Microsoft: All XML, all the time

Jim Allchin is not buying the argument that there's any confusion about Microsoft's message on Web services. If anything, he says, it's just the opposite.

Earlier this summer, Bill Gates allowed that Microsoft's .Net Web services strategy was progressing more slowly than anticipated. But Allchin, who is responsible for Microsoft's platform strategy, maintains that the industry adoption of XML (Extensible Markup Language) Web services is proceeding apace, any speed bumps notwithstanding.

"In terms of the overall vision of XML Web services, I think it's been quite successful," he says.

Allchin, who recently oversaw the first update to Microsoft's Windows XP operating system, sat down with CNET News.com to discuss XML and the future of consumer-oriented Web services, a segment that was once promoted as the future of online commerce.

Q: Bill Gates said a few months ago that .Net has gone more slowly than expected. What do you think is the problem?
A: I actually think things have gone pretty well. There's one thing in particular that I don't think went well, and that was this thing called Hailstorm, which became .Net My Services. But in terms of the overall vision of XML Web services, I think it's been quite successful.

Bill was primarily pointing at two things: On .Net My Services--there was some feedback we got both technically as well as businesswise--that we had some issues. Second, when we look out on the landscape, we don't see enough Web sites, and in particular customer-facing sites, that have XML Web services interfaces that people can take advantage of.

Even at this late date?
Sure. From my perspective, I would certainly like to see a lot more. There are some already out there. There are more inside companies. This is happening inside companies, no question. But in terms of customer-facing--does Amazon have a Web services interface? No.

Do you think there is a technology issue to explain the reluctance for adoption?
I don't think there's reluctance. People love this XML Web services stuff--I mean they really love it. It's really awesome. Today, you have a home page; and you may think you've got some customization, but it's really pretty weak. Weak in the sense that somebody else has designed what areas of the screen you can put things and what things you can choose--like weather can go in this area, and you have to add stock quotes in this particular area--when you could build it yourself, choosing from any of a myriad of sites, and build your own portal, if you will.

But we're now a couple of years in since being told XML would be the next wave.
It is the next wave

Look at the Linux kernel. Do you think there's a lot of innovation there?

Yet the adoption has been uneven.
I don't know if I agree with you. Within the customer-facing, OK. But inside businesses, there are tons of apps being written. If anything (is) surprising me, it's the opposite. Think about what we have today: We have only the basics--SOAP (Simple Object Access Protocol), UDDI (Universal Description, Discovery and Integration), and three or four others--and we do have WS security. If you go back to the beginning of this year, we didn't even have that.

With Hailstorm, there's no question we can have a long conversation about it. We made a push there, and we learned a bunch of stuff and we really retrenched. With XML Web services, we came out with Visual Studio.Net. It's a rock-solid product. We've had lots of good feedback, and in terms of interconnecting devices of all types, I think it's the right vision, and I think it's happening.

What's your current assessment of Linux? Last time we spoke you said you were concerned. Any change?
I'm just as concerned. They are a very serious competitor. We think very hard about it.

Can you be more specific, beyond it being a challenge? What will Microsoft do? You have China and Latin America embracing Linux quite openly.
There are a number of dimensions we think about. Their community is very, very good, and we're hard at work trying to follow that model. The exchange of source within a certain set of licenses, we've also learned from and we've done a lot today.

Obviously...the way (Linux) work is done--I think about it in a more componentized way. I believe in integration because I believe it makes peoples' lives simpler. On the other hand, I consider componentization to be a great attribute from an engineering perspective. Then there are things I don't want to learn from them. It's very hard to innovate when you're in a decentralized mode.

There is a counter-argument to that.
OK, but in all my years of experience, this is what I believe. Look at the Linux kernel. Do you think there's a lot of innovation there? It's not that they're not great developers--they are. I have great respect for them, but I think it's hard, because of the model, to do. They can do innovation--great innovation--in small pockets.

Microsoft's under attack all the time; in some cases, there are very serious attacks.

But let's suppose you want to add handwriting recognition throughout the whole OS; it becomes more difficult in terms of flow and work. You don't have the system componentized in the right way beforehand. Then you back up and have to do a lot of work, vs. somebody who does the innovation--like Microsoft or some another company that has significant R&D dollars. Then they can see, "Ah, that's the way I can componentize that, or how I can break that up to clone that." And I think they're very good at that.

There's quite a dichotomy there and there are some advantages. But in terms of some of the innovations for the future, I don't want to adopt that model.

Has your thinking about GPL, the GNU General Public License, changed?
No.

Last week there was a critical flaw in Windows announced. Are you satisfied with your Trustworthy Computing effort to date? How much longer before you'll be able to make this stuff rock-solid?
It's a journey. Am I satisfied? Yes, I am. How long before it becomes rock-solid? As I said, it's a journey, and we feel very good about the reliability of where we're at.

At this point, Microsoft's Palladium project to enhance security is a white paper and will take several years before it becomes tangible.
We're not ready to ship it, but it's a lot more than a white paper.

How long before becomes something you're able to get to the market?
It would be premature right now for me to say. It's not just retraining our developers and putting them through a more rigid program and doing threat analysis and all the other things we've done. We're also trying to think ahead in terms of WS security, in terms of this truly global environment of interoperability between lots of different vendors.

Did the attacks on 9/11cause you to reconsider the way cyberterrorism gets treated at Microsoft?
I think everyone should be concerned about cyberterrorism. We already had the religion. Microsoft's under attack all the time; in some cases, there are very serious attacks.

The last time we spoke, you were just finishing preparations to get XP out the door. What's at the top of your agenda these days?
Quality is at the very top: Excellence in design, architecture and relationship to customers is very important. In terms of actual products, there's a set of technology in the Yukon wave. Then there's an update of the .Net framework, which will let you do storage procedures as part of doing databases. But the majority of my time is being spent on Longhorn (update to Windows), which we're many, many years away from shipping. A significant part of the team is already switching over (to work on Longhorn.)

I haven't heard much lately about software as a service. What's your latest thinking about the subject?
I still believe in the concept. Are there going to be business models? Sometimes yes, and sometimes no. Perhaps the largest software as a service we have is Windows Update.... There are businesses that are in today with subscriptions. MSN is moving toward that model, as are all the antivirus companies. Could Office-type productivity products be done that way? I don't know. Maybe for small businesses. From a technology point of view, we want to be able to do that.