same. 
I could build component A, you could build component B, we could build 
it with two different releases of the same C++ compiler, try to link 
them together and guess what. They don't work. So what happens is, pro
grammers go in and they write that glue code. And then secondly, you 
have to ship the source code to them so they can recompile it, and dig 



into it. Well, what's the problem with that, this glue code was so hard 
to write that most programmers just toss these objects out and said it 
wasn't worth it. It's not worth doing, they just rewrote the code from 
scratch. So that sort routine, it was written for the 11th billionth 
time in this industry. Same function. Just 11 billion different imple
mentations of it.



There's two big things that changed. With the object management group 
agreeing on a specification called corba, the common object request 
broker architecture. And they are currently up to corba 2.0 as the lat
est level of the specification. Defines how you can basically provide a 
logical bus that you can plug in components and they can be created by 
any vendors and different languages. And they'll be able to communicate. 
It's very powerful. We've demonstrated it recently at, not the most 
recent one, but the previous net world plus inter op. With Sun and 
Hewlett Packard and I-B-M running objects on their systems, collaborat
ing with one another over a corba pipeline that was connecting them. And 
all we were doing was plugging our objects in. And it worked. It could 
demonstrate that then multiple vendors can get to this point.



The second thing is, they were creating these objects in different lan
guages. Yes sir.



(question from audience. How does this differ from visual basic exten
sions.)



Well visual basics extensions can do the same thing and is based on one 
standard also. As long as the Microsoft V-B-X, or now Active X, support 
is implemented everywhere. And by the way, it went from O-C-X or I can't 
remember all the evolution of the products now. V-B-X's to O-C-X's. To 
Active X implementations and I think it'll change again when they come 
out with their common object model. And it was disruptive each time. Is 
that that's available on whatever platform you want as long as Microsoft 
puts it there. Which is basically Windows today. Not across a heterog
enous platform that many different vendors are implementing. So that's 
one difference.



The other dif



ference is, is this is based on object technology. Microsoft will say 
no, we're on object technology. We support encapsulation because we can 
wrap stuff around it. Let's see, do we support polymorphism. Yeah, you 
can change the interface so when you invoke a method, time A it does one 
thing and time B it does it does another thing. That's the second char
acteristic of an object.



And the third is something called inheritance. You ask Microsoft, do you 
support inheritance. They said no we don't, that's bad stuff. It's too 
complex. Programmers will never use it. Well, the biggest benefit of 
objects is actually inheritance. Because the number one nemesis that 
we've got in this industry is maintenance of code. And what happens is, 
the traditional way that we used technology in the past is we take a 
routine, we say we like it, you know, our neighbor's got it on his desk. 
We make a copy of it. And then we make modifications to it. And that 
permutation happens four billion times. So that object has four billion 
different versions to maintain. It becomes a nightmare.



Microsoft instead of using the word copying sections of code, calls that 
aggregation. And that's its third characteristic. Where this truly does 
support both single inheritance and multiple inheritance capabilities. 
And if you go to most computer scientists, they'll say you have to 
exhibit polymorphism, encapsulation and inheritance to be truly object 
oriented. So that's the two differences. V-B-X's are based a proprietary 
standard and on procedural programming.

So with that, the industry really can build these objects that we're 
demonstrating. I-B-M is committed to implementing the latest level of 
the O-M-G spec in all of its operating systems. In fact we think every 
vendor in the industry is implementing except for one. And that's the 
one we just talked about. So because Windows isn't committed to it, 
I-B-M has built already for Windows 3.X and we're building for Windows 
95 and Windows N-T. These basic set of object services. So you'll be 
able to build in



dustry standard objects in a Windows N-T or Windows 95 environment. And 
have it collaborate over a network. The other thing, V-B-X is only 
remote, really work locally. They don't work well over a network envi
ronment, whereas this is designed for a network environment.



The second thing that we're doing is we're building a standard set of 
objects. It's called I-B-M Open Class. About three weeks ago we just 
released it on Windows N-T. It's open class, it ships as part of I-B-M's 
visual aid C++ product today. We're doing something, I-B-M's implemen
tation, this is called som, or the system object model. We're in the 
process of what I call sominizing these objects. Or putting a som wrap
per around it. Instead of a C++ wrapper, when you wrap this with som or 
the corba specification, it truly does give encapsulation. Because if 
you're a programmer, you know that with C++ the methods are in there but 
the header file is outside of the object and you can access the header 
file so everything is not encapsulated. And it gives you language 
dependency. Where we want to make this language independent, so by 
wrapping it around we provide the language independence. With very 
little overhead. In the range of 3-5 Intel instructions of overhead.



