-----BEGIN PGP SIGNED MESSAGE-----
From: Bill Gates
Sent: Wednesday, February 19, 1997 9:46 AM
To: Jim Allchin (Exchange)
CC: Nathan Myhrvold; Paul Maritz
Subject: RE: "Losing a Franchise - The Microsoft Windows Story" (a new Harvard
I agree that making sure applications are primarily on Windows is something we
have lost site of.
I rail against the people who want to just give things away like DirectX. I
think exactly the same thoughts in your message.
However I think the problem is much less what we are doing than what we are not
I can cancel every give away and we would still have a major problem.
We need a vision of what a Windows application looks like - for example how can
it sometimes run on the server and sometimes on the client. What will we have
that the Java Runtime will not have?
Cross-platform demand is not coming from statistics. It is coming from the
free-lunch syndrome we have allowed to develop. All of a sudden people think
there is no drawback to being cross platform. No drawback in size, speed,
interface, richness, testability. To some degree this is true because
machines have enough memory now that a "duplicate runtime" is not
We should have people laughing at the idea of 100% pure Java whether they write
in JAVA or not. However we have nothing along these lines. The fact that
their runtime is changing, will keep changing, will be subsetted on some
machines, will have to make someone money, will have to choose UI, etc...etc
is just lost because we are not there driving a positive agenda for Windows.
The fact is that applications can be run on the server against an HTML client.
I would prefer this to be a Citrix like client but that is not the key issue.
Most applications will have very little client code in the future. For
example Federal Express giving you their package status.
The fact is there will be lots of machines where HTML/some level of Java is all
they will have in common. Cheap devices and old PCs will be like this. I
makes it very easy for people to think they should just program to this.
Our installed base not migrating is a major drag on our ability to promote
Lets work together to find the solution to this. It's critical. I can say I
am more scared than you are but that is not what will help us figure out where
we should go.
- ---Original Message---
From: Jim Allchin (Exchange)
Sent: Tuesday, February 18, 1997 5:17 PM
To: Bill Gates
CC: Paul Maritz
Subject: "Losing a Franchise - The Microsoft Windows Story" (a new Harvard
I'm sure this subject got your attention. It's what I worry about every day
when I shower, run, eat, etc. Paul knows how worried I am, but I don't think
I've told you before.
We all know we have many challenges. I think about the challenges however in
two buckets. First we have challenges that we have a unified strategy for and
perhaps an implementation plan underway. Some of these are big issues, like
TCO, but they are much less scary to me than the other bucket. You can say we
aren't doing things fast enough for TCO or marketing hard enough against the
NC (and I'll agree), but our strategy and path is clear. We know that if we
can execute fast enough we can stop this disease. The second set of
challenges are the ones where we do not have agreement on our strategy within
the company and the company is often working cross-wise internally. The cross
platform vision and keeping Windows as the platform and the center of
innovation fall into this category. In my opinion Windows in the process of
being exterminated here at Microsoft.
1. Why are we doing so many things cross-platform? Are there more Macs, OS/2
or Unix clients today as a % or less than a year ago?
I assume the argument is that we have to do things cross-platform because
Netscape is (or says they will). So, we move our innovations cross-platform
and dillute Windows. The alternative is to say "NO" and push even harder on
Windows. I know that we have to do some things cross-platform. But our
default today is to assume functionality needs to go cross-platform instead of
assuming it doesn't (and then later reluctantly moving it
- -------PAGE 2-------
when necessary). I consider this cross-platform issue a disease within
Microsoft. On our current path, IE 4 will not be very integrated into
Windows. The IE team is not focused on this problem and I was requested to
shut down my UI/shell team. Windows will get what the other platforms get
UNLESS the IE team finds out they can't do it easily on the Mac, etc. This
is the wrong approach. We should be asking for specific innovations to be
restricted to Windows. I can't fight this disease alone. The problem is the
company is not unified on the strategy.
I am convinced the path we're on is the wrong one. We are playing into
Netscape's strengths and against our own. I hear lots of words about how the
software will be "better" on Windows because we have more people working on
Windows, but I can't sell abstract statements like this. We focus attention
on the browser battle where we have little marketshare instead of focusing on
the battle at integrating things into Windows where we have marketshare and a
great distribution channel. When IE 4 first was discussed we
were "integrating the browser into Windows". That is what we told everyone.
That was a strong message for Windows. That message is now gone since IE is
going onto all platforms. It won't be as "integraded" (whatever they [sic]
means technically), but all the words about WebPC and the like convince me we
are determined to put a gun to our head and pull the trigger.
I se the same pattern here as with Novell a few years ago. Some people
believed we should drop our work in TCP/IP and only do only [sic] IPX/SpX
work. It took significant effort in order to convince the PSD team to accept
TCP for Windows 95. Why? Because we were in copy mode of Novell. We are
doing it again. There is a time for this clone strategy, but the better long
term approach is always to attack from a more strategic perspective.
2. The platform is Windows isn't it? duh.... it would seem obvious. (But is
it a browser/IE? or maybe Outlook?)
We are in a head-head competition with Netscape. They claim the platform is
Communicator, etc. so that's why we say IE is the platform. However that
fights against our strength and plays to their strength. We are marketing IE
as a platform. That is a mistake. The whole "WebPC" concept will totally
confuse the platform story. This is another nail in the coffin for Windows.
No amount of marketing will fix this. The meeting I had with the PR team
convinced me of this today.
When Paul asks me what's innovative for developers in Windows I really
struggle. I finally have a team working on this in the areas I control but it
isn't enough. There are two things that are really critical for code running
at the client: UI and storage management. We are working on articulating
directory access, multimedia (whatever remains non-cross platfrom), etc, but
UI and storage are the key pieces. We have done a bad job with storage
innovation. But, we appear to be just giving up on the UI since it's all
going to go cross-platform. I don't support this. I believe smart people
here could find ways to do things beyond what AWT or IFC can do if we tie it
more into Windows. Remember these class libraries are layered. For example,
no one is working on integrating Trident into the OS in a fundamental way.
Maybe this won't amount to much but it's the type of investigation we have to
do. (Actually, I think there will be many gains in terms of performance and
consistency throughout the system). We should move as little cross-platform
as possible. Without a specific UI focus on Windows I think we are in series
[sic] trouble. Davidcol agrees he can't support the OS changes we need and he
is suggesting that we create a team again. This is not enough if the strategy
isn't synchronized - both marketing and development-wise.
I'm available to talk about any of this. On our current path, I just don't
feel that Windows can win. Given the ecec retreat I thought I would send this
to you for your thoughts.
ATTORNEY'S EYES ONLY
MS7 011793 CONFIDENTIAL
MSS 0218906 CONFIDENTIAL
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
-----END PGP SIGNATURE-----