You worked at Microsoft for ten years, then left the company two-and-a-half years ago. From your perspective, do you think Microsoft has fundamentally changed as a result of the antitrust lawsuit?
My short answer would be "No".
There were many positive things about the Microsoft work environment. But there were some negatives. People use the term "enabling environment" to mean a situation that encourages someone to act in a negative way, such as drinking alcohol heavily, by mitigating the negative impact of the behavior, and providing tacit approval for it. Well, Microsoft constructed an enabling environment for socially obnoxious behavior: it was welcomed and rationalized into positives. If you were late for meetings it meant you were busy doing important work, if you were extremely confrontational it meant you were passionate about your job, if you required subordinates to work long hours it meant you were committed to the product, if you turned down everyone you interviewed it meant you weren't soft, and so on.
So Microsoft had this system that encouraged and rewarded people who acted a certain way. And some of that behavior trickled out into meetings with customers and partners, where they were correctly seen as negatives and helped foster the anti-Microsoft attitude. But since Microsoft kept hiring and promoting obnoxious people, they kept being obnoxious.
Furthermore, some Microsoft groups have always had a bit of a bias against having too many dependencies. The idea was get it done and don't worry about other people in the company. That's why different applications wrote their own Print dialogs, and why networking between Windows NT and Windows 9x isn't a seamless as it could be. That attitude helped shipped some products sooner, but it permeated outside the company too. Now Microsoft has pushed profit and loss responsibility further down in the company, so the groups within the company should become even less likely to collaborate. So how will that increase the prospects for collaboration outside the company?
Asking if Microsoft can really change is the same as asking if its people can change, and I don't think people can change. Maybe Steve Ballmer can soften his public statements. But the rank and file will keep acting in the manner that has made them successful. It's possible that the hardass people will gradually leave the company and be replaced by kindler, gentler types, but it would take a long time.
And it's not clear that Microsoft would want to change. There have certainly been massive speed bumps on Microsoft's road to success, but the company seems to have overcome them. So the impetus to change may not be there.
But don't people at Microsoft realize that the company is a convicted monopolist?
Actually, I really don't think they have internalized that. The distinction is still drawn between having been found guilty, and actually being guilty. It is easy to blame the initial judgment handed down by Judge Jackson on a bad legal defense strategy and a biased judge. So I don't think people view the final settlement as dodging a bullet, but rather an affirmation that the previous behavior really wasn't so bad. The feeling is that whatever was done was judged to be wrong only because the company was found to be a monopoly, and since that finding of being a monopoly wasn't made until the actual trial, people had no way of knowing what they should or shouldn't have done.
Plus, I think Microsoft people view the actions that got them in trouble as the work of just a few people. Call them bad apples, or overeager employees, but in either case it is not viewed as a pattern of misbehavior that came from the top executives. When you see Disney strong-arming a competitor, you get the feeling that Michael Eisner has personally approved it and a platoon of lawyers has gone over it. When you see Microsoft strong-arming a competitor, somehow it comes across more as a couple of 23-year-old hotshots acting clever. A lot of people may view Microsoft as a single borg-like entity, but of course it's not, it's a group of 50,000 people each trying to do their job as best they see fit.
Anyway since the vast majority of Microsoft employees were not actively involved in negotiating OEM licensing deals, meeting with Netscape, or any of the other activities that came under scrutiny, there is not anything they personally could have done differently, so there's not any reason for them to act differently in the future.
You say there was no pattern of misbehavior. But certainly executives were involved. There is a famous quote where a Microsoft executive threatened to "cut off Netscape's air supply."
The prominence that is given to that quote is somewhat baffling. I suppose it's an unusual and somewhat graphic metaphor. But is it really worse than saying you are going to kill your competitor, which is a fairly typical comment in a business environment? Or that you want its executives' children to grow up poor? What do people think that Sun and Oracle say about Microsoft in private?
Or perhaps people are thinking about the meaning behind the quote, that Microsoft was trying to take away Netscape's revenue and put them out of business. Isn't that how businesses act? I mean, are other US airlines sitting there thinking, "Gosh, we must be bad companies, we competed with United Airlines and now they are in bankruptcy?" Of course not.