Now if your method in here, when you go in is, is a six instruction 
method, and you've got five instructions of overhead, that's a big per
formance penalty. I think Marty has written a lot of object oriented, 
Marty Cawthon's written a lot of object oriented som code and can give 
you some good guidelines. And I believe he's talked with you previously 
on that.



But we're building these object classes, there's about 500 objects or 
classes that I-B-M ships with Open Class today. And we're growing that 
dramatically. There's a very interesting object class that's in there. 
You might not be familiar with. How many of you are interested, raise 
your hand, in object technology. Is this worth talking about. Okay.



One is called the Compound Document Framework, that I-B-M just started 
shipping, as part 



of the most recent release of Visual Age 3.5. What that does is it's a 
higher level abstraction where you can now write software and generate 
O-C-X's or O-L-E code. In the future it's going to be adapted so it will 
generate either open doc or O-C-X's. So with one high level abstraction, 
you can be more productive in creating objects and those objects can be 
used either for O-A 2, or Active X now. And open doc objects. So there's 
a wide range of visual components and computational components that can 
be assembled together to build an application. And what we're doing is 
putting these across all of I-B-M's operating system platforms. O-S 2, 
A-I-X, M-B-S, O-S 400. And we're putting them on Windows N-T. We've got 
work going with H-P on H-P-U-X. We intend to put it on Sun Solaris. And 
on other operating systems. What it means to you, is that you'll be able 
to write an application once. If you write to these object A-P-I's, and 
through recompilation, port to other operating systems. Literally making 
your application independent of hardware and the operating system. Yes 
sir.



(question from audience. Can an object become real time.)



Can an object become real time. Mean with dynamic linking. Rather than 
static linking. Marty, you better jump in here. I'm going back to some
thing I'm recalling from a long time. But I know we support the static 
linking now and I think there's work underway for dynamic linking in the 
som technology.



(question from audience. Eventually I guess you'll want to have som 
objects distributed over a network. ...)



Well I take that back, I got it wrong. It's dynamic linking that was 
first in it. It's static linking that's being added, I believe. The 
dynamic linking has a performance penalty versus static.

(question from audience. Dynamic linking in som objects and with dis
tributed som there's dynamic linking across the network but you can link 
the som object across the network ... and that's called ... )



Let me run through a little bit more of this because we're going to run 
out of time in about



 25 minutes. And I haven't gotten to Merlin and a few of the products 
that we have under development. We're still at a very high level and I'd 
like to drill down if that's okay with you. Because from the comments I 
got before hand, a lot of you are interested in what's coming.



The fifth of the five key technologies, and this is going to address the 
Win 32 question that came up earlier. Is the fifth pillar is what we 
call investment protection technology. We know that a lot of our cus
tomers have invested billions of dollars buying licenses for software. 
Developing software. Putting in networks and training users. And I-B-M 
has been long known for protecting investments. In fact if you're a 
programmer in the early 1960's, and wrote for the 1401, 25 years later 
you could still run that code because we kept emulating the 1401 
instruction set on main frames. Well, in the O-S 2 environment we are 
committed to continuing to enhance and support the dos 16 bit environ
ment. Now that's important to a lot of our customers because they view 
that other dos manufacturers are abandoning dos. We like to have all 
those dos users come to O-S 2 warp. So please get that message out to 
dos users.

There's another one. We are committed to continuing to enhance and sup
port the 16 bit windows environment. There's at least a strong feeling 
amongst a lot of windows 3.X users that their former vendor who provided 
the product is not providing support for windows 3.X. They claim there's 
60 or 80 million of those users over there. We're welcoming them all to 
the church of O-S 2 warp. To come in and run their applications in an 
O-S 2 warp environment and we're committed to continuing to provide 
support. We'll provide bug fixes. Because we've got the source code. And 
we've got rights in perpetuity. Which is a pretty long time. To be able 
to provide support for that.



