Interview: Bowstreet founder extols Web services

Frank Moss is now back at the helm at Bowstreet Software, which this week launched Version 5.0 of its Business Factory offering. Bowstreet helped pioneer the Web services category and is now leveraging J2EE-based application servers to provide a new layer of Web services automation. In an interview with InfoWorld Editor in Chief Michael Vizard and Test Center Director Steve Gillmor, Moss, who played a pivotal role in establishing client/server computing, shares his views on why Web services is going to be an even bigger phenomenon and why the major vendors are not to be trusted.

Q: What's your goal for Bowstreet today?Our experience has been that if you look at how people are using Web services, and at the kinds of applications, portals, and applications they're assembling, this is really a style of computing. The way in which businesses use software is going to be revolutionized by Web services and we want to be a big part of it. Although times have been difficult, the premise on which we started the company is coming true more than ever. We started the company with the belief that Web services were going to be big, and we continue to carve out our own unique spot. But I don't even think that we dreamed of how significant the movement would be among the big vendors in the industry. It's our job now to carve out our position in that and ride the wave. The way I see things going, the big picture of the world is that Web services are what computing is becoming.

Q: So in the same way that client/server was a new style of computing, so is Web services?At that time, it wasn't that everything was client/server, but it was the new style. It was the way in which developers thought and business people thought, and it really defined an expectation. Web services, quite frankly, is going to be a bigger wave because I think it packs more potential. The problem is there's a lot of hype behind Web services. The big guys have gotten behind it hook, line, and sinker. And a lot of what they're saying is absolutely true. Web services has tremendous potential for really changing the rules about how businesses use computers and the network to do business. However, I do think there's some peril to the way things are developing right now, to be candid with you. The big guys are talking a good game about working together and about the joys of interoperability and openness, but I do think that each and every one of them is plotting their own strategy [for] how they're going to leverage Web services to obtain global hegemony. I think that could be detrimental in terms of whether or not Web services are really transformational or just another marketing campaign. I think they all know that if they're not the winner, then they stand to lose a lot. So the stakes are high for these guys. I'm delighted to see them working together in these interoperability forums and trumpeting the joys of XML and WSDL [Web Services Description Language] and SOAP [Simple Object Access Protocol]. But the reality of it is, in my view, they're trying to own all this stuff all the way up to the top of the stack. I think there's fine print here that needs to be read by customers to understand what's good and what's bad about all this.

Q: So what is that exactly?

Web services standards are going to be implemented and enable the middleware of these vendors to interoperate very effectively. But the real battleground is going to be at the level above, which is the process by which Web services get assembled into meaningful business portals or applications. Each major vendor is trying to not only own the middleware layer, but they're actually trying to own the top level, which is effectively how the applications are realized. They're taking 30-year-old hardwired application technology, and they're using that in the form of portal service and application frameworks to assemble these Web services. I think this is very detrimental in the long run because first of all the assembly technologies that they're using are outmoded. They're literally hardwiring together Web services. I like to say that they're pouring the wet cement of hardwired applications all [over] Web services. The second problem is that once they get end-users to buy into their portal servers or application frameworks and build business applications out of Web services, then those applications will be tied to their platform. They're only going to run on their application server. I think that's the ultimate form of lock-in that these guys are looking for and that customers need to be aware of.

Q: What role will Bowstreet play in all this?What's really needed is a new, innovative platform-independent layer which we call Web services automation. The platforms for creating and deploying Web services are wonderful. But what the world needs is not for each of these vendors to try and lock customers into the way of using their platform to build applications. I think we're going to need a new category of software, which is platform-independent assembly. That brings us to the point where Bowstreet lives. We are a Web services automation layer that lives on top of these application servers. It enables customers to assemble and automate the assembly of these applications in a very unique and very powerful way. And most importantly, none of this stuff is hardwired. Bowstreet developed a very unique technology and architecture that literally automates the process of assembling these applications, through capturing the process the developers use and then really building these applications at run time. This is a technology whose time has come. It's the automation of software. That's what Bowstreet brings to the table and I think is going to be seen as a whole new category of software.

