DBMS Interview - August 1995 DBMS Interview - August 1995
Powersoft Corp. CEO Mitchell E. KertzmanThe Accidental CEOBy Theresa RigneyMitchell E. Kertzman explains how he fell into the development tools business and sets the record straight about the Powersoft/Sybase allianceIn 1968, Mitchell Kertzman dropped out of college to be a disc jockey at WBCN in Boston. He was a folk musician in Boston in the '60s, and thought it was likely that he would make his living in some part of the music industry. However, in order to eat, Kertzman took a job at an educational software company specializing in audio-visual production. Intrigued by a time-sharing terminal connected to a commercial time-sharing service, he tried his hand at programming. The rest, as they say, is history. In two years, Kertzman was running all of the development in this company.In 1974, Kertzman founded Powersoft Corp., a small software consulting business for MRP applications. In the mid-1980s, Kertzman began to see a shift to client/server, and hired David Litwack, former head of research and development at Cullinet, to create new software for this burgeoning architecture. In June 1991, Powersoft released its flagship client/server application development product, PowerBuilder. Powersoft is now set to release the fifth major incarnation of this product, PowerBuilder 5.0.Perhaps the biggest news for Powersoft, however, was its acquisition last November by Sybase Inc. In a stock trade worth almost $940 million, Powersoft aligned itself with this database vendor, a move that has raised concerns about Powersoft's (and PowerBuilder's) continued database independence. One of the main reasons for PowerBuilder's success was that it let developers build applications without concern for the underlying database(s). In May, Features Editor Theresa Rigney sat down with Mitchell Kertzman at Sybase's users conference. In the edited transcript that follows, Kertzman confronts the database independence issue and discusses PowerBuilder's future.DBMS: I read that you kind of fell into the database tools business. How did your career begin in this industry?Kertzman: I was actually going to be a disc jockey for a living. I dropped out of college to be a disc jockey at WBCN in Boston. But I was an unemployed disc jockey in 1968. At the time I think I was pumping gas at a gas station, and winter was coming, and she decided to try to find me an indoor job. So through a friend of a friend, she found me a job at an educational software company doing audio-visual production. This company had a mainframe with which it ran its business. It also had a time-sharing terminal connected to a commercial time-sharing service and I thought, "Gee, that programming looks kind of interesting," so I asked my boss's boss if I could try it. He said -- and this was one of the great decisions affecting my life -- "Sure, you can use $100 of time, and then tell me what you did." So I used my $100 of time, and he let me do some more, and in two years I was running all of the development in this company.Programming was something I took to. I had no background or training in it. It was something I was good at and really loved. I still remember when my first program ran successfully. All it did was print the numbers between one and 20 in the left-hand column of a page, and the square of those numbers in the right-hand column. When that thing ran on an old Model 33 teletypewriter, I was hooked. It was like I had a physical addiction to programming.I started my first business in 1972, because I decided I wanted to be my own boss. My first idea for making money through programming was to create a program to print biorhythms. I took out an ad in Rolling Stone saying, "Send me your birthdate and I'll print you a biorhythm." That wasn't a big money maker.My second business was called "The Music Machine." This business matched rock-and-roll bands with gigs and bookings. I had a database of bands, and colleges or clubs could book a band through me. They could tell me what kind of band they wanted (a certain style of music, a female lead singer, and so on) and I would massage the database and come up with a list of bands that matched the criteria, and I would send them the bands' phone numbers. The fault in that business plan and the reason it failed was that it had a business model that relied on the business integrity of rock-and-roll musicians. This was a bad foundation for a business. Rock-and-roll musicians don't pay their bills.At that time, I was in a position where I still wanted to run my own business, and I was a pretty decent programmer, so in 1974 I incorporated the company that became Powersoft. But my business plan could be effectively summed up as: "Will program for food." If someone had a business and wanted something computerized, I would help them. Back in 1974, people couldn't afford their own computers. Even minicomputers (which were just coming out) were pretty expensive. Most companies were using batch service bureaus; they would fill out paper forms with customers' billings and things, and these service bureaus would punch them in and send back reports. Time sharing was getting more popular, and so I found a company that had some excess time on a time-sharing system and set myself up in business to do custom programming and sell time-sharing time: I would write the programs and my customers