Another question is, is there's thousands of Windows 95 applications on 
the market. Thousands of them. All I ask you to do is go into the store 
and turn that shrink wrap on its sid



e. And read that fine print that says will work on Windows 95 or windows 
3.11. That strikes some, who, what'd they do in here. Some magic. Well a 
lot of those we find are win 32 S applications. And a lot of them what 
we're finding are written to the win 32 S 1.25 specification. Now in O-S 
2 warp, we support the 1.15 level of win 32 S. We're upgrading that in 
some beta test now. And in Merlin we plan to support the win 32 S 1.25 
level of the spec. And, yes, Microsoft is currently at the 1.3 level 
with the latest spec that's come out. And some applications, I think 
it's Pagemaker 6, supports 1.30. We do have that on our list for future 
enhancements. That 1.30 you should not expect to see in warp. So that 
would address the large number of those win 32 S applications or win 95 
applications that are branded win 95.



If you think about I-S-V's building those applications, now why would 
they build it that way. Well they've got this market of, if they're 
going to sell to the corporate market, of win 95. What we're finding is 
that a large percentage of the corporate and larger governmental agen
cies and institutions are saying, keep window 95 out of my offices. If 
you want to bring one in, bring in a 32 bit operating system such as 
Unix or O-S 2 or N-T. Now there's not a lot of applications written for 
N-T or not of a large installed base. So from an I-S-V perspective, 
targeting the corporate market, there's not many out there. O-S 2 has a 
much larger market share.



But what's the big one out there. It's win 3.11 or windows for work 
groups is the large installed base. So it's only wise for the I-S-V's to 
build for that environment for some period of time. And that's what 
we're seeing happen. So we're supporting that... (END OF SIDE 1, TAPE 2)



...I-S-V's. Call them D-S-V's. Dependent software vendors. That they're 
dependent on one particular vendor. But we recognize that they'll prob
ably be successful in attracting some important applications over there. 
What we also find out, I wish I had a piece of paper. Assuming you a





ll studied set theory in school at some point. Think of a circle repre
senting all the Windows 95 applications. Think of another circle, and I 
purposely moved over here, representing the windows N-T A-P-I's. You 
know what, those aren't a perfect overlap. They do intersect. But they 
don't have a perfect overlap. What we found is that within that inter
section where there's a common set of A-P-I's and if you were an I-S-V, 
wouldn't you build your applications to run on windows N-T and Windows 
95. Therefore meaning you run in that little intersection.



What we did is we analyzed 9 million lines of application source code 
for applications written for that environment. And found that there is 
737 A-P-I's that are used in well over 80% of the cases. And we've 
implemented them. They went by the code name of dax. D-A-X. We're not 
going to use that as a formal product name when it comes out in respect 
of our friends at Paramount. But dax is the name of those A-P-I's and 
300 some messages. What it means is two things. One is, if you've 
already written an application to the win 32 A-P-I set. It means porting 
over is much easier now. Dramatically reducing labor because most of the 
A-P-I calls will simply recompile. What it also means if you're a really 
a smart vendor, is you're going to write to the dax A-P-I because with 
one set of source code, you can compile for Windows 95, compile it again 
for windows N-T and compile it a third time for O-S 2 warp. And we're 
seeing that happen.



We had one vendor in Toronto who's going to come out with a desktop 
publishing package on O-S 2. They had decided before we told them about 
dax, that they were going to build an O-S 2 version of their desktop 
publishing package. They estimate it's going to take them at least six 
months. We provided them an early version of the dax code. They recom
piled and it worked. It ran. There were bugs in it, the bugs were in our 
dax code. We're still fixing the bugs. We're still working on the per
formance. But it points out how powerful it is.



Lotus with Fr



eelance, the next generation of Freelance when it comes out. The one 
coming out on O-S 2, one set of source code. For Windows 95, and for O-S 
2 warp. Simply recompiled. The next generations of 1-2-3 and organizer 
and others, we're writing the same way. Adobe's acrobat was written for 
O-S 2 using that technology and others. So this dax technology which I 
won't use publicly, once the product is announced, is in there.



Now, the question often asked is, are you going to take dax to its full 
conclusion of supporting all of those A-P-I's. And can you do that. Well 
the answer is, yes we can. We have the technical resources, we have the 
skills, and we have the legal rights based on the termination clauses, 
you understand what those are, in the separation agreement that we had 
in our contract called the J-D-A. Or joint development agreement with 
Microsoft. So we have all the rights and skills that we need to do that. 
And right now we've been spending the last several months really lis
tening to our customers as to what they want, and what's important to 
them. So we are taking a very serious look at that and how we do it and 
really welcoming your input on what we need to do. What we find is that, 
with the win 32 S support, the 16 bit windows support and the dos sup
port, we're supporting the large majority of the applications now and 
we're going to continue to drive into that. But we are not precluding 
adding other A-P-I's.



