At next weeks VSLive conference in San Francisco, Microsoft Corp. will be announcing the status of such tools as Visual Studio 2005 Beta 2, SQL Server 2005 Beta 3 and the first Community Technology Preview of Indigo. S. “Soma” Somasegar, corporate vice president of the Developer Division at Microsoft, who is one of the shows keynote speakers, sat down with eWEEK Senior Editor Darryl K. Taft to discuss the companys plans for these tool releases, as well as the developer ecosystem, dealing with open source and more. Prior to his current role overseeing the developer division, Soma served as corporate vice president of the Windows Engineering Services and Solutions group within the Windows Division. Prashant Sridharan, senior product manager for Visual Studio Team System, joined Somasegar in the interview.
Whats the holdup on getting Visual Studio 2005 Beta 2 out?
Somasegar: Were still aiming to ship VS 2005 and the .Net Framework 2.0 this summer. But more important is delivering the right product with high quality in a timely manner. We expect our early adopter customers to go live with VS 2005 Beta 2, and we need to ensure the right level of quality to enable these early adopter customers to deploy in their production environment. But in the end, we want people to remember the quality of the final product—the one they use on a daily basis for years. We have to get the quality bar right.
Whats the importance of support for smart clients in Visual Studio 2005?
Somasegar: Smart clients are easily deployed and managed client applications that provide an adaptive, responsive and rich interactive experience by leveraging local resources and intelligently connecting to distributed data sources. The idea of smart clients is to bring together the good things about local applications like support for running offline and rich user interface with the things that are good about Web applications like easy deployment. Were working to make this vision happen by introducing technologies such as the .Net Framework 2.0s ClickOnce that make application installation and update simple. The goal of Visual Studio 2005 is to make it simple for developers to use these technologies so they can deliver better experiences to their users.
From your perspective in this “new” role for you, whats your message to developers on how youre going to deal with the issue of open source and the movement toward the open-source community?
Somasegar: One of the things I have personally learned from the open-source community or movement is transparency. The reason Im excited about transparency is if Im a developer what I really want to know is the internals of the system, I want to know when decisions are getting made, I want to know why the decisions are getting made and—assume Im a developer outside of Microsoft—that Microsoft can provide a way for me to interact on a regular basis with the product teams that are building the technology. Then I feel good about using the technology and taking a bet on the technology. In some sense the thing open source has done very well is having a rich, vibrant community. Thats what weve learned over the last three or four years, that having a rich, vibrant community of customers is absolutely critical and is even more crucial in the developer space.
Over the last two or three years weve put a tremendous amount of effort and focus inside the division now, in terms of connecting with the customer, communicating with the customer and creating this sense of a close community as we make this huge step forward. Look at things like the Community Technology Preview [CTP]. It looks like thats just an obvious thing to do, but I can tell you inside the team there was a lot of angst because historically we were in a mode where wed write some code, wed go through an extensive stabilization period, wed call it a beta, and then wed show it to customers.
We were sort of not ready to air our dirty laundry so to speak, because every build that comes out wed have some problems, wed have some glitches, but thats how software gets developed—from a build to build perspective. And we were trying to say, Hey, if I can share it with you as an engineer in the team, I want to share it with my close community of customers. Because I want them to see the progress we are making. And if they think we are making the wrong decisions they can tell us right then and there. I dont want to wait another year before I can get feedback from the community.
I think trying to keep the close set of customers, whom I call the community, an integral part of the development process is what transparency is all about, and I think thats what developers want and thats what we want. And thats the only way I know how to build more effective products for our customers.
The CTP Process
How is the CTP process working? Ive heard complaints that there are too many CTP releases and things are becoming confusing for testers and leading them to be less thorough. … What are you hearing from developers on this?
Somasegar: We consistently receive very positive feedback on the CTP program—developers tell us every day they love the frequent code drops so they can monitor progress were making on VS 2005. Developers love this level of transparency and are feeling more like an extension of the VS development team as a result of the CTP program.