Right, but the other airlines are not a monopoly.
Sure, but when that comment was allegedly made, nobody knew that five years later Microsoft was going to be found in court to be a monopoly. I should point out that I don't know for certain that the quote is accurate, but given how widely it has been reported, I see no reason to doubt it.
OK, enough about the antitrust suit. What about Microsoft's future plans. The company is involved in a lot of side projects. Do you think they will all pan out?
Personally I dispute the notion that the personal computer is dead, and the future will be small customized devices. Microsoft is trying to cover all the bases by continuing to work on Windows, while also getting into all the other devices.
Many people consider Microsoft to be this all-powerful entity that always gets its way and succeeds in whatever it attempts (whether through technical ability, luck, marketing, or illegal means). But in fact, if you look back at all the initiatives Microsoft has attempted, most of them have failed. Remember Windows at Work, an attempt to unify all office machines on a network? Or TrueImage, a clone of the PostScript page description language? Or ACE, the Advanced Computing Environment, designed to replace the Intel PC architecture? If you look at all the things that Microsoft has boldly announced with a gaggle of industry partners, the success rate is maybe 10 to 20 percent. So pick any ten things that Microsoft is attempting now: Auto PC, cell phone operating system, set-top box, Xbox, electronic books, Internet Gaming Zone, Tablet PC, Pocket PC, speech recognition, MSN. The odds are that only one or two of those will actually succeed.
Not that this means Microsoft is making a mistake doing all those things. I'm sure the same debate being played out in the industry now, personal computers vs. smaller devices, is going on within Microsoft. So Microsoft can hedge its bets and see what really becomes important. I mentioned the 10-20% success rate. But I think the ones that failed basically did so because the industry wasn't ready for them or there was no competitive advantage in Microsoft pursuing them. Microsoft rarely abandons a market just because of strong competitors. So if one of those ten I listed above becomes a major industry with active competition, expect Microsoft to keep working away at it.
The only thing I worry about with Microsoft is the reason that some of these initiatives get off the ground. Microsoft is an interesting place because a lot of oldtimers have enough money that they don't have to work. Rather than retire, they kept working at Microsoft because they enjoyed it. But eventually many of them decided that it would be more enjoyable to be working on a fun, speculative, cutting-edge project rather than something dull and profitable like Office. So they went off and came up with some fancy project and then presented it to Bill Gates, and what is Bill going to say to these people who have served him so well for so long?
When I interviewed to work on an interactive television project that Microsoft was doing back in 1994, one of the guys in charge had been working on compilers and he decided he wanted to work on something trendier. He told me, "I want to work on the sizzle, not the steak." The project was basically hopeless from the start, but Microsoft worked on it for a few years, spent a lot of money, had many employees working on it instead of something else, signed a bunch of deals with partners, and eventually the whole thing just went away.
The combination of people hyping projects because they think they are cool, and executives worried about missing the next big thing, results in Microsoft funding some dubious projects. For example, one of the critical upcoming tasks for Microsoft is making the next version of Office compelling enough and useful enough that people actually want to upgrade to it. Yet within Microsoft working on something like that is generally considered boring, while helping Xbox lose $100 million or whatever is seen as cool.
How can you criticize employees who work on projects like that, when you yourself left Windows NT to work on an interactive television project?
It's true, and then I worked at Softimage for a while. I spent about two-and-a-half years away from the Windows NT team and never shipped a single product. But after that I went back to NT and spent three more years working on Windows 2000 and XP. People there used to refer to the gap as my sabbatical.
You mentioned electronic books. Do you think e-books will succeed?
Not for a while anyway. I am shocked when people claim that reading an e-book is equivalent to reading a printed book. It's a vastly different experience, just like reading a news story online is quite different from reading it in a printed newspaper. Now there may be some advantages to the electronic versions, but you can't deny that it is very different also.
Maybe at some future time when we have flexible displays with a resolution approaching that of printed pages, e-books might work for some limited uses. I picture something that is the same size and weight as a current bound book, and has 300 pages or whatever, except each "page" is actually a display that can be updated, so the entire contents of the book can be changed. There is something about being able to flip pages in a book that must be preserved.