It could include something like even spec 1170. Which is the unit defi
nition for the Unix A-P-I's and A-I-X is the first system branded. X-P-G 
4 compatible meaning it passed the test with those A-P-I's. O-S 390 has 
now implemented the spec 1170 A-P-I's. It's a Unix system. We expect to 
get X-P-G 4 branding on O-S 390. The M-V-S system this year. A-S 400 or 
O-S 400 has about 90% of spec 1170 implemented today. And it's on our 
candidate list so we'd like to hear from you, if that's important in the 
O-S 2 environment.



It means also supporting networks in the future. We know there's a lot 
of Novell netwar



e out there. And one of the things that's becoming popular is netware 
4.1 with what's now called, Novell directory services. Was netware 
directory services until a couple weeks ago. And Merlin is going to have 
the netware client built right into it with N-D-S support. So it will 
have the netware directory support. So if you're a netware user, you'll 
love it. Yes sir.



(question from audience. ...)



I have to go back and check at that level, but as I understand we're 
doing endis and O-D-I support both.



(question from audience. ... maximum advantage of ... wouldn't be ...)



Oh, you mean the netware on endis. I'm not aware of that because this 
code is coming from Novell and as far as I know it only supports the 
O-D-I support. I have to go back and check on that. T-C do you know the 
answer to that. Okay. You got me on one. Years ago I would have known 
that, but I'm fading fast. Yes sir.



(question from audience. Isn't dax kind of a two way street, might dis
courage developers from developing to the native O-S 2 environment...)





Dax is the native O-S 2 environment. Dax is not a translation or emula
tion layer. Those are native O-S 2 A-P-I's. What we took a look at is, 
when you look at the different A-P-I sets. We have a subset of our 
A-P-I's in O-S 2 are called the G-P-I, the graphical programming inter
face. Microsoft's got an analogous called G-D-I, graphical display 
interface. Take at look at them, functionally, semantically, syntacti
cally they're very similar. They're just enough different to cause 
severe headaches for people. Like we reverse the X and Y coordinates 
between the two on where you put a dot on the screen. So what we're 
doing is, we just said, we concede, we're going to take away those 
headaches and we're just going to put native A-P-I's. The message is 
right for the dax A-P-I's. They're native O-S 2. They happen to be 
native N-T and native Windows 95. The reason is, we don't care. I don't 
say that belittling, this is very important. We don't care because we 
think these are procedural A-P-I's and those



 are the past. We think the future will be dominated by object orient 
and that's where we're putting open class across all the platforms. And 
what we're encouraging people to do is write to open class. If you still 
want to write procedurally, write to dax. It's high performance, it's 
native O-S 2. And it's an insurance policy because if you ever want to 
go to N-T or Windows 95, for some odd reason, you can recompile. And 
that is a tremendously powerful selling message when someone finds out 
how easy it is and how protective and independent of the operating sys
tem they are with their application.



It also, from an I-S-V, opens up a much larger market opportunity for 
them. With 13 million O-S 2 users out there, who we think are going to 
be moving quickly to Merlin.



Let me move on because I've got a lot in 15 minutes and they're giving 
me the 15 minute hand sign. I assume that middle finger was the one and 
the other one was the five back there.



The message I had is, all of these technologies we're building, where 
they exist, on industry standards. And we think that the solutions in 
the future, applications, will be built on these distributed micropro
cessors connected in networks. Therefore the primary programming model 
will be client server. And that's why we're building the object tech
nology to be operated, so objects can execute anywhere and be easily 
constructed into a client server model. We're focusing primarily on 13 
industries. And I won't name them now, but I will be happy to name them 
afterwards. And if I get them all, I win the prize.



I'll jump down to our product strategy. Our strategy at I-B-M is to 
employ these technologies across all our systems. So when we build user 
interface technology, like speech technology, we put it on P-C dos and 
windows and some people it M-S dos and windows. And they can do that. 
Put it on O-S 2 warp and on A-I-X end on warp server. When we build 
object oriented programming plumbing and object oriented programming 
class libraries and object oriented tools, we put them on 



all these systems. And when we build networking and network management 
technology for the system administrator, we put them on all these sys
tems.