But its critical for testers to also understand that VS CTPs are intended to only provide a snapshot of progress at interim junctures of the product development cycle. Developers should decide how much time they want to spend on each CTP release. We do love the feedback that we get from developers on our CTP drops, and that absolutely helps us build the right product. If there is something new or they care about a particular feature/bug that is fixed in a CTP release, by all means they should pick up the CTP and give it a test drive. On the other hand, depending on what a particular CTP includes, some developers may choose to just give a cursory look at that. We dont expect every developer to go deep every time we release a CTP.
The changes in Longhorn announced in the fall, how will they affect Orcas?
Somasegar: The previous plan of record, before we made all the changes, was that we would ship Longhorn and then we would have a tool set called Orcas that would target that system. Nonetheless, we also said, in the old plan of record, that if a developer wanted to develop a Longhorn application using the new managed interfaces and the like, you do not need to wait for Orcas. What you can do is take the Whidbey or Visual Studio 2005 tool set, and you can start writing a program using the Longhorn SDK—thats a Longhorn application taking advantage of the new managed interfaces, and that application will run on Longhorn. The things we are missing in Visual Studio 2005 are some specific designers targeted at the Longhorn components, namely Avalon and Indigo and the like. But to be able to write code, you can get started on that with Visual Studio 2005. So that is the old plan.
Now the changes that we announced in Longhorn in my mind, if I had to summarize, there are about two or three changes that we announced that are meaningful changes. One is we said we are going to do whatever it takes for us to have a high confidence plan of delivering a very quality Longhorn in calendar 2006, so that if we have to cut features well cut features, if we have to do something else well do something else. But it is super-critical because we think there is so much value that we think we can deliver to our customers that we really want to get into a plan that will let us deliver Longhorn. That is one thing that we announced.
The second thing that we announced was if you think about WinFX as sort of the next-generation API layer that we want developers to target … there were three components that we talked about before: Avalon, Indigo and WinFS. What we said was the feedback that wed been getting from our customers was its great that Longhorn is going to be this great, new platform, but hundreds of millions of installed base customers arent on Longhorn, and if you tell me that my Longhorn application runs only on the latest and greatest hardware, then I have a problem because I really want the reach.
Reach has always been an important thing for ISVs of the world. And for a while we were thinking that Longhorn is so great that everybody will just get that. So we took this opportunity to say that Avalon and Indigo, when we deliver the first version of that in the Longhorn time frame, those things will not just run on Longhorn, but also run on Windows XP and Windows Server 2003, so that we really have a solution for the reach issue that ISVs have. That was the second thing we announced.
The third thing we announced was the third pillar of WinFX, which is WinFS, the new file system and storage system, we said that will not be ready in the year 2006. We said if you want to be realistic about it, we think it will take until 2007. And so we really had two options. One was to say were going to slide everything until 2007 and wait until WinFS gets ready so we can ship it all at one time. Or we said we had the option of saying that WinFS will be delayed until 2007, but I can get everything else done in 2006, and not only that, I can get Avalon and Indigo to support down-level systems as well. And obviously the latter plan looked more interesting for our customers and for us, so we went with that.
So in the new plan, from a tool set perspective, its pretty much the same. Whidbey is the tool set you want to use to start building Longhorn applications. You can start with the Whidbey beta and Longhorn SDK. If you want to wait until Whidbey RTM, thats fine too, but Whidbey is the tool set you want to get started on.
The next version of the tool set, which we are internally calling Orcas, will still contain the new designers that we want to have in place for some of the Longhorn components. But beyond that it doesnt fundamentally change us. Well still shipping Whidbey this year, and then well turn around and do an update with Orcas that fundamentally is Whidbey, but includes some new designers and some other things. And we are still in the early phases with Orcas.
So Orcas was originally planned for release in 2006?
Somasegar: We always said it was Longhorn plus probably three to six months. That is it would be part of the Longhorn wave. And in the new plan, Orcas is still part of the Longhorn wave.
Microsoft Business Framework
What about the Microsoft Business Framework [MBF]? When will we see that? I heard that was delayed because of the changes as well.