E-books also seem to be one of the battlegrounds for Digital Rights Management, so I have a philosophical opposition to them right now.
The fact that Microsoft is working on Digital Rights Management software leads some observers to place them firmly in the "evil" camp. Is this fair?
I think Microsoft the platform company is generally agnostic about issues like this. Many people are talking about DRM, so maybe there will be money to be made providing DRM software. In a situation like that, Microsoft is always going to be investigating it and working on software to support it, in case it pans out and becomes important. If it goes away due to consumer revolt or new laws or whatever, then I think Microsoft will move on to the next thing.
Meanwhile Microsoft the content company I think has been pretty quiet on DRM issues. I don't think they feel this is a great time to be collecting arrow wounds leading the charge on an issue like that.
There was a recent article in the Seattle Post-Intelligencer about Longhorn (the next version of Windows), describing it as a "radical change" from Windows XP. What do you think?
The only real evidence presented of a "radical change" for Longhorn is the claim that "Its guts will also be radically different from Windows XP's, because they're based on XML". That sounds impressive, but I'm not sure what exactly it means. The internals of Longhorn will be based on the same NT kernel that XP was, and that will continue to be primarily written in C.
Longhorn is also supposed to have a new file system, which I predict the average user won't know or care about, and also a new unified way of interacting with data. The article states, "Obviously this means a thorough overhaul of not just Windows but also the Office software suite." To me, it seems more like a thorough overhaul of Office only.
In fact it looks like Longhorn might be a fairly minor change to the bulk of the operating system, just as XP was a fairly minor change from Windows 2000. This would actually be a very good thing, since it means Microsoft won't have as much new code to shake the bugs out of.
Microsoft is criticized for bundling applications with Windows XP, such as Media Player or Messenger. Do you think it should be allowed to do so?
I think it should be allowed to do so, to the extent that the antitrust settlement allows it to. But I'm not sure why Microsoft wants to do so. Look at Media Player. Microsoft has spent a lot of money developing it, and has never charged a penny for it. Furthermore the company has risked serious legal trouble to include it with the operating system. And I can't really see what the point is. Is there some huge advantage in having people use your media player instead of Real's? To me it would make much more sense for Microsoft to simply bundle Real Player, AOL Instant Messenger, and so on with Windows. It's cheaper, it avoids legal headaches, and everyone is still going to pay for a copy of Windows with (almost) every PC they buy.
I think Microsoft had a moment of panic where it thought that AOL or Real was going to commoditize the operating system. But I think that moment has passed.
A leaked Microsoft document was critical of Windows 2000 as a platform to run Hotmail on. How can Microsoft convince other corporations to use Windows 2000 if it can't even convince its own people?
Well, in many ways it can't. The document makes a key point, which is that Windows 2000 has a "GUI bias," and that makes it very hard to reliably manage a machine remotely. This is something that companies have been telling Microsoft for a while, and it appears .Net Server, the server version of Windows XP, will start to address some of those issues.
You could spin something like this two ways. You can laugh at Microsoft for having dissension in the ranks. Or you can point out that this is a first step towards Microsoft improving its software. If I was a competitor of Microsoft's, I would be much happier if the company was oblivious of the faults in its products.
This reminds me of companies who have rules that you have to use your own products, such as car companies that require employees to drive its own cars. That's nice, but wouldn't it be better if employees actually voluntarily avoided competitors? So if this document spurs Microsoft to improve its software enough to convince the Hotmail team that Windows 2000 is better than FreeBSD, then it makes things tougher for its competitors.
What's the story on Palladium? Is this Microsoft's latest attempt to regain control of the industry?
Palladium is evidently this year's candidate for generating the most overheated Microsoft-related verbiage. The argument seems to be that Palladium involves Microsoft and security, so it must be some nefarious plan by Microsoft to impose draconian rules on users of its software.
Palladium is at its heart a fairly simple idea, which is hardware support for storing keys and performing cryptographic operations on those keys. It's true one of the uses of this could be for Digital Rights Management, but Palladium is just one component that a DRM system could potentially use to make itself more reliable and hack-proof.
