Thursday afternoon at the Web 2.0 Summit wrapped up with a conversation with Microsoft Chief Software Architect Ray Ozzie. A year ago Ozzie sent a famously leaked memo about how the world in which Microsoft software is released has changed. Ozzie talks to program chair John Battelle about Vista and other complex software projects.
You can download the audio as an mp3 or download the video as an mp4, or you can subscribe to the audio podcast or to the video podcast. Check out the entire set of Web 2.0 Summit podcasts.
This episode is sponsored by the Intel Software Network.
Transcript created by Casting Words.
ANNCR: Thursday afternoon at the Web 2.0 Summit 2006 wrapped up with a conversation with Microsoft Chief Software Architect Ray Ozzie. A year ago, Ozzie sent a famously leaked memo about how the world in which Microsoft software is released, has changed. Ozzie talks about Vista and working with other very complex projects. Here's Microsoft's Ray Ozzie with Web 2.0 Summit Program Chair John Battelle.
John Battelle: Many of you may recall the memo that was written by Ray a year ago, which was kind of one of these great Internet tidal wave memos that so famously come out every decade from Microsoft. With that I'd like to bring Ray Ozzie, the Chief Software Architect of Microsoft, up.
John: Thank you for coming and for being willing to talk to me.
Ray Ozzie: Thanks for having me. Congratulations on a pretty successful, at least measuring by the buzz, Web 2.0.
John: I'm glad to hear that. I'm too busy taking notes and trying to catch up here, to know what's going on but I hope you guys are enjoying it.
So, I promise I won't ask you about quarterlies and DNLs, twice.
Ray: Thank you.
John: Maybe once, but not twice. Let me start with the preamble that I brought up, this memo that you wrote, which was preceded a few days before with an announcement that Bill Gates sent around. We've all seen the whole email. It's on the web if you just search for it, "Bill's emails" on the web thanks to Dave Winer and some other folks, where he announced that you were going to take over the service-driven portion of the business and you were going to be responsible for this transformation at Microsoft. Some days later, this memo was leaked. Was it really leaked?
Ray: Yeah, it really was leaked. Once it leaked then we had to fully...
John: OK. It was very successfully leaked.
Ray: Oh no, it leaked. It wasn't leaked.
John: It leaked, it wasn't leaked. I feel like I'm talking to Bill Clinton here.
John: Depends on the definition of leaked.
The memo was eloquent and honest and it talked about some of the issues that Microsoft might... and the language was very aware of, it was really almost you were a stranger in a strange land and you were trying to really shake these guys up a little bit but not freak them out. It's been a year. You wrote it in September of last year. Since then, your role has changed again. Two-part question: How has your role changed? And how is it going after a year of shaking up things with that memo?
Ray: It's going well. It's a large organization. There are many, many products. It's a very broad challenge and a very broad opportunity. When I came to the company, I could see that some people really got it with respect to the shift that the industry is in right now and some people were heads down, working on Vista, working on Office and working on whatever product they had. I really felt the need to get a broader message across that this is something that everybody really needs to pay attention to because it's fundamentally going to impact every one of the offerings over time. The intent of that memo wasn't to prescriptively say, "you should do X, you should do Y, you should do Z", but more or less to start people thinking about it. I think in June, when Bill announced his transition, it more or less caused a bunch of people to go back and re-read the memo.
John: I bet it did.
Ray: I think the first go around, people probably read it and said, "interesting" and then the second time people read it, people went "hmm." Now, we're at a very interesting juncture with Office done, with Vista done. We're...
John: Well, let's define done.
Ray: Released to manufacturing.
John: In a world where people update their software every day or every week, what does that mean, "released to manufacturing?"
Ray: "Released to manufacturing" means it went through an immense amount of testing. It went out to many, many, many people. It means that when it goes out, it will not be perfect, it will be updated. There will be a great set of drivers, more drivers than there were for Vista at that point in time, but there will still be missing drivers. At the point that it is right now, it's feeling really good in terms of the amount of internal deployment that we've had. I've used it for quite a while. Is it perfect? No. It's software and it will have flaws, but it's fulfilling the role that I believe an operating system really, really needs to have in this era.