Somasegar: The way this happened was previously we used to have object spaces and entity persistence because we wanted a way to have a relational store for MBF. We made a decision to say we are going to align our technologies here and really take a bet on WinFS where WinFS is also going to provide a high-quality relational store in addition to providing a high-performance file system. So in some sense we said MBF was going to take a dependency on WinFS. Now with WinFS shipping in 2007, under the current plan of record, well ship MBF with WinFS. But we are also starting to look into whether there is a way to accelerate the delivery of MBF. And the teams are working through plans.
So no date?
Somasegar: No date today. If we continue with the dependency on WinFS and wait until WinFS is ready to ship, then it will be 2007 whenever WinFS ships. But if we can accelerate the relational store and we can get done earlier, then well get done earlier. And our hope is … to get done earlier than later.
The big thing that we want to do in the MBF context is we have … a program called TAP, the Technology Adoption Program, where when we are developing a product we pick a finite number of large customers who are going to work with us hand in hand through the development process, and these guys are committing to building something on the new platform and deploying that in their production environment before we ship the product. In some sense if you look at the ship criteria for our products today, one of the ship criteria is sign off by the TAP customers. So for MBF weve got about 30 TAP customers around the world. The key is for us to give them early drops of our code so they can see what we are building and start building their application on MBF. So our plan is we are working on the M3 milestone and well provide an early drop of that code to the 30 TAP customers so they can stop working and give us feedback.
Ive heard MBF is being refocused as a DSL [domain specific language] tool now. … Is this true? If so, why the change?
Somasegar: We have always thought of MBF as a framework that is focused at the line-of-business [LOB] ISVs. That still is absolutely how we think about MBF. In that sense, you can think about MBF as more specifically targeted at business applications, but it is not a DSL tool. The set of graphical designers that we ship with VS 2005 [code-named Whitehorse] is an example of a DSL tool.
When are we going to see the Software Factories stuff in Microsoft developer products?
Somasegar: The way you want to think about it is its really a combination of technologies or tools. Its a combination of building in some prescriptive guidance, some process guidance and a bunch of what I call solution accelerators or templates and the like to make it easy for people to go through the software factory notion of building software.
And with Visual Studio 2005 you are going to see the first instantiation of the tools, the process guidance, the prescriptive guidance, and templates and solution accelerators coming together where they can really execute on the notion of software factories. So thats the first time youll see that coming together.
So well see some in Whidbey?
Sridharan: What youll see are components of it. But there is a larger vision associated with what software factories are.
Somasegar: The first step is what youll see in Visual Studio 2005 when the Team System stuff ships. And then Orcas will take it forward, and the next version will take it forward.
But when will we see the vertical tools focus of this stuff?
Sridharan: I think youll see that in the near term. Im not sure exactly when, but thats something were working with a number of our partners on.
Somasegar: Thats something to remember because some of it we will do, but a lot of it well likely expect our partners to step up and work with us in delivering the tools for the vertical space.
Sridharan: The key thing is software factories are not a Microsoft thing, theyre an industry thing. I think its in our best interest to make software factories an industry thing … to shepherd it but not to exert a lot of control over it. And to get a lot more people involved and interested in it. And that needs to happen in its own sort of vacuum first. That ecosystem needs to thrive first without a lot of scrutiny from the outside world. Then once they have critical mass they can go outside. And thats sort of where we are right now.
Somasegar: And I think once people hear our story theyll realize there isnt really a fight about it, but that were thinking about it in a broader way than what some people are thinking about. And we see a world that is a broad world where those people who have a little bit narrower focus than what we have, they can still plug into that world and the world is one happy place. I think as people begin to understand more of what were saying about domain-specific languages and the broad spectrum of things you can do in modeling, there isnt going to be this “us versus them” attitude. We just need to do a good job of articulating our story and backing it up with the tools and technology.
Sridharan: Certainly from our perspective theres going to be no “us versus them” attitude. Were going to do UML, and were going to do a lot more.
What is the status of Sparkle? [Sparkle is an authoring tool meant to compete with the Macromedia Flash Authoring tool.]
