Having the last laugh

As group vice president of Microsoft's Platforms Group, Allchin was ultimately responsible for shepherding Windows XP, the company's latest operating system, through to a successful completion last fall. Arguably the company's third most important executive behind Bill Gates and Steve Ballmer, Allchin, a soft-talking Southerner, can be excused for enjoying a moment of schadenfreude at the expense of critics who said XP's debut would fall far short of the hype.

But in the computer business, the XP launch was an eternity ago; these days Allchin is wrestling with the equally daunting challenge of helping transform Microsoft .Net and HailStorm from white-board sketches into commercially successful ventures. In addition, there's the not-so-small matter of trying to bulletproof Microsoft software against security holes during a companywide code review, taking place during the month of February.

Allchin recently sat down with CNET News.com to offer an update on life after XP and how the company plans to hurdle the growing privacy challenge presented to its software.

Q: It seems more of your energy and focus is on the enterprise level these days.
A: Me, personally?

You, personally, as well the company.
If you go back 10, 11, 12 years when I came here, we were nonexistent in the enterprise. We've certainly made a lot of progress since then. That doesn't mean I don't spend as much time on the embedded space, or thinking about Linux or the client. I'm personally very locked into the client because I think that's where you can touch more people.

How do you feel about the Windows XP launch? What went right, what went wrong?
I think it's one of the best launches that we've ever done. And I'm basing that on whether the partners are super happy...when I talked with them, they were super happy. The OEMs (original equipment manufacturers) appreciated the visibility they got and the programs we did with them.

Early on, though, there were mixed reviews.
There was some in the press who said, "Oh, they're not selling as many copies." I think that was pretty misdirected in a confusion of retail vs. OEM (sales). I was saying that our expectations are being met or exceeded, which came out in our quarterly filings. It is what I said it was. What could have gone better? I'm pretty happy with the launch.

Did you finish spending the entire $200 million or so in your XP marketing budget?
We're finishing off spending it. We haven't quite spent it all.

Back to your expectations for XP: Are they being met at both the OEM and retail levels?
Our internal ones?

Yeah.
Absolutely. We're also happy with the mix (comparing XP Professional vs. XP Home Edition).

Was it bigger on the Professional side or the Home Edition side?
I'm not going to give you the numbers, but relative to what we had planned for, the mix was good. And when I say the mix was good, that means more sales of Professional than of Home Edition, which is what we had planned for. We try to do financial plans here very conservatively--sometimes not conservatively enough--but we try to do them conservatively. And they met the expectations I met.

I hadn't expected that Professional would sell as well as it did at retail. We still have our analysis going on as to why that was. Was it because you had more memory being added to those packages, or was it because of the special deals we did? There's probably a team that knows the answer and just hasn't met with me on it. But that was a surprise--though it's probably no surprise to you that we make more money with Professional than (with) Home Edition.

What about in terms of the technology?
I can think of many things that we probably could have done better. The product activation was one that we didn't do a very good job of communicating. If you truly understand it, you can still dislike it. But the problem I saw was that people didn't understand it. I've spent a great deal of time in newsgroups watching what people said, and it's very clear to me that our communication about the way it worked was inadequate. It was inadequate before the launch and it was inadequate after the launch.

Can you offer a couple of concrete examples?
Well, before the launch, it took a company in Germany or someplace to crack the protocol and say, "OK, they're not really sending any private information." We'd been saying that, but they did that. That made me so angry because we should have done that. Why not be transparent? There's nothing proprietary about this. We should have just documented exactly what the protocol is. Why did we have to have a company in Germany make us look like we're nuts because we didn't document it?

I have a saying about being transparent, particularly in the role that we play in the industry and the scrutiny that, rightly or wrongly, we are under: Let's just be transparent. It's OK to be angry at us if we do something that is really bad in the perception of the community. But I swear, so much of the time, people are guessing what our motives are, or there's not enough trust, which is something I personally want to have us work on.

There's still a perception about lax treatment of privacy when it comes to Microsoft software. What are you doing to counteract that?
In this case, we're trying to be very good about privacy. We really are. And so here's a case where we didn't get a positive for it. You can still hate activation, and think that we're not doing anything bad in terms of privacy, but it didn't turn out that way. I had a technical guy go and document exactly what's there, and it's published on our Web site. You can go up and read that. But the fact that we had to go through this is an example that it just wasn't handled well.

Another example: We're just not getting after the lack of information that still exists on this thing. I'm still seeing comments in newsgroups about it. We should be in those newsgroups, pointing them to the Web page, explaining what we're doing. We're getting better every day. But there's just such a huge benefit for us to help customers if we can provide the information for lots of people to see.

Where else do you think you fell short on launch?
I think we did pretty good, given that we were moving to a new code base. I wish there was more device support the day of the launch. As hard as we worked, people were still having problems. Also, we didn't put enough people on (phone support). We did do something positive. I said, "I don't care about the money. Just make installation free." Then we didn't execute perfectly and didn't have enough people on the phone line to handle the calls on the free installation.

We're very critical. You guys may be critical, but trust me--internally, we're very critical.

It would seem that Bill Gates' privacy memo deals mostly with your group.
I'm in trouble most of the time here to some degree (laughing). We're not trying to do everything in a month. I don't know of any other company that's done something like this.

Will the development of .Net My Services require getting more businesspeople involved in addition to the technical people working on the project?
We haven't done the best thinking there.

So it's still not figured out?
My opinion?

Yes.
Absolutely. It's not figured out. We had some technology stuff that went into the (Microsoft Professional Developers Conference). Even there, it wasn't thought through from the tech side like we needed it to be...I think we just got ahead of ourselves and didn't get clear enough thinking. We did have smart people working on this, and they've done some incredible innovation. But the business side was confusing, and didn't fit with what, I think, many of our customers wanted. It's pretty obvious, in my opinion, that customers wanted to run this inside their companies.

As opposed to consumers or individuals?
Yeah. And it didn't make any sense. And the business thinking was all tied to that. It really put us in a weird, weird world. So we're working through that now.

Did you backtrack from where you were before? Sounds like you're rethinking things.
You bet we are...We'll soon swing around to the business side.

Any predictions about when that's going to get ironed out?
(Long pause.) I don't have any--meaning that I think we made a lot of progress today, and that's the key. I want the technology to be right and we'll swing back around.

What's your impression of Apple Computer's latest iMac?
It's a warmed-over Mach (kernel)...The technology didn't blow me away.

What are the things that still keep you up at night?
Linux keeps me up at night in terms of the energy that IBM is putting on it. The thing that's a corollary to that--but much more important--is the intellectual issues associated with that, GPL in particular. I worry a lot because it goes to the heart of whether you can have a business selling software. Whether (or not) we're innovative keeps me up at night...The complexity, the legal restrictions keep me up at night. As to why I'm still here, I don't know. I do love technology.

From the development point of view, what's your long-term ambition about where to take your code?
It should be pretty clear, I would hope, that in our dream, we're going to move to a managed set of APIs...We think there are great programmer productivity gains in going to managed memory. There are some downsides. But then you think about expanding that set of APIs to encompass other parts the apps might want to get to...We may have unmanaged, too, but our long-term dream is managed memory. We're still learning, but we have a whole new graphics system being built with managed memory. So we'll have to wait and see with performance and everything else.

That's going to happen when?
Who knows. The future. Unless stuff is so intertwined, we start projects long before we tie it to any particular product shipment, and we see how far they can make it. And at some point, you have to say "commit"--at which point you know you're not going to roll it back.