Back when XP was launched, the Internet was a far different place than it is right now. The amount of attacks that are going on, just very, very commonly right now, out on the net, is tremendous. The biggest single thing that Vista can really do is to provide a safe environment within which you can do your interaction on the web. A safe environment that you can load code onto and try to load code in a way that doesn't leave itself as open as Vista and other operating systems did before it, in terms of just installing Trojans and things like that.
Vista was really secure by design. It's undergone a tremendous amount of static analysis of the code to make sure that we did the best possible job we could, given the surface area of the product to make sure that vulnerabilities were removed before it was shipped.
John: When is it coming out?
Ray: It is released to businesses by the end of November. It'll be released to consumers by the end of January and things are on track.
John: And same for Office as well?
Ray: Office is released to manufacturing. The launch event is I think November 30th.
John: Who's playing?
Ray: Who's playing?
John: The Rolling Stones' "Last Time." They're still around.
John: I bet they're not as expensive as they were last time.
So, let's go back to the thrust of the memo. There's one thing. First I want to talk culturally and then get to the specifics of the service-based architecture. There's this great quote, which of course I can't find now, in which you gave one concrete example I thought was a really great kind of cultural anecdote. I found, in my experience, when you're working with teams, it's good if, for example, they sit together in the same space. For a memo that was really laying out a lot broad strokes and explaining some big ideas, to then drop in to this very specific example of let's not actually have meetings and then do emails afterwards but let's just sit in the same space and work. That sounds like what you were doing when you were running your company before it was acquired by Microsoft.
Ray: I've been working in the collaboration software business for quite a while and CS is just technology that is intended to help organizations reduce co-ordination costs. This goes back to Ron Coast and many many people have studied it and technology is not a panacea for reducing co-ordination costs. A lot of it is leadership and process and management practices. Unless you kind of step back sometimes, when you have very complex projects by design... unless you kind of step back and question what are the ways that we're restructuring to revisit them sometimes you don't take advantage of opportunities that you might be able to have to make things operate more seamlessly. In some cases, depending on the nature of the product, if a lot of innovation is required... if you really need to do a lot of white board brainstorming, then being together in a physical space is really a great way to do that... oddly enough in some projects, breaking them apart, and putting them even geographically distributed introduces a barrier that you forces you to have to overcome and sometimes constraints like that can be liberating and sometimes you can get more net work done by introducing those constraints than by having lots and lots of little interruptions in a large project.
John: Do you think that we're talking about a significant cultural change with Microsoft? Are you going to be seen as the standard bearer in a significant shift in how things get made?
Ray: I think you'll have to ask people progressively over time who work there. I believe that changes like this have to happen locally, that they can't happen by mandate. You more or less have to go to a group. Different groups have different work styles. The Office workgroup has a radically different way of structuring its processes, than the Windows group and that's different from the Xbox group. But each one, as I have an opportunity to talk with them, they'll hear my input. I've been listening to the pros and cons of different ways of the structure and I think things have changed already in certain ways and things will change over time, but it's not one person who makes this happen, it's a conversation.
John: Now when eleven years ago Gates wrote a memo called "The Great Internet Tidal Wave", I recall the coverage, I was working at Wired at the time and I recall the coverage was essentially a year later that Microsoft had "turned on a dime", the great aircraft carrier of a company had turned around and steamed over Netscape within 18 months. This time, is this that kind of a memo, is moving towards, migrating toward a web services, and a new business model of advertising and other potential models that might develop around services... Is it that big a deal again?
Ray: I think it's as big or bigger deal from a business perspective but it's not that necessary to turn in one day. I believe that it's important for each product, for each product group to look at the customer, to focus on the end user, you have to focus on the customer first and foremost and say what kind of experience... in this era, given the bandwidth and the type of devices that we've got, what in this era is the best way to deliver that value? If it's productivity value, if it's CRM, examine what the best delivery model for that type of value is... what kind of things do people want to do on the web, what do they want to do when they're around here and they have their mobile phones and then to reshape and re-architect that product.