Q: For a long time, it was never really clear what Bowstreet was trying to accomplish. Why was that?It hasn't been easy for any of us to kind of find our place in the universe, as the world has been moving so quickly. But I could tell you, it's absolutely crystal clear today. We are not an application server or a directory, we run on top of application servers. The Bowstreet Factory, Business Web Factory runs on top of WebSphere and Web Logic and iPlanet. We will run on top of .Net Server in the near future. We are a layer that sits on top, which we call Web services automation. Think of it as really being an alternative to the portal servers and application frameworks that the vendors themselves are trying to establish on top of their platforms. Customers can use the Bowstreet Web services automation to create applications, to assemble them, and to change them over time. But it's been tough because I think what we're seeing emerging is a whole new level of software that didn't exist before, which in a sense it's really a dynamic way of assembling applications. We're replacing hardwired applications with application generation, if you will. That's what Bowstreet does. We use directories to store meta-data, and we sit on top of application servers. I really believe very strongly that you'll see the category of Web services automation emerge from this whole Web services revolution. I would say that we are development and a deployment platform. Developers can take Web services and capture the process by which they connect together those services in an application. It really is a development and deployment platform that sits on top of existing application servers. In a sense, it captures the entire lifecycle of an application and enables line of business people to literally change those applications dynamically without reprogramming.

Q: What tools do you provide to make that happen?We have our own set of tools that programmers use. We call it the Designer, for literally capturing the way in which the Web services are assembled into an application. It uses a process which we call Parametric Modeling. The designer or the developer literally captures the process of designing an application in something called a Builder. It's like a robot that literally automates the process of assembling the application. Then we have a run time that runs on top of application servers, which is the Bowstreet engine that automatically assembles the applications according to the formula captured in that model. Then we have a customizer ' for literally modifying the model without making changes. The whole purpose here is that designers or developers can capture at the very beginning the design of the application. But as this application composed of Web services changes over time, you don't have to go back to the programmer. Line-of-business people can modify it. That's really where the juice is here. In our experience, 80 percent of the cost and complexity of deploying and maintaining these complex Web applications out of Web services is in the add/delete change process. It goes on forever. New Web services are identified and have to be assembled into the application. Web services change and new sources of data are added. All these add, delete, and change are where the cost is.

Q: Why use parametric technology?We have focused on the automation technology that enables these applications to be designed in a completely reusable and dynamic way. It comes from CAD/CAM technology. Our CTO actually came out of parametric technology, where you literally build a model of the application at the very beginning, and that model can be dynamically changed. In CAD/CAM, the analogy to that is where you actually model an aircraft engine out of all the complex pieces using builders. And then if an engineer wants to change the diameter of a shaft, he does that, but then the entire jet engine is updated automatically. In a sense, our applications built out of Web services are like the model for that jet engine. But you don't get automation for free. I mean there's a tremendous, tremendous benefit for businesses to be gained by figuring out how to automate and reduce the manual operations involved in programming these complex Web applications.

Q: There are not too many people familiar with parametric programming models. How do you deal with that?That's just the challenge that we've faced with. We've been dealing with that problem now for over three years, and what we've managed to do, particularly with our new release, is to hide as much of that as possible from the programmer and to make this stuff as easy to use and familiar to use as possible. Programmers do not have to leave the J2EE [Java 2 Enterprise Edition] or the .Net environment. They're really not moving away from their traditional programming environment and we're working with them to create these builders.

Q: Much of what you're saying approaches the area of business process integration. How will Bowstreet play in that space?There's a certain amount of workflow and business process modeling that occurs with our tool. When you assemble these applications and create these builders, you capture your workflow and you model your business processes. There also a lot of work being done right now to do more sophisticated business process modeling, particularly in the Web services arena. So that's an area where we'll probably work with a partner who specializes in that area.

Q: How does Bowstreet tie into legacy applications?Most of what's being assembled today is not orthodox Web services. As people use the Bowstreet Business Web Factory, they're taking a lot of legacy processes, applications, and business logic and assembling those together. The bulk of these Web services are really legacy applications and processes that are coming from within the company. Because Bowstreet has been highly focused on sort of a business Web view of the world, where services can come from anywhere, our customers are really beginning to implement or integrate Web services from outside of the company. But today it is dominated primarily by assembling existing components into applications dynamically.

Q: So for Bowstreet, the key then is to not let the big guys dominate the entire stack?I'm not going let that happen. That's always the risk of a early-stage company that finds itself in a whirlwind of activity. The way in which we deal with that is we have ground wars, where we're going in there. We have a product that not only works today, but it's about to be in its fifth generation. We have big customers and the big guys are still talking about it. We, as an industry are going to miss a tremendous opportunity for Web services because there's a possibility that absent the innovation and the independence of players like Bowstreet, that Web services will come and go and not fulfill as promised just as client/server did not. I wish that this does not end up being just another battleground by which the big guys can fight one another and lock customers in once again. I think it's a tremendous opportunity for them to build their applications in a way that relieves the cost and the burden of management change that's been around for years. Web services could be the biggest event that I've ever seen in 30 years in the industry. I'm hopeful that it develops in a way that the big guys don't own it all and stifle innovation and independence in a way that could make a big difference for customers. And in every generation there does emerge an independent software company that has an important place and kind of leads the pack in spite of the fact that the big guys would like to beat them down. It's a lot of fun to play that role.