Here's an eye chart. That's the next slide. Same diagram, different 
titles down here. We're doing the same thing on O-S 400, O-S 390, and 
our leading competitors. When I say windows, it primarily means Windows 
N-T. I made the statement I think Windows 95 is stillborn for most and 
that Microsoft will abandon that. Sorry to use that term, but I really 
think they're going to switch people to something else and they're going 
to try to move them to whether it's N-T 40 or something else. We think 
that's where they'll invest their money as well as Telarus and H-P-U-X.



So the thing is, you can get user interface technology. When you write 
an application, if you use this technology, you'll be able to go across 
platforms. For the heterogenous environment this has been very attrac
tive. Now this is warp, you know this. I'll go on. In an O-S 2 user 
group, why should I talk about warp. You've got enough people talking 
about that.



The product by the end of 1995 that was selling best was warp connect. 
Which is basically the warp product but with these networking features 
added into it. And I assume that you're familiar with warp connect, but 
what we came here to talk about is Merlin. That's why I was rushing.



Merlin is the next generation for these two products. Where to avoid 
some of the confusion, we think that almost everyone, and based on our 
sales, almost everyone's going here anyway. That we would have warp and 
it will have connectivity built into it.



The big feature is, warp is going to be designed to be the best system 
to connect it to a network, regardless of what it is. With Internet 
built standard right into it, a much later level of T-C-P-I-P protocol 
stack built into it then is currently in warp connect. A better browser 
technology. Not just tacked on. Integrated into the operating system. So 
that when you go to an icon on the screen, on the workplace shell, 
double c



lick on it, or do something to it, it could start up a program on your 
hard disk and load it and open up a window. It could start up a program 
somewhere out in the network and run that and return the results in a 
window on your screen. And the user won't see the difference other than 
that time lag. So we're building it in. When you want to browse your 
files, you open up your file folder, you see the A drive, the B drive, 
the C drive, the D drive, the X-Y-Z drive and the local area network. 
The F-T-P drives, out within the Internet somewhere. Show up in a drive 
window and you'd be able to just take a look at them and drag and drop. 
Find a file on an F-T-P server, drag it over to your C drive and drop it 
in there and you got it.



It's a very powerful networking technology with some significant 
enhancements in the user interface technology. Because we think 
Microsoft did some things good in Windows 95. And that they particularly 
did a lot of some nice cosmetic work. And we're working on that as well.



Security is an important part, particularly for more of our corporate 
customers but more and more anyone who gets on any network wants secu
rity built into the operating system. So we're putting the basic 
enabling security features in it so that with third party products you 
can plug them in and meet the requirements of various companies and 
governmental agencies.



I talked about the support for more applications. This means enhanced 
win 32 S support, that I described. And the dax A-P-I's. Something oth
ers have been asking in performance improvements. And some pretty 
substantial ones, I think.



Our target is sometime after mid year. This does not mean December 31st, 
it's much close to the other mid year of June 30th.



In the next few weeks if you watch a world wide satellite video broad
cast we have, you may see the first public glimpse of O-S 2 or of 
Merlin. We haven't given it a formal name yet, that's the code name. Yes 
sir.



(question from audience. Is Merlin going to have ...)



No. It won't have a 32 bit H-P-F-S be



cause that comes with a financial relationship with Microsoft. We still 
have a financial relationship with Microsoft. So it's only going to be 
in the warp server product for right now. Secondly is, that it's prima
rily designed as a network file system, in that the big performance 
benefits are when you access it over network, because everything stays 
at ring zero. And that's actually more important than the 32 bit aspects 
of it. The biggest performance benefits come from its caching out rhythm 
is much better than the standard H-P-F-S. And the fact that when you 
come in over net bios or T-C-I-P, stay at ring zero, you process your 
file request and get the results at ring zero. Your caching is at ring 
zero. And you get it back out of the network over ring zero. So you're 
not going through the security ring switching that's built into the 
Intel architecture. And you save the performance. The fact that it's 32 
bit code versus 16 bit code, really doesn't contribute that much to 
performance. Because some 32 bit elements actually slow it down and 
other improve, shorten some algorithms and speed it up. Overall, it's 
the caching and ring zero support.



(question from audience. Will some of that aspect then be ... and will 
... zero ...)



There's quite a few things we haven't announced about Merlin yet that I 
can't announce tonight. I explained to several people earlier, this is 
the week that I-B-M announces, today actually, is the day we announce 
our earnings and we put a blackout period around any new product 
announcements for a week around that, so I just am obligated not to say 
anything for any new features we haven't described. But that might give 
you a hint anyway. Yes sir.