John: This is what you call "scenario based design"
John: Because often times large companies say what can I make out of the sausage factory I've currently got that I can foist upon the public. But you're saying, your note in the memo is that all these innovative companies have very lightweight design principles that turn something out and then improve, improve, improve. That is not the... how long did it take for Vista to get made... years? and office to get updated... three years? Do you feel like that is really taking?
Ray: It's definitely taking. I mean just look at the nature of any organization, particularly one that has been as successful as Microsoft. MS success began in the mid 80s and this was a pc era about personal empowerment, using the PC as a tool. And the most natural thing is to build upon that success and then it expanded into servers and now to build on that server success. The people in those groups unless they are self-motivated to look externally and say what's going on on the outside are going to continue seeing things through a PC based lens, through an enterprise server base lens. What I've been hoping to do, and what I've been trying to do is to help people step back, to step outside the actual delivery mechanism of that value and say "for the experience that we're trying to deliver to the customer, what's the best piece to put on the pc, mobile device and web?"
John: Let me ask you a question. I'm asking it later than I did everyone else but I'm still going to ask you. What do you make of Google?
Ray: Very successful company. They've taken what was pioneered by Bill Gross and Overture and have done incredible things. They've delivered good value to the users and they've built a great ad eco-system and they are a force to be reckoned with for many people in the industry. I think that there is immense opportunity in the core space that they are in and I'm actually surprised at this point in time that we have not branched into, I believe there are a number of ways of doing search, contextually, particularly in the realm of refinement of search and things like that can satisfy the user's intent more than we've seen thus far, but they've stayed focused on the user and that's something that I very much believe in, and I think that as long as we build properties that are focused on the user we're going to have satisfied users.
John: Now Google proved the advertising model for web services. They are proving it, certainly for search, possibly for other services... that seems to be where they are headed. What's the general zeitgeist on the Microsoft campus as Google rolls out one and another web-based word processing, web-based spreadsheets, web-based mail, web-based calendaring... starting to look like the Office Suite. Eric was here earlier, claiming that he is only interested in consumers, strikes me as everyone in the world. Do people start to get angry when you talk about Google up in Redmond?
Ray: I think, depending on the personality of the people in the company, people are used to winning.
John: There's a chair-throwing incident, for example.
Ray: I know nothing...
John: I'm just, I mean that.
Ray: Let me just tell you from my perspective.
John: Let's look it up, Ballmer throwing chair...
Ray: From my perspective, when I look at what Microsoft has in the productivity space. We've got about, I mean, depends on how you count, but probably close to half a billion users out there who use office actively. This is a dress-able market. Some of those users are paid users, some of those are prospects for other services that we might be able to up-sell them into. From my perspective, I look at this thing, and I go "what kind of an opportunity." I don't have to go out and buy companies and do this and do that to acquire a new audience, they understand the value proposition already. All I have to show them is that we get it and that we can deliver value in a way that is relevant to them and that's relevant to this era that they want to use.
John: How long is it going to be until my interaction with Word as a writer or Excel as a spreadsheet jockey is going to completely web native? I'm not saying it won't also exist on the client and there won't be client software. You just created in my mind this vision of me as someone that is connected to Microsoft the way I feel connected to Google when I use their search service.
Ray: I think it depends on; sorry I'm going to come back to this, it depends on the scenario of usage. I don't see the way that the way things on panning out right now and the nature of how the technology works. I don't see that it's the right thing to do necessarily to just take the PC interface and functionality and just port it up onto the web. I think what you have to do...
John: So you differ quite significantly from Eric, who in the first day of the conference pretty much declared that's where it's going, end of story, game over.