would run them using my service, and I made money at that.It turned out that one of the companies interested in betting its business on a one-person software company (especially when that one person had long hair, a beard, and drove a 1964 Rambler American) was a small manufacturing company. I automated this business with job costing, accounts receivable, accounts payable, and so on. When I successfully completed that project, the president of the company had a friend who had another manufacturing company, and I ended up, by happenstance, specializing in software for manufacturing companies.So how did Powersoft move from being a manufacturing applications vendor to a development tools vendor?It was a one-person business for almost five years, and it gradually grew into the minicomputer MRP business on the HP 3000 platform. In 1987 or 1988, I felt that we had the opportunity in that business to take a leading role by rewriting our software to be what we now think of as client/server: graphical, distributed, relational, and object oriented. We couldn't find any development tools with which to do that, so we brought David Litwack into the company, developed PowerBuilder, and the rest is history. We sold the original manufacturing business in 1992.Did you have any idea that client/server would take off like it did?Not as much as it did. Powersoft's success certainly exceeded my expectations. The history of the computer industry showed me that the opportunities in software exist when there are paradigm and platform shifts in computing. Big businesses stumble and small companies get opportunities to succeed. An example of that in our history would be Cullinet, which absolutely dominated the mainframe database software business. When Oracle came along, [Cullinet] didn't take it seriously, saying, "Don't bet your business on a little relational database; it doesn't have all the features we have." But Cullinet didn't understand that there was a fundamental paradigm shift to relational (and, along with that, a move toward minicomputers and Unix computers). As a result, Cullinet stumbled and Oracle succeeded.It turned out that the way to build a big software company is to be the leader in a product category during a paradigm shift. Think of Lotus: It was the leading spreadsheet on DOS and it missed the move to Windows. But Powersoft took advantage of that opportunity by really coming out of nowhere to be the leader on this new platform when it emerged, and when client/server emerged. Nobody expected the leading development tools company to come out of some minicomputer-based MRP company. So the growth of client/server was a pleasant surprise, and the growth of Powersoft was an more pleasant surprise.What factors are changing the application development market now? How will it be different in three to five years, for example, and how will Powersoft adapt?The application development market will change and mature along with client/server. One of the things that is happening now, and will happen over the next few years, is that the client/server platform is becoming more stable, robust, and mature. For the first time, things like system management tools are becoming available, which allow people to manage this distributed, heterogeneous platform the way they used to manage the mainframe. I think that is one of the key things customers are looking for in order to move their larger applications to client/server.Powersoft's view is that client/server will move from the tactical arena to the strategic arena; that is, people will really start to rely on it for their strategic applications, their so-called mission-critical and large-scale applications. And, whereas client/server really started out in the departmental and workgroup arena (that's where we have been most successful), the action in the next three to five years will be in the enterprise -- building larger, more robust transaction processing applications. Powersoft is committed to extending its leadership as the platform becomes more robust.People have mistaken Powersoft's strategy for a fundamental architectural limitation. But our strategy has always been to aim our products and their capabilities at what we think is the "sweet spot" in the market. We believe in developing products that are oriented to where people are buying. Then we add capability as people want it. Powersoft has always been very good at understanding the dynamics of the market, and understanding the functionality that people are ready and willing to implement (and that the platform can handle).PowerBuilder 5.0 will be a surprise to a lot of people in terms of its enterprise capability. That's where we're focusing our development energy now. As DBMS readers read this, we will be shipping, or ready to ship, the Macintosh and Unix versions of PowerBuilder. The product will then be cross-platform, which has been an important strategic consideration for our customers.Now, where will Powersoft be in the future? I hope that -- and our job is to make this so -- in three to five years we will still be the number one supplier of development tools for client/server.You mentioned that PowerBuilder 5.0 is imminent, but how long can you extend the current code base without starting over?That is always a concern eventually, but I don't think we're near the limit of the current code base yet. PowerBuilder is only