(question from audience. How much more resource hungry is Merlin going 
to be compared to warp ... )



Well, one of the things we're going to do is we're going to get rid of 
that little statement that says it will run in four megabytes. And we're 
just being realistic. We will recommend eight megabytes minimum. So for 
the base set of colonel function



s, if you keep the same level of functions, about the same amount of 
disk and it's about the same amount of memory. Because we are just going 
and fine tuning algorithms in that area. We are adding in a lot of 
additional functions. And if you add some of them in, they take up con
siderable amount of space. We'll probably only test it also on 486 and 
faster processors. Because we find very few of our customers running 
386's. And very few buying new 386 machines. But that's primarily a 
testing statement because our testing budget is huge. And that's a big 
part of it, so we can eliminate it, we can focus on other things. Yes 
sir.



(question from audience. Is there going to be any printing organiza
tion.)



Yes there is. I can't tell you about it though. Yes sir.



(question from audience. Are there any plans to run support for the 16 
...)



There's no plans right now to drop the support for the 16 bit O-S on the 
X 86 processor. We'd love to see it atrophy. Because as soon as it 
atrophies, and if our customers say, we don't need it anymore. And if we 
could pull it out, it would reduce the size and improve the performance 
of the base operating system. So we would love you to be able to get all 
of your I-S-V's, your applications converted to full 32 bit and tell us 
you don't need the 16 bit code. We would love to rip it out. But we have 
no plans to do that. We've got a lot of customers running 16 bit O-S 2 
applications and we want to keep them happy.



The other product is warp server, just came out in the market. I think 
you had a session on warp server, including a demo here. Right James. So 
I'm not going to cover that again.



I want to cover the philosophy very importantly. Warp server was the 
first product from I-B-M where we're taking something like, what we have 
today, about 20,000 software products in I-B-M, which I know everyone of 
our sales reps understands everyone of them in depth. We're going to 
reduce it down to a fewer number of products that make sense. But warp 
server really is a packaging of about a dozen pr



oducts. And what we found is that when one builds a network operating 
system, whether it's Novell netware, N-T server, O-S 2 and lan server, 
they usually went and bought a data management product, a remote access 
product, a tape backup product and a variety of other products and 
integrated them all together. And they found all the ones that release. 
Well if you think about the cost of that, and think of it as an iceberg. 
Buying the license is the tip of the iceberg, from the cost of owner
ship. The labor costs associated with procurement, going out and finding 
all the releases that work together is high. The cost of integrating, 
whether you do it yourself or through a consultant or a service agency 
is high. The cost of testing is high. The cost of administering 12 dif
ferent license agreements just from I-B-M, if it was, with 12 different 
terms and conditions, is high. And having 12 different points of contact 
for service and support, drives up your cost.



What we've done is integrated all of these, tested them all together as 
one product, one installation program. There's one license agreement and 
one point of contact for service and support. What we're trying to do is 
take that huge portion of the cost of ownership and shrink it. So we 
introduce this product, shocking our competitors by pricing it at $629, 
well under our competitor's prices for a network operating system, add
ing in much more function then they do. Even capabilities of attaching 
in a lan, and support and have a big printer like 300 lines per minute. 
If you don't know what a 300 line per minute printer, if you haven't 
seen it. It's the one that takes a redwood on one end, sticks it in and 
comes out photocopied on the other 300 pages a minute. We put in main 
frame style data management, hierarchical storage management, can be 
attached on to this one. So some very main frame type environments now 
on a server. It's very rich.





Why do we do that. We got into the market, zero market share. Novell was 
dominating it. Well what's Novell dominated. File an



d print serving. They're great at it. Has a great file and print server. 
Well we saw the market going differently. We saw network operating sys
tems becoming application servers. And in 1995, and I'm going to get 
this wrong by about two percentage points so I apologize to I-D-C. But 
of the entire network operating system market, about 56% or 58% was file 
and print servers, in 1995. 42-44% were application servers. Well guess 
what happened. In 1995 O-S 2 and lan server became the world wide market 
share leader for application serving in the industry. Ahead of Unix 
systems, ahead of netware, ahead of N-T. Number one in application 
serving according to I-D-C and of course, according to our numbers.