I have seen many magical abilities being assigned to Palladium. One is that Microsoft will use it to restrict access to Office documents to users running Microsoft's Office, as oppose to StarOffice or other clones. The problem is that I haven't seen any explanation of how this would actually happen. For cryptography to work you need some cryptographic secret that is shared or matched between the encrypter and the decrypter. So DVD decryption for playback works because there is some information contained in the players that is related to some information used when the DVDs are created.
So how could Microsoft encrypt an Office document so only another copy of Office could decrypt it? And how would Palladium be involved? Certainly Microsoft could simply embed a cryptographic key in the binary of Office and use it. But it could do that today on current hardware, and what would happen is someone would step through Office in the debugger and figure out the key. So the key would need to be stored in the Palladium hardware. But how would Microsoft be able to have that key available on all systems running Office, yet somehow unavailable to a copy of StarOffice running on the same machine?
I'm not saying it's not possible to construct such a scenario, I'm just saying there's been a lot of hand-waving going on by folks who are predicting things like that.
So will Palladium make Microsoft's software more secure?
Actually I really see no particular way in which Palladium will make software more secure. If you look at why software is unsecure, you have bad design (such as Outlook allowing macros to run by default), bad administration (people not properly configuring their system security settings), and genuine bugs (such as buffer overflows). Palladium doesn't really address any of those directly, although I suppose Palladium might help a user notice some of those problems (if a user tries to play a DVD on a system with the wrong security settings, it might refuse to do so).
What Palladium is doing is going after a security problem that really isn't addressed by current software, and trying to solve it. Microsoft has to fix all the other problems first: make design decisions that favor security over ease of use, make the system easy enough to administer that people actually do so properly, and cleaning up all the bugs. Then it can attempt to write a Palladium system that is trusted.
Incidentally, I think one of the steps it will have to take is releasing the source code. Not as open source, but as "read only" source.
Do you think Microsoft can accomplish those initial steps: better design decisions, easy administration, cleaning up bugs?
Well, I think the design decisions are probably the easiest. Microsoft didn't set out to make Outlook vulnerable. It's a series of steps: first thinking it would be nice to allow macros in email, then deciding they should be turned on by default so novice users could see fancy email, then wanting to make those macros more powerful for some particular reason, and so on. Eventually you arrive at a system where a macro can easily read the address book and send out mail. But it's not like any of the individual decisions was necessarily dumb, it's just that nobody ever stepped back with an eye towards security over ease of use and thought about what Outlook had evolved to. Or if they did, they were voted down.
And it isn't always easy. Look at the buffer overflow in Universal Plug and Play. That was new code and everyone said, Microsoft should turn off features like that by default. But if you require a user to select a setting to enable something like Universal Plug and Play, then it won't "just work" out of the box, which defeats the whole purpose.
Making a machine "easy" to administer is another thing that is hard to do in practice. Of course you would like to minimize the number of settings that users can select. But a lot of times you need to give the user a choice of what do do, because computers are so complex. So you have to allow two or three or ten choices for some particular setting. Then you start making assumptions like, If the user selects this, then they probably want to do this also. So do you do that for them? You may be right and save a novice user, or you may be wrong and annoy a power user.
Removing all bugs will be the hardest. There is no way Microsoft's recent code scrub removed all the remote exploits in Windows, and I don't know why executives are bothering to claim it did. The only way to remove the bugs is to stop changing the code and let them shake out over a period of years, which Microsoft isn't willing to do with its main releases (although it does continue to release service packs for older operating systems).
Personally I don't know why Microsoft doesn't just offer a bounty for bugs. Say $10,000 for each one found, on the condition that Microsoft be notified with a week's warning. Even if 1000 bugs are found, that's only $10 million, which is spare change given what Microsoft spends developing Windows. Make the offer apply to its own employees too, and the eyeballs scanning the source code will be a lot more attentive than they are during enforced code reviews.
How about the recent prediction by the Meta Group that Microsoft will offer its server products on Linux by 2004?
Well, that article seemed to be pure speculation. Still, do I think there are people at Microsoft pushing for the company to port Exchange to Linux? Undoubtedly, and that's a good thing, because the Exchange team should be thinking "What can we do to sell Exchange?" And if it turns out that porting it to Linux is the right move, then they should do so. And if the Windows team complains, the Exchange team can reply that the Windows team should have done a better job of making Windows a platform for enterprise applications (see the Hotmail memo above).