four years old. Once you have built on an object architecture, which is what PowerBuilder is built on, my guess it that your old notions of the life of a code base start to change, because the entire architecture is much more object-oriented and modular. If you've done it right, you shouldn't have to tear up the streets at some point to rearchitect your monolithic code base.Over the last 10 to 15 years, several high-profile products have enjoyed and lost market leadership. (You mentioned Lotus.) If this were to happen to PowerBuilder, what do you think would be responsible?If I were to put myself in the future and ask, "Why did this happen?", I guess it would happen for the same reasons it happens to other companies. Market leaders can get arrogant and complacent, and they believe that they can extend their market leadership forever. They never see the two people in the garage who have a better product, they never listen to their customers, or whatever. One of the things that has always been a part of the Powersoft culture is that we're paranoid (the only other place where I see it as predominately is at Microsoft). We don't get up every day and say, "Isn't it wonderful to be the market leader? The sun shines only on us." Rather, we get up every morning wondering who's out to get us and assuming that we need to keep on our toes in order to be successful.Not long ago, PowerBuilder and Gupta's SQLWindows were considered the two market-leading application builders, but more developers seem to be choosing Visual Basic and PowerBuilder. Who do you see as your major competitors now and two or three years from now?Microsoft and Microsoft. We have felt for a couple of years that Microsoft is our number one competitor. If you look at the marketshare, PowerBuilder and Visual Basic are neck and neck. (We are a little ahead in the client/server space.)Are Oracle Power Objects or Borland Delphi anywhere on the horizon?The client/server tools business is interesting because it's theoretically supposed to be open, which means that customers have a choice. If you lose your leadership and cease to be innovative, people will switch. They can use Delphi for their next project instead of PowerBuilder. Yes, there's some learning curve there, so people are less productive in the beginning, but basically they can pick and choose.We keep our eyes open for other competitive products. The single competitive advantage of Delphi is compilement; it can compile directly to code. But PowerBuilder 5.0 will have a complete compiler, so that advantage will go away. And then, of course, you have all the Borland issues.Oracle has a tool du jour. It has Power Objects and it has CDE2, which is now Developer/2000 and Designer/2000. Oracle seems to be focusing Power Objects on Visual Basic, but I don't see how it can compete with Microsoft in the Visual Basic space. Microsoft is really good at the low-end part of the market. When we compete against Visual Basic with our PowerBuilder Desktop, the primary advantage we have is that PowerBuilder Desktop is compatible with PowerBuilder Enterprise and it scales up. Visual Basic doesn't scale up. Power Objects doesn't scale up; if you want something more powerful, you have to move to Developer/2000, and there's no compatibility between those two Oracle tools. This is the weak part of Oracle's strategy, and I don't think it can "out-Microsoft" Microsoft. Nobody can.When Powersoft merged with Sybase, your competitors criticized -- and still criticize -- you for losing your database independence. How do you respond to those criticisms?It's not surprising that they would say that, because that's what they have to say. What I would ask anybody to do is judge us by our actions, not by the threats, words, and activity of our competitors. Since we have merged with Sybase, we have announced a new native database driver for Microsoft SQL Server 6.0 -- a direct competitor with Sybase. We are in discussions now with other database companies that compete with Sybase to provide direct drivers to their databases.Basically, if Oracle comes up with a new version of its database and we don't support it, come up to me and tell me that we are not being open. But, so far, our track record is that we are continuing our strong commitment to open database.During my first conversation with Mark Hoffman [CEO of Sybase Inc.] in July 1994, and all the way through today, there has never been a single discussion about limiting or restricting PowerBuilder in any way. We haven't discussed either cutting back support for other databases, or even making PowerBuilder work better in some exclusive way with Sybase SQL Server than it does with any other database.Some people also criticize PowerBuilder for creating "fat clients," with too much logic in the application. Do you think the concept of three-tier architectures is the right solution for this problem?In the sense that, to date, PowerBuilder runs on the client, then that is where the code resides. However, we recently did a demo in which PowerBuilder ran on the client, and PowerBuilder ran on an application server in the middleware. Because Powersoft supports the concept of a non-visual object (an object that does processing but does not interact with the user) PowerBuilder has always been able to run on a server. However, we haven't had a server-based strategy. That is coming in version 5.0. The