We go into the year with market share leadership. Now what happens. 
I-D-C estimates that file and print serving will grow at 11% compound 
growth rate. They estimate that application serving will grow at a 33% 
compound growth rate. Now for those of you who are mathematicians, 
project that out to the year 2000, says, 70-80% of the network operating 
systems in the year 2000 will be application servers. So here we come 
in, starting 1996 with the market share leader for application serving, 
we put in tremendous application serving functions to build off of that 
and price it well under our competitors. And our goal was to make this a 
competitive, meaning we'd have comparable file and print performance on 
this product. And darn it, we missed the objective. P-C Week on April 
1st published their net benchmarks, and warp server was 26% faster than 
the number two product which was N-T server and which was faster than 
netware. So we overachieved as being a file and print server on top of 
having a very, very robust application server. So we think we've got a 
tremendous product and this product had a great first quarter. And it 
was just introduced on February 26th.



That's the philosophy. That philosophy is carrying forward in another 
product I'll describe briefly. I want to just spend a very quick moment. 
We introduced O-S 2 warp for



 power P-C in January, and the beginning of February we said we're 
stopping all development in 1996. And we're stopping all marketing of 
the product in 1996. Now we may resume in the future. Some people call 
that dead, I call it, not in a coma, but it's suspended animation for 
right now. It's great technology, we may use it in the future.



The reason being, we sold 6 million of these puppies in 1995. We were on 
track to sell maybe tens of thousands of these. Yet we were putting over 
half of our R and D budget into this one. And we were putting over half 
of our marketing budget into it. Now when we talk to our customers, what 
do they want. They wanted the X 86 based product. In fact they wanted 
features that we weren't putting in it. So we got a couple thousand 
programmers working on it. Imagine if we put a couple thousand program
mers all on the X 86 base, and started building the features that you 
wanted. Then we could drive it through and that was the reason for our 
decision. We do plan to use this technology in the future and may get 
back on to the power P-C.



Along with this application serving, you know we need more than operat
ing systems, so we built a set of servers that will run on warp, A-I-X 
and on Windows N-T. These were 57 separate products from I-B-M that have 
been designed down literally to work as one, but they're seven sepa
rately licensable products. For a database server, communications notes, 
Internet transaction directory, and system management. So you can build 
a very robust application on this.



These functions are also available on O-S 390 and O-S 400. So you have 
the basic platform. Of these shipping, seven of seven are available on 
A-I-X today. Five of seven are available on O-S 2, the remainder will be 
available before the end of the year. And two of seven are available in 
N-T and by 1997 all seven will be available on N-T.



So we're building the strength of this as an application server. Now 
applications are key on people's minds. And what our focus is is getting 
the right applications f



or the customers that we're serving. I want to go back to the market 
segmentation. Our focus over the last five years has been selling to the 
world's 10,000 largest corporations. That's where almost all of our 
money went. We did make it available in the retail channel. We got a 
fairly large number of home users, P-C users, most power users using it 
and we appreciate them and that's a very important group to us.



What we're planning to do now with warp server is to move into the small 
and medium sized business market segments. We've recruited 5000 business 
partners to help us. When Merlin comes out, Merlin will be sold for the 
first time, not just in large corporations, but to departmental users, 
medium size businesses, small businesses, the so ho market segment or 
small office, home office market segment, as well as a sub set of the 
home users. That sub set of home users are people who do work at home. 
Not just entertainment. But people who want to be able to access infor
mation in the office or work as a remote office. In fact there's a lot 
of mobile features going into Merlin so that it is an ideal mobile 
office. So that you can, through a network, I-S-D-N or analog, link into 
a remote network as if you were attached directly to the token ring or 
ethernet and have all access to the services. And by the way when you 
separate, and you go off traveling in an airplane or train or bus, and 
your changing data in files, and back in your office they're changing 
the data in the file servers. When you connect back in, the files will 
synchronize by themselves. So there's mobile file synchronization capa
bilities going into.



So we're focusing on those market segments. As we go to those market 
segments, O-S 2 has very strong position in an area I call tools. Of the 
2500 applications, there's probably about 800 who fall into category of 
tools. Tools for developing applications, tools for running and managing 
a large complex network. The second category is there is a large number 
of industry specific applications. Teller appl



ications, casualty claim applications, graveyard management applica
tions. The applications that you don't normally find on Egghead shelves 
or a CompUSA shelves. But they sell a lot. And that's where O-S 2 is 
being used a lot. The third category is group ware. If you look at all 
the major group ware products, whether it's Notes or C-C mail or group
wise. Even Microsoft mail, they're on O-S 2. Or the major database 
servers with Oracle and Cy-base and D-B 2 on O-S 2. So in the group ware 
category there's a lot of strength in O-S 2.