Somasegar: The plans for Sparkle have not changed any with all the Longhorn plans. The reason is because one of the things we decided early on with Sparkle is to say that we are going to take a bet on Avalon. On some sense the updated plan has only helped us because now there is an opportunity for Sparkle to run wherever Avalon runs. So if you think about Sparkle as an ISV thing, or a tool on top of the platform, then reach is important. So in some sense its made our lives a little bit easier.
So is it still viewed as a [Macromedia] Flash killer? Thats how its been characterized by some.
Somasegar: The way I think about it is if you want to really present information to your users in a highly compelling way, Avalon is the platform that lets you do that. Now are we going to compete with other people? Yes. But really what we are trying to say is for the breadth of customers that we have, and for the breadth of scenarios that we want to enable, a much richer UI experience will go a long way in simplifying the user experience and in enabling them to have richer UI experiences. And thats the primary reason we came up with this new presentation subsystem that we think is also going to let you take advantage of the latest hardware advances in the graphics world.
The other thing this decision has done is that if you think about internally in the company, almost everybody that I talk to and the teams I talked to, people are excited about this plan. Because they feel that they have a high degree of confidence that they can deliver on this plan.
Has it been considered that Avalon might be dropped from Longhorn like WinFS?
Somasegar: We absolutely thought about that. We always think about that kind of possibility because, as I said before, the feature decisions that we make we constantly revisit them because at the end of the day there are three endpoints to the triangle. Youve got features, youve got quality, and youve got time to market. Theres always a good balance you need to strike between all of those things. So every feature we think about whether its the right feature, is it a feature we want to hold the product for, or is this feature less important in the grand scheme of things when I think about time to market and quality and adding value to our customers?
So in the Longhorn process we thought about pretty much every feature. But we decided that the only feature that we wanted to delay until the following year was WinFS. Particularly because we were trying to build something that was really ambitious. Its both a high-performance file system and a high-performance relational store. And trying to get both together done along with everything else we were doing in the platform, we just thought it was going to take a longer time.
So we thought about that, but as of today Indigo and Avalon are going to be in 2006 with Longhorn.
In terms of the Microsoft developer ecosystem, how do you balance the competition with partners along with the cooperation with them?
Somasegar: Thats always an interesting balance. Because on the one hand any time we build a new tool or a new piece of software, you always think about “Hey, what does the competition look like here?” And a lot of times the same people who are our partners also compete with us in one dimension or another, so how do we handle that kind of thing? We always think about it and sometimes deeply agonize about that before we make decisions.
Just to give you one example … This happened probably about 18 to 24 months ago, which was about the time inside Microsoft that we decided we were going to get into the enterprise tools space. So, prior to that, wed been thinking about that being the customers demand for a cohesive set of enterprise tools targeted at team development. People wanted these tools to really be integrated. And if you talk to large customers theyll tell you they have a variety of solutions from other vendors and a lot of them have homegrown solutions, so they really wanted an integrated set of solutions. And we thought that that was a place we could get into and add value for our customers.
But then when we started thinking about this area, we said it was a place where a lot of our partners and competitors were already doing work. So the dilemma we had was how do we get into this and still do the right thing by our partners?
So we decided to bring in about a 150 to 160 other companies—people who we thought were involved in the enterprise tools space—and brought them into Microsoft at Redmond [Wash.] and said, “Well give you a complete briefing of what our plans are in this area. And youll have at least a two- to two-and-a-half-year lead time because of how long its going to take us to get into this market in a realistic way … so we want to share our plans now with you so that you have a choice. You can make a decision of how you want to play.” Because though we are entering this space there is still a ton of opportunity for our partners to bring complementary offerings to what we are doing to fill out the suite.
We always think about how do we effectively and fairly compete with our competitors, and at the same time give them opportunity to partner with us, so that together we can build great products.
Were there any protests about your encroaching on their space or anything like that?
Somasegar: So far, from what I understand, I dont think weve had complaints from anybody, particularly because we really opened the kimono and shared with them what we were thinking about and have along the way given updates. Most of the partners that I talked to who are in this space really feel good about this, and they actually feel excited about what they can do to continue adding value for the customers with the tools they are doing. And they feel that with the extensible platform that were are building in VSTS [Visual Studio Team System], they can add more and they feel good about the business model that they have, and they feel good about the value they are going to add. So I havent heard any complaints.
