How Microsoft Thinks
Tuesday, November 3, 1998 by Dave Winer.
I just finished reading the "Halloween Memo" which got a lot of play on the net yesterday, along with comments from Eric S. Raymond.
I found Raymond's comments distracting as I was trying to understand what the memo was saying, so I wrote a little script that removed Raymond's comments, and posted the result here:
If you are a user of Microsoft software, or a developer who builds on their system software and/or apps, I highly recommend reading this piece. I believe it's real. If so, it provides a glimpse into the thinking process at Microsoft.
It's about more than open software, there's probably a similar paper inside Microsoft on the Department of Justice and one on IBM, and quite a few, over the last few years on the web and Java.
Here are my notes after reading the piece.
As you might imagine, Microsoft is looking to embrace the ideas of the open source world, and is thinking if it will diminish their role as the dominant operating system software company.
I agree with their major conclusion that right now open source software is powerful on the server, and not in strong position on the desktop, on client machines or on content workstations.
They note IBM's recent increased involvement with the Apache web server. They also note a falloff in interest on the Mozilla mail lists, which they archive internally. A reminder to people who believe Microsoft isn't watching and listening, it's better to assume they are.
The piece talks openly about FUD, Fear Uncertainty and Doubt, and acknowledges that the leading open source products, Linux, Apache and Perl, aren't vulnerable to FUD, since their continued existence isn't in doubt.
One of the most interesting sections is towards the end, talking about ways of blunting attacks from the open source world. While XML isn't mentioned, it's behind the discussion on WebDAV, with a prediction that the Apache organization won't pick the right way to invest in WebDAV, and presumably this will provide an advantage to Microsoft, server-side.
I find this very interesting because my company has been trawling on the edge of WebDAV, experiencing the same kind of confusion. It may be that WebDAV is just plain confusing, equally so to the people at Microsoft. Or there could be an ace up their sleeve that we can't see.
There's no mention of XML-RPC, but it does talk about MSMQ for Distributed Applications. What is that? I don't know.
To understand this memo and to understand Microsoft, remember that this is a very big company with thousands of employees. Unless they're organized by an outside threat, they often move slowly and in a confusing way, both internally and to people outside Microsoft.
This memo represents one slice of the thinking of this organization, both in time and in people. It's almost certain now, three months later, that Microsoft has a different focus, a different set of priorities and a different set of issues they're working on.
I see it differently from Raymond, who polarizes things as open source versus closed source. I think, as Microsoft does, that many of the supposed discoveries of open source are actually discoveries that apply to all software, independent of their economic model. In fact, what's really going on is the opportunity of the Internet to displace centralized software companies, like Microsoft, is being wasted because of insignficant lines between non-Microsoft development organizations.
Earlier this year I made a bet with some Microsoft people that before long they would be releasing source code as a normal part of their development and release process. They said it will never happen. In the memo it's clear that they're now considering releasing parts of the Windows source code, for example, the TCP/IP stack. Although people on both sides of this fence doubt it will ever happen, I remain sure that Microsoft will adapt and in some ways match Linux, Apache and Perl, and will probably embrace open source to a greater extent than anyone now thinks possible.
And that's the way it will work for all software companies. The ones that are starting around SendMail, Linux, Perl and Tcl (what else?) will release proprietary software in binary form (many already have) and Microsoft, Adobe and Symantec (and UserLand) will release parts of their technology base in source form. This is a transition that began a long time ago, and it's accelerating.
Microsoft has had recent great success with its Embrace and Extend strategy in its competition with Netscape and Sun. It's virtually certain that they will respond to open source with the same tactic, that they won't resist open source, they will embrace the idea and then take it where they want it to go.
For the first time, this memo reveals the process at work, it's visible and it's pure Microsoft. If the memo isn't real, it was written by someone with a very deep understanding of how Microsoft works and thinks.