Having said that, I doubt Exchange will actually be ported to Linux. People who talk about this are only thinking of the development costs. They ignore the testing and support. I don't know if Microsoft wants to be supporting Exchange running on home-built Linux systems. How can they reproduce bugs like that in their labs? Plus I think now that Microsoft has recognized Linux as a genuine threat, it will pull its head out of its rear and make Windows a better enterprise platform.
And remember, if Microsoft ports Exchange to Linux it might not be a great thing for Linux. Microsoft's plan would be to migrate users on to Exchange on any platform, while it waited for Windows to catch up to Linux. Then it could start migrating users from Exchange-on-Linux to Exchange-on-Windows. It could remove the price issue by throwing in a free copy of Windows server with every copy of Exchange. And Exchange-on-Windows could always stay a few features ahead of Exchange-on-Linux. Porting Exchange to Linux could be more of a kick in the pants to Windows than any validation of Linux as a long-term proposition.
It sure did seem to take Microsoft a long time to realize the threat that Linux posed.
Indeed. I think Microsoft got sidetracked by the open source thing, because its response to Linux has been really bizarre. First it ignored it. Then it tried to attack the GPL. I'm having a hard time picturing a CIO who views buying software as a way to express their personal philosophy on software source licenses, but I guess that's what Microsoft was banking on.
Now Microsoft has finally decided to treat Linux like it treats other competition, by figuring out why people buy it and trying to make its software better in response. But it's been a long time coming. And with the slowdown in innovation in PC hardware, it may be too late. Linux as it stands now may be plenty good enough for people, and there may not be any paradigm shift on the horizon that is going to dislodge it.
I have often said that Microsoft has been lucky in its enemies. Novell got too focused on file and print sharing and then got Bill Gates Envy, Wordperfect and Lotus missed the shift to Windows (how much of that was due to Microsoft misdirection I won't discuss here), Netscape became obsessed with Java. The few companies that have not made major strategic errors, like Intuit, are doing pretty well. Oracle was lucky in that its mistakes came in the early 1990s when Microsoft had no competitive database product.
Anyway Linux seems to have been quite lucky in having Microsoft as a competitor. For whatever reason, Microsoft has stumbled badly in its response to Linux, and Linux has not made any mistakes. By the way I am really referring to Linux on the server. Linux on the desktop is a fairly hopeless case for the average user, and should remain so due to the lack of application and driver support.
Microsoft made a lot of noise about .Net and Hailstorm, but that seems to have died off now. Any idea what is going on?
For whatever reason, Microsoft was for a period of time obsessed with the notion of subscribers, whether for software, or Internet access, or whatever, but the idea of people paying an ongoing monthly fee rather than buying packaged software when they wanted to. Now they do have some subscription plans going, with MSN and Xbox Live. But with .Net, they tried to put the cart before the horse in two ways, first pushing Hailstorm instead of .Net, and then talking up the subscription features of Hailstorm. This made the company seem greedy and also confused developers who were trying to figure out what .Net actually was.
I think .Net could still be a useful platform for web services. But Microsoft has blown the explanation of .Net so completely that it may have to drop the whole thing, rename it, and start over.
Anything else Microsoft has to worry about?
There's an external issue, which is that Seattle might become a less desirable place to live than it has been. The highway system is terribly congested and the public transit system is inadequate, but the state is in an anti-tax mode so there is no money to fix those problems. The public education system is also running low on money, and as an added bonus has been hijacked by advocates of standardized testing, so it should crater in about ten years. All this may make it tougher to recruit people to live in the Seattle area, where most of Microsoft's product development takes place.
But still, Microsoft has some incredible advantages, in terms of the market dominance of its existing products, and the technical talent it can bring to bear. The main antitrust lawsuit has been (mostly) settled, and it looks like most of the remaining legal issues can be made to disappear with a small dip into its $40 billion cash hoard. So things are not looking too bad for the company.
OK! Thanks for your time and have a great new year.
Thanks, same to you.