The area we're criticized, is a fourth category and fifth category and 
that's personal productivity and edutainment or education and enter
tainment. And there our strategy has been run the existing dos and 
windows applications because we needed a large enough install base to 
track I-S-V's. We made two key changes to our strategy. One is that, we 
thought we would not compete with any vendors and just get O-S 2 out 
there, provide support and great applications will come. Well that 
didn't happen. And we tend to think that the Microsoft model, as much as 
some of us don't like it, is really the one that probably works more 
closely in that you need a great office suite and a great core set of 
applications. And if you build those, people are going to buy the oper
ating system along with these industry specific. And when there's enough 
all the other vendors will start coming in. So that's our strategy.



Now as part of the reason why we bought Lotus, we spent $3.52 billion 
dollars, to get Notes and SmartSuite and all the Lotus applications are 
going on O-S 2 and windows. But we're working closely with other vendors 
to make sure that they're applications in key categories. The office 
suite being one of the most important. SmartSuite will be out in mid 
year. And then it will be updated with SmartSuite 97, where we'll get 
full parity with what's on windows.



Star division from Munich, Germany has announced that they're going to 
be bringing Star office into North America, I heard it was going to be 
Jun





e. They have a clone of Microsoft office, it's just smaller and faster, 
but functionally it looks like the same product.



The leading office automation in Europe is I-C-L's Teamware. Client and 
server on O-S 2. They just opened up offices in the last several months 
in Santa Clara where Teamware is now being brought into the United 
States so we get three very powerful applications as well as applica
tions for a variety of other vendors that we think are very important. 
So we're targeting now, specific applications.



The other one is on Lotus. We think Notes is very critical to us. We 
made a decision, the Lotus products are our strategic products for the 
office. Notes is our strategic E mail document management product, in 
fact Notes 4 has the C-C mail interface built into it for E mail, which 
is a dramatic improvement over the version 3 Notes mail. We declare that 
SmartSuite is our office desktop standard. And we stopped projects in 
I-B-M that would build any competing technology. And we provided all 
that technology to Lotus and they're integrating it in because what 
I-B-M had, we've not been well known for award winning user interface 
technology. Lotus has with Organizer and Word Pro and other products. 
We've been pretty well known for building great plumbing, but a lot of 
end users don't buy the plumbing, they'd rather rely upon vars and sys
tem integrators to put it in.



What we're building, for example, with organizer technology so it links 
into the network and you can organize calendars and schedules in a net
work environment. And we bring that capability.



Approach is a great desktop database with ease of use. We provide great 
plumbing to hook it into Oracle's database and I-B-M's D-B 2 and Cy-
base's database. So it's a great front end for it. And other technology 
is being linked in. We built in technology plumbing that solve problems 
in Lotus Notes. One problem is, Notes had trouble scaling up. But with 
I-B-M's M-Q series technology some asynchronus messaging, we can now 
scale up to very, very large networks 



and we've got that into the Notes technology as well as I-B-M Internet 
technology.



And I just want to leave you with the message that we are committed to 
being open in cross platform. On O-S 2 and A-I-X on the desktop, and 
with Notes across a wide range of products. Not just I-B-M's operating 
systems. Yes sir.



(question from audience. ... I-B-M doesn't really support the I-S-V ... 
Microsoft does. ...)



We absolutely believe that people don't buy O-S 2 warp just because it 
has a fantastic solitaire game in it. And one of the best C prompts in 
the industry. But that it is applications. And that that drives the 
business. What we're doing is, right now we're focusing most of our 
energy on a select set of vendors where our customers for the markets 
that we're going after, say these are important. Not only for today's 
market that we've been going after, but also for the market segments I 
described earlier. But it's not supporting all of them equally any 
longer. So we made two key decisions. One we're going to compete with 
some of them, particularly in the office automation area. We're going to 
make sure there's a great set of office suites out there. Secondly, is 
we're going to select then a set of vendors who have the key applica
tions rather than just let the flowers bloom everywhere. Now we 
recognize that we've got areas that are weak in support like technical 
support provided electronically over CompuServe or Internet. And we're 
going to fix that. And we are going to invest more... (END OF SIDE 2, 
TAPE 2)



PAGE65









] -@0 -n_O@0 - yy1