Sridharan: Its not just about Microsoft demonstrating some kind of leadership; its about responsible leadership, thats the key. And 20 or so months ago when we brought all these partners out to campus and we said this is what were doing in terms of lifecycle tools, we laid it out there and said this is where we think the industry is going wrong with life-cycle tools, theyre not very well-integrated, they wont be able to share data. … And we laid it out for them and said this is the problem and the company that can fix it is us. Were going to deliver a platform, were going to build great life-cycle tools and top of that platform and that platform will be extensible. And were telling you a good two to two-and-a-half years before we actually ship, so that you can make any kind of business decisions and technology decisions as well.
Somasegar: And the one data point that I can look at to see how our partners are viewing our entry into this space is looking at how much they are willing to take a bet on VSTS and be able to deliver complementary offerings. We already have a number of partners that have committed to working with us on this and build additional tools to round out the offering. And over the next several months youre going to see more announcements from us as we make progress in terms of getting the partner ecosystem up and ready so that by the time we launch Visual Studio Team System at the end of the summer well have a strong set of partners and offerings from them that will complement what were doing.
Sridharan: And thats certainly at the high end. If you look across the Visual Studio product line, we intend to build out the entire partner ecosystem up and down the product line. If you look at [Visual Studio] Express, we have Amazon and eBay as content partners and our goal is to get as many new people involved in programming as possible. New people dont want to program for the sake of programming. If I ask a random person on the street, “Do you want to learn to program a computer?” theyll go, “No, The Apprentice is on. Id rather go watch that,” right? But if we give them compelling content that attracts them, then we can accidentally teach them how to program. And that is the goal.
And we will build out the entire partner ecosystem, from the bottom with the Express line all the way up to Team System.
Any new partnerships on the Express level?
Sridharan: Not yet, but in a few months. I think there are some really cool ones Im excited about.
Id heard there might be partnerships with game software providers and digital camera makers …
Sridharan: Yeah, imagine an 8-year-old kid wanting to build a game, thats exactly the target market.
Somasegar: The thing that is interesting about the Express editions is when you think about the developers weve been focused on so far with Visual Studio .Net 2003, we sort of targeted what I call the professional developers. And we think there are about six million professional developers around the world. Thats a good number of developers, and they are core to what we are doing, so we love them and we want to take care of them. But if you think about the rest of the Windows customer base … and the Windows customer base is 600 million … my vision is that someday we can build tools that are relevant to the 600 million people or whatever the installed base then is, and not just to the six million professional developers. So if I think about it that way, Express Edition is sort of a first big step in the direction of expanding the number of people who can use tools to do things they really care about getting done in a simple, easy way. Thats the thing that excites me the most. And we think literally there are probably three times the number of professional developers, or about 18 million people, that are potential customers for the Express Edition.
Back to the ecosystem, I totally see where youre going with it and see where companies like Compuware or folks like that might go for it, but then theres IBM and, to a lesser extent, Borland. How do you gauge that whole competition issue with them?
Somasegar: Thats sort of the balance that we walk. IBM is absolutely a competitor for us. At the same time IBM is absolutely a great partner for us. So the key is for us to figure out how when I build VSTS and make it an extensible platform, I really want IBM to play in the platform as well. I will be excited for IBM to come and say, “Hey, we have the Rational developer tool set. Lets figure out how we can make that add more functionality to what you guys are building and really have a complementary offering.”
The reason why Compuware and Borland and Mercury and others are excited and are working with us is because they see those places where they can add more value than we are adding as well as fill in the gaps that we dont have. I would love for IBM to think about it with the Rational developer tool set. And actually IBM should decide how and when they want to get started on this. I want to do everything possible to get IBM to bring their tool set to our platform as much as anybody else. But at the end of the day theyll have to make a decision as to when and how they want to do that.
Check out eWEEK.coms for Microsoft and Windows news, views and analysis.