The Halloween Document

TheHalloweenDocument is a startling analysis by EricRaymond on an e-mail that was sent around Microsoft's inner circle in an attempt to point out that de-commoditizing the market will allow them to snuff out Linux.

The original e-mail was sent by a Microsoft viper named Vinod Valloppillil on August 11 1998. It scares and horrifies people naive to the common workings of competition over business markets.

http://www.opensource.org/halloween/

David vs. Goliath, or the Bazaar vs. the Cathedral...

Order vs. Chaos, Anarchy vs. Empire, Bre'er Rabbit vs. The Tar Baby ... Celebrity Deathmatch was never like this.

Also cited in TheHalloweenDocument is open-source development vs. closed source development. Eric warns folks to expect "Microsoft's propaganda machine to be directed against the process and culture of OpenSource, rather than specific competitors. Brace for it..." The real big question is whether the Microsoft juggernaut can adapt and overcome an entirely different paradigm with its marketing machine like it has in the past with competitors (insert the usual laundry list of casualties here...).

Order vs. Chaos, Anarchy vs. Empire, Bre'er Rabbit vs. The Tar Baby ... Celebrity Deathmatch was never like this.

[AnonymousDonor retracted his comments at this point.]

How can you say that it's "free" development? Linux is developed by programmers working at real for-profit companies, or at schools and research institutions supported by my tax dollars. Someone must pay for the roofs over their heads. -- another AnonymousDonor

OK, I'll bite... Of course this software isn't free (pace Stallman), but its economics are very different from the conventional vendor model. Microsoft people are not stupid but they do have tremendous resources behind them, and some people believe that these are not always deployed for the greater good. My personal experience as to the relative solidity of NT and Linux (one of the packaged versions) differs from yours, YMMV. For myself, I believe that if NT had 1) been cheaper and 2) included more of the underlying features that we old Unix hands had found so useful (e.g. a hierarchical file system that spanned servers) then Linux might have become less successful than it appears to be. --SteveFreeman

Is Wiki free? By your logic, if I wasn't employed by a REAL company, I wouldn't be writing this right now or spending the time I've spent on other Wiki-based projects I'm involved with. And when you say "a company rationally responding to a challenge," what challenge? To write better software or to monopolize the marketshare? I feel like you're comparing apples and oranges here, and Eric's point in his annotations is that Microsoft will need to challenge a process (open software), not a product.

EricRaymond's ultimate concern is Microsoft's quote "...we raise the bar and change the rules of the game" and other comments that imply a strategy of de-commoditizing the industry (see "Beating Linux" in the document). That means

Don't get me wrong -- I'm not anti-Microsoft. My entire professional career is based on using their products & technologies, I worked at Microsoft twice, and I too am (or was) acquainted with a few people on the author's list. However, if Microsoft monopolized innovation, would we have the world-wide web, the Internet, and other innovations that have come from outside Microsoft?

Forces in the industry caused Microsoft to perform an about-face and take the Internet as a serious threat. This caused subsequent innovation at Microsoft (and I'm sure there were memos flying around like the one TheHalloweenDocument annotates), but this innovation did not originate from them. I hope to wait a long time for the day when I need to download service pack 5 of Visual Perl, or read a memo describing a strategy to "raise the bar & change the rules of the game" in something as commodity-driven as hog futures. --PhilipEskelin

I'm so confused. Of course Microsoft looks at things like Linux to see what they can learn from them. And of course they want to make the world a better place for Microsoft - it's their job. As I read the Halloween document, I saw a rather rational analysis of what is going on, flawed as are all human works, analyzed by someone with even more of an ax to grind than has Microsoft. What makes Microsoft so successful is that they do learn, and their stuff is "good enough". I do wish they didn't pressure their big customers in some of the ways they do, and if that's illegal, the law should stop them. And, quite likely, will. --RonJeffries

I learned a lot about GNU and Linux from the Halloween Documents. But, of course, so did Mr. Gates. --PhlIp (And BTW my home and work platforms are now 100% Linux, thank you...;)

[Someone signing "AuthorSquelched" relates his recent positive experiences with Linux, while NT was churning his box for no apparent reason. Later, he retracted his remarks.]

Interesting, Mr. Squelched. I think you've just demonstrated what many Linux advocates (at least, the sane ones that I pay attention to --- how's that for narrowing the field to exactly the subset I want to talk about? :-) have been saying: it's not that Linux is already better than NT, except in some narrowly defined senses of kernel size or speed or code quality or whatnot. The issue is that the development process is better. The argument is that open-source has numerous advantages over more traditional methods for producing quality software that meets users' needs in a short span of time.

I'm not 100% convinced of the universal validity of that argument. But something is clearly going on with Linux.

And (to answer another of your points) while it's true that someone is paying for all of this, the cost is certainly spread around quite differently, and so far I like the change. (I have been a contributor to a few open-source projects here and there, so I'm not just benefitting from others' largesse.)

Too bad all of your applications still look like Windows, though. (And if you ever figure out why NT churns away at the disk all the time, please let me know.) --GlennVanderburg

I want them to look just like Windows. That is what makes KDE and ultimately Linux useful for me and for others. In terms of applications, it still needs many more like a Quicken to be a serious platform.

I know. I was just being silly.

If you ever figure out why NT churns away at the disk all the time, please let me know.

Could be FindFast? if you have Office installed. Look on the Processes tab in TaskManager? and sort on CPU while it's happening. -- PhilGoodwin

No, but it is Windows 2000 and is still a beta. There may still be debug code in it. The process that continues to run (in Task Manager sort) is called ntfsfr, which may be an NTFS disk defragmenter. I have to look into it. I have Office 2000 on it. I'll have to check about Find Fast.

If it is a defragmenter and you're a developer that could be trouble. Large builds write tons of files most of which are at least partially fragmented (at least the defragmenter I worked with thought so). That plays havoc with the defragementer of course. If does turn out to be the defragmenter don't turn it off completely if you can help it, those large builds really do fragment the heck out of your drive, but set it to run once a week at a time when you are not using the machine.

Read what KenThompson had to say about Linux (ouch! and he should know!): Computer recently (date?) visited Ken Thompson (co-creator of UNIX) at Lucent's Bell Labs to learn about Thompson's early work on Unix and his more recent work in distributed computing.

Thompson: I view Linux as something that's not Microsoft's backlash against Microsoft, no more and no less. I don't think it will be very successful in the long run. I've looked at the source and there are pieces that are good and pieces that are not. A whole bunch of random people have contributed to this source, and the quality varies drastically. My experience and some of my friends' experience is that Linux is quite unreliable. Microsoft is really unreliable but Linux is worse...

Some people suggest that this is sour grapes that Linux is a zillion times more popular than PlanNine. A more moderate view is that Linux is useful and practical, whereas PlanNine is just interesting. From a research point of view it's not remarkable, just a good implementation of a traditional design.

The quote makes facts sound like criticism: every system arises in response to the prevailing situations (such as Microsoft); nothing lasts forever; every system has some parts that are better than others; every project has some programmers that are better than others.