fact that PowerBuilder runs on Windows NT, will soon run on Unix, and supports non-visual objects, means that it will support partitioned applications, all of the three-tier topologies (DCE and all the transaction processing monitors), and distributed object computing (OLE, CORBA, and so on).Just as we have been database independent, network operating independent, and are soon to be client operating system independent, our strategy is to be application topology independent. Whether you want to write two-tier applications, which will be the majority of applications, or whether you want to be three-tier with TP monitors or DCE, Powersoft will support you.Is PowerBuilder built on an object repository?Not today. One of the benefits we get from our relationship with Sybase is that it had been working on a product called Enterprise Momentum that included some of that kind of technology, and you will see some announcements about that technology in the future.Visual Basic has enjoyed a huge third-party market. Does Powersoft plan to grow a similar third-party market?There already are a lot of frameworks and tools available for the PowerBuilder developer. We have a robust third-party environment out there, although not as robust as Microsoft's. We intend to continue that, but you can use several VBXs today in the PowerBuilder environment. More important, we support the creation and utilization of OCXs for OLE 2.0. We expect to have not only our own PowerBuilder controls and add-ons, but also to take advantage of those developed for Visual Basic.Powersoft recently purchased SDP Technologies' S-Designor CASE tool. Why S-Designor and not another CASE tools such as Logic Works' ERWin, with which PowerBuilder already has a relationship?We still have a relationship with ERWin, as well as other modeling tools, even though we own S-Designor. We thought it was important to have an offering there, because database design is becoming more and more important to the success of an application development effort. Because PowerBuilder is so widespread, we've seen a lot of successful development projects, but we've also seen a lot of people struggle with it. So we spent a lot of time asking, "Why are people successful, and what gives them the most problems (particularly with performance)?" What we found was, in many of the cases in which people had performance issues, they had ignored database design. Part of the reason was the illusion of the ease of use of the GUI -- that with a few point-and-clicks and drag-and-drops, you've got an application. But people were forgetting that database design wasn't any less important to the performance and functionality of an application.We now offer training courses in database design, but we thought it important to also have a tool to sell. We had a number of discussions with companies, but the chemistry was right with SDP, the value was correct, and it shared our vision. The fact that it was strong in Europe was a plus.CASE seems to be making a comeback in the application development spaceżPeople are discovering the difference between CASE and I-CASE. What gave people fits was this concept of super-large CASE environments, in which everything came off of some high-level model, and down at the end you got maintainable code. But it turns out that there is a more tactical approach to CASE; the individual approach to modeling and design turned out to be very useful. There's a new generation of tools that have a client/server mindset; they're PC-oriented, graphical, easy to use and install, and quite powerful. I think we're seeing a resurgence of interest in the individual components of CASE, but not in some big methodology.In addition to Lotus Notes, will PowerBuilder support other nonrelational servers, such as OLAP or object databases?Yes. PowerBuilder has always had a view that it wasn't just a client talking to a database; it was a client talking to network services. Just as we aren't restricted to two tiers, we're not restricted to pure SQL relational databases. And not only OLAP, we're also looking into extending the network into a broader network such as the Internet, where you can have applications running with servers on the Net. We're also looking at wireless client/server applications. The platform is still expanding.What is your main priority for moving PowerBuilder into the future?The number one priority is to stay focused on customers and what they're doing and asking for. In the near term, we're going to be very focused on taking every opportunity to demonstrate our continuing commitment to openness. We will not rest until we put that issue to bed.Another priority is extending PowerBuilder into these new application architectures, whether it's three-tier, partitioned, or whatever. We will also continue to explore the positive aspects of the combination with Sybase.Do you still program?I don't do development any more. In fact, my old software code is used as a negative example to new programmers -- what not to do. "If you write code like this, you will most certainly be fired."Powersoft Corp., 561 Virginia Rd., Concord, MA 01742; 800-395-3525 or 508-287-1500; WWW address: http://www.powersoft.com.Return to the DBMS home page. (http://www.dbmsmag.com)Copyright © 1995 Miller Freeman, Inc. ALL RIGHTS RESERVEDPlease send questions or comments to mfrank@mfi.comUpdated Saturday July 1, 1995 |
|
