Responses to Tim O'Reilly's "Open Letter to Microsoft"

Note: You can find Tim's Open Letter to Microsoft here.


Thank you for forwarding that Open Letter. While I doubt that Microsoft will be very interested in Tim's advice, it did cause me to read the Halloween documents in full, and to think about the implications.

The history of computing tells us that proprietary standards favor the creator but only when enough people are willing to adopt them. Consider IBM's MVS, DEC's VAX VMS, the IBM PC, Apple's Macintosh and the IBM PS/2.

The openness of the IBM PC architecture and Microsoft's willingness to license MS-DOS to any PC maker were fundamental in the adoption of the PC architecture rather than any of the myriad previous microprocessor architectures. In the short term, Apple did well out of the proprietary nature of its Macintosh architecture, but, in the long term, we now see that a more open stategy might have served it better.

Of course, where the market leader has a commanding technological or marketing advantage over its competitors, or where competition is only with proprietary systems, the proprietary system can win. Good examples are MS-DOS, Windows and Intel's processors. Yet, where the advantage is less clear (PS/2, Compuserve and perhaps NT Server or Intel's Slot 1 architecture) a proprietary architecture could be its downfall. Unix illustrates both points: its open nature led to widespread adoption but the proprietary differences between implementations have held it back.

As it happens, just after reading the Halloween papers, I have been researching the widespread problems with the Windows TCP/IP stack. Clearly Microsoft would have done a lot better to have simply released open source code of service and interface routines like this: it would have produced bug-free code at no cost to them and any resulting improvements in Windows performance could only have been to their advantage!

The only possible downside could be if so much code were released that it would make it significantly easier for competitors to bring out a Windows-compatible rival. Since there are already open source implementations of TCP/IP stacks, device drivers etc. this shouldn't be a concern. Ironically, as the examples of OS/2 or Sun's WABI shows us, the only rival likely to be able to muster sufficient resources, and to have a chance of securing enough market share to survive, would be an open source Windows API binary implementation.

Paul Mullen, Computer Shopper UK


Perhaps it is truer than we thought: absolute power corrupts absolutely.

And this may explain why monopolistic organizations seem to do 'stupid' things. Surely you can think of many examples:

  • IBM & PCs (when IBM made the PS/2)

  • AT&T and long distance (inspite of lowering costs they didn't want to lower prices)

  • GM and its car designs (went from 70% market share down to <30%)

Brandon Fouts

Dear Tim,

I strongly agree with your comments about the "Halloween Memo", the books you publish enligthen and focus my work on the area of open source support and development using this technologies.

I am worried about the ways Microsoft cares about their customers and the low levels of quality on their products.

It is amazing how the open source software community produce production strength software and the effort your company has on giving the knowledge tools to work with OSS.

Keep up the good work, and thanks to spoke in the name of an unheard community that follows any new title you publish.

Your task is enormous, but when you think it is not worth the amount of effort you put, think again, people all over the world belive on the work you and your company is doing.

At this point I have 40 of the titles Oreilly has published and I thank you for the tools you have given us to compete in this changing world.

Whenever you think on putting this books in spanish please let me know I can help.

Truly yours,

David A. Trevino Rodriguez
Technical Support/Soporte Tecnico
CITI/TRIT (Technological Research on IT Corporation)

Right on!!!!! You are right on the beam with respect to the process that brought the internet to success. You and I are absolutely positively aligned on this one. Design and decree is what I have been fighting for 15 years.

When you get a chance (in betweenanswering all those feedback letters) check out the feedback page I posted resulting from the Chicago Tribune/David Strom editorial

Particularly the first part. Note that it's a few days old.

Bob Denny


Your open letter responding to Microsoft's "Halloween Document" about attempting to "embrace and extend" the open source community was very good in both intent and execution.

Rather than heckle Microsoft or even explain how Redmond could learn something about robust and powerful software from the GNU crew, you made an argument that ought to appeal to the Microsoft's informed self-interest. You even avoided the joke about how Bill Gates can't afford to crush his competition because he'd have no one left to steal from (though your letter does dance close to that line).

Similarly, I appreciate the spirit of constructive criticism in the Annoyances series. Well done all around.

Dave Demko

Tim's open letter puts an interesting spin on the Microsoft memo. Of course, the implications in the letter (e.g., that Microsoft's R&D has produced -- and, really, is capable of producing -- nothing of more value than "Microsoft Bob") may be a bit unpalatable to Bill and his friends. But I guess that's Microsoft's problem (:-).

I have seen several articles which discuss the fact that Microsoft is hiring large numbers of leading computer scientists. In fact, the company has been equated to kind of an intellectual "roach motel": scientists go in; nothing useful comes out. Microsoft has also been accused of distinctly predatory hiring practices.

The Halloween memo's suggestion on hiring OSS leaders is clearly in line with this analysis. The interesting thing, of course, is that Microsoft doesn't NEED to get anything good out of these folks, as long as they aren't working for the competition. Of course, there's always the danger that a scientist may eventually get so frustrated and bored that he will smuggle useful code out onto the Internet!

As Eric Raymond notes, however, the key issue is Microsoft's abuse of standards. If Microsoft is allowed to poison publicly available and useful standards with proprietary and obscure "features", we all will deserve the predictable results!

Rich Morin, President
Prime Time Freeware

Dear Tim,

I'm sure this note will be lost in the mass of email you must receive daily, but I want to take the opportunity to express my appreciation for your "Open Letter to Microsoft" posted today.

There are few in your position with the courage to speak the truth.

I am an O'Reilly book fanatic; the only books I haven't read are the ones still on order at Borders. Keep up the good work!


Brad Nelson
Software Dev. Mgr.
University of Kansas/SPED/SCR*TEC

Hi Tim,

I applaud your open letter to Microsoft. It is a very sensible document and I hope they see it as good advice rather than as an attempt to capitalize on their compromised position.

I don't think you went far enough. You should have advised Microsoft to release the source code of Windows NT, Internet Explorer and the Office Suite. What other chance do they have? :-)

Lincoln D. Stein

Thanks, Mr. O'Reilly, for being a voice sanity and clarity within this massive swirl of disinformation, greed, and zealotry.


John Kordyback
Senior Systems Analyst

P.S. I will continue to purchase ORA computer books for the very same reason. Currently, I have two ORA books on my bedside table, three open on my desk at work, and a couple of dozen on various book shelves. Needless to say, quality still counts.


Thanks again for your stance and leadership. And thanks for mentioning SGI.

I promise I'll do everything in my powers within SGI to exceed all your hopes regarding our support of free software.

Peace, Ariel