Ray: I believe that you have to look at what the web is really uniquely good at, universal access, sharing scenarios. It's good at quickly going in and out of something if you happen to be online. Then you have to look at the things the PC is good that the web doesn't necessarily have as it core strength; such as, really really flexible fast UI regardless of the connection speed. A certain amount of reliability in a world of Wi-Fi where you are going from hotspot to hotspot you end up with dropouts. We're actually going from a world in the 80s that was focused on the PC for document editing to a world where we're embedding different media types into what we do. This is where the PC really has amazing strength. We're going to have terabyte hard-disks, like JVC Camcorder has a 60 GB hard disk in it that I take videos, drop it onto the PC, will those things be streamed up to Internet services? Absolutely, but we have constrained upload bandwidth and the media editing functions, that's what the PC was designed for. If it's list-sharing scenarios, scenarios where you've done some work and want to organize people into a workspace very quickly and do some things like that, some of the things that we're doing in Office Live. Those are the things that the web-based scenarios are going to just nail and they'll complement with common file formats and the whole nine yards the things that are on the PC.
John: Guys, please come to the microphones for some questions if you have any, feel free to. One last from me, it strikes me that when Bill Gates was chief software architect, Bill Gates go into a room, any room, and say, in fact he was quite famous for this, his style I think may have been different from yours, just a guess, I never sat down with him like this, I wouldn't know, but matter of fact, one time he did storm into the editorial offices of Wired and tell us how dumb we were. That's pretty much my one encounter...
Ray: I'll try to live up to that.
John: He had the ability to pretty much mandate something to happen. He not only was chief software architect, he happened to be the founder of the company. Do you find you have the authority you need to get things done in an organization as big as this? Or are you finding that you're going to be doing a lot more consensus building than Bill would have done in the same role.
Ray: That's a really good question. There is a certain mythology, of course, around any leader, particularly Bill, who is a really talented guy. He is very broad and very deep. He can engage in some very deep technical conversations even given his breadth. The way that Microsoft works, it is very rare, very very rare that Bill actually gave an order and said, "Go do this." Bill had, has and will always have an amazing level of soft power within the organization and people want to follow. To be a good leader, you have to have good follower-ship. The organization reveres him and wants to do what he says would be the right thing to do. I was given a free pass coming in because I've known Steve and Bill for a long time, my companies have had a good relationship with various technical groups within Microsoft, and so I came in with the benefit of the doubt. But I've got to earn that follower-ship, and that takes some time. I'm really fortunate that Bill has given two years to this transition because people can see us together, us interacting together in a number of different meetings. Will the organization miss Bill? Absolutely. Are things challenging? Sure. My interaction style is significantly different.
John: So you don't rock; rock in the chair?
Ray: Ah, no.
John: All right, questions back here...
David Ellingston: Yes, David Ellingston, SAP, so if the theme for this generation of Windows and Office is safety, and I really hope reliability. What will the theme of the next generation, if you look out five years to a new generation of Windows and Office and desktops?
Ray: That is a good question; I think it is different for Windows and Office. I'll just say first that the way that the organizational processes work, really the product groups will define what that is what I can tell you is kind of my intuitions as to where opportunity might be, but on the office side again I think the biggest opportunity is that the, it's a world where mobile devices are exploding, smart phones are very shortly going to be all phones not just the differentiation between phone types that we've had in the past is changing, the web is pervasive and I think that the scenarios, there is tremendous opportunity on the office side to take advantage of those, of different types of productivity scenarios involving multiple different types.
On the Vista side and on the Windows side I think there are if look at the natural hardware trends just leaving safety aside, if you look at natural hardware trends what's really happening is right now is we are about to, we are on the cusp of going from multi core to many core processor environment where pc's are going to have just many, many, many, many different processors and I think the system needs to help application programmers be able to consume that in a reasonable way without teaching everyone on earth to be, to understand that, how to factor their code that way.
I also think that are, I think many people here with laptops would probably agree that just in something as straight forward as power management, there is tremendous opportunity to innovate and that really requires innovation at all levels of the system.
There are other things that are kind of like hardware problems that have been in the Windows domain for a long time that we are going to take the opportunity to address such as state separation which is keeping one app completely separate from other apps in terms of it's execution but also it's settings and things like that, so that deployment models can kind of be brought up to date this concept of inserting a CD and setup that exe. From my standpoint is dead, it should be gone, you now we are in an era where everything should deploy from the web, things that you use on the client.
I mean, when you go to a web page you are bringing java-script down and you are bringing code down. A client side install is bigger than flash and bigger than java-script but you shouldn't need a DVD to do that, and so I think Windows supporting those kinds of scenarios, there is tremendous opportunity.
John: We have time for a couple more questions.
Questioner: Hi, in a world that's obviously more browser centric there is a big event going on the IE seven upgrade, do you personally think it was a good idea for Microsoft to bind the upgrading of IE seven so closely with license validation? And do you think that will drive the gray market Windows users more quickly on to Firefox?
Ray: I don't know, I don't know, the entire issue of license validation and what do you do if the license doesn't validate, is a very dynamic issue that is discussed, there are many, many pros and cons of it, we are a company that makes money by selling our software but on the other hand my view is that unpaid users, if Office and Windows were services today we would treat those unpaid users as prospects as anonymous users, so I think there are multiple, different ways of viewing it and it is a conversation that continues.
John: Over here.
Questioner: So, Bruce Chisson was here earlier and pretty arrogantly claimed victory and said that in the realm of PDF and swift-files game's over, you lost, they won.
So, I wonder if Microsoft is no longer going to play in that sandbox and if you are, what is your response to Bruce's comments earlier?
Ray: Well, PDF, many PDFs are generated from Office products and we would like that to be a default capability of the products.
There are complex reasons why people don't want that. I think customers really do want that, but I would love to that to be a base feature of Office.
In terms of having a competitor to a file format my basic attitude is, as an industry we've lived with a number of file formats for many, many years and to the extent that we're now in an era where customers are storing data for long, for forever, and that the data outlives any application that generated that data. It is imperative that we as, we all, all vendors have products that create formats that give customer a choice and give, and have longevity in the data.
Whether PDF is that storage format or XPS is that storage format or there is a or there are other persistent formats for other types of things, I just think it's going to be a, the great thing about XML is at least it makes very transparent what's going on inside, is much more transparent than binary formats and I just think that is the world that we are in right now.
John: Last question here.
Questioner: Yeah, I had a question about Zune. My understanding is that Zune will not be compatible with Plays-for-sure MP3 players and the store won't be compatible either.
I was wondering if you could kind if explain how Microsoft came up with the decision to go with the closed system for music at the risk of antagonizing partners and OEMs and why that made sense for your business?
Ray: There is, there was one very very strong focus with Zune and given the time frames, the development time frames and things like that, that were involved, we really needed to stay focused on that, which was, the goal of Zune was to build an and-and experience because customers have clearly stated that's what they want.
Now, there are technologies inside of that and-and experience related to Windows Media and things that are useful far beyond our particular implementation of that experience.
And we license it, it is a, that is the Plays-for-sure platform and we license those technologies far beyond it. Whether in the future we could investigate making Zune a Plays-for-sure device or having the store license it's music in that, those are all decisions that are nothing is a forever a forever decision, but we never would have succeeded, I'm being very, very frank here given the size of the team, given the time frames involved, we never would have succeeded at the end experience if we put too many complexities and too many dependencies on the team at once.
John: Thank you, now Ray, thank you very much for coming and speaking with us, we appreciate it.
Ray: Thank You.
ANNCR: Microsoft Chief Software Architect Ray Ozzie at the Web 2.0 Summit 2006.
Daniel H. Steinberg is the editor for the new series of Mac Developer titles for the Pragmatic Programmers. He writes feature articles for Apple's ADC web site and is a regular contributor to Mac Devcenter. He has presented at Apple's Worldwide Developer Conference, MacWorld, MacHack and other Mac developer conferences.
Copyright © 2009 O'Reilly Media, Inc.