Umang Gupta, founder of Gupta Technologies Inc, was recently in the UK to explain his company’s strategy for growth in the brave new world of distributed computing and object-oriented technology. As he sees it there are two types of object-oriented trends: one coming from database vendors and one coming from tools vendors. From the tools […]
Umang Gupta, founder of Gupta Technologies Inc, was recently in the UK to explain his company’s strategy for growth in the brave new world of distributed computing and object-oriented technology. As he sees it there are two types of object-oriented trends: one coming from database vendors and one coming from tools vendors. From the tools point of view Gupta feels that his company is very much on the ball with its SQLWindows product cutting it in the market alongside Digitalk Inc’s SmallTalk/V Windows and Micro Data Base Systems Inc’s Object/1. Gupta is in no doubt that object-oriented programming is the way both of the future and of today and programming tools have to take account of this. He argues that SQLWindows is object-oriented to the extent that data and procedures are combined and encapsulated, that variables are defined non-procedurally, that programming is event-driven and that there is an element of inheritance, although not as complete as in SmallTalk. However, the company decided against touting the product as object-oriented for marketing reasons.
In Gupta’s opinion, dBase and Cobol programmers don’t want to hear about object-oriented programming, but do want to hear about enhanced productivity by using graphical tools for writing commercial applications for the SQL market. Gupta Technologies has been accused in these pages in the past of ignoring object-oriented database technology in its SQLBase product (CI No 1,598). This is true to an extent since object technology has gone unremarked by the company when it comes to SQLBase; however, Gupta himself says his company is far from ignoring the potential of object-oriented databases. It is just that in his opinion databases have long life cycles and the SQL market is less than 10 years old, while database cycles tend to go in 20 to 30 year cycles. While taking the point that legacy systems necessarily slow down the introduction of new technology, how scientific such cycle deductions can be in such a young industry is more of a moot point. Nevertheless, quibbling over time aside, most database vendors – apart from Sybase Inc agree that relational databases will evolve into object-oriented databases as there is nothing particularly sacred about tables and rows. However, most of the small relational vendors are waiting for leadership from the major players such as IBM Corp and expect to implement the technological changes via SQL extensions determined by the various SQL standards bodies. SQLBase can currently store objects but the company is looking for guidance as to how an object will be retrievable from a database by an application programmer – this is something the Object Management Group is working on with its discussion on how to formalise an object model.
Is object-oriented programming the inevitable wave of the future or the ultimate turn-off? Gupta Technologies founder Umang Gupta joins the debate in conversation with Katy Ring.
Gupta says that while his company will contribute to the development of object-oriented tool technology, it is not a significant enough player to attempt to establish technological leadership in the database area. This makes sense as the company’s turnover splits between its three products as follows: 50% from SQLWindows, 25% from SQLBase and 25% from SQLNetwork. Gupta is not surprised that the SQL server market has not taken off in the way people expected. These false expectations were fostered, he thinks, by pundits in the personal computer market that are used to the fast take-off of products in a mass market. Whereas the shift to client-server computing will slowly gather momentum and will be driven by the availability of appropriate programming tools. Indeed, Gupta thinks that the SQL server market will grow in a kind of back-handed fashion through the sale of workgroup database software, in which the server is not being sold as a separate entity but is the engine enveloped by user-friendly development tools. In other words, at the programmer level, SQLWindows will drive the growth of SQLBase. While at the end-u
ser level, end-user tools akin to offerings from Borland International Inc, which Gupta will launch shortly, will drive SQLBase sales among end-users. The advantage that Gupta believes his tools have over those of Sybase and Oracle Corp is that SQLWindows is a tool that works with multiple database engines: Oracle, DB2, SQLServer, OS/2 database manager, AS/400 database manager, NetWare/SQL and Informix – meanwhile Teradata Corp and Hewlett-Packard Co are both building connections respectively to DBC/1012 and AllBase.
At the moment 50% of sales are beyond North America, with the bulk of these being in Europe. Over the past six years the company has had a compound annual growth rate of 100% and now has a turnover of $20m. Gupta expects tools to account for nearly two-thirds of the company’s business over the next two or three years as it expects to become the market leader for front end tools – a market that Gupta doesn’t expect to see throttled by standards and commoditised as the server market has been. Gupta doesn’t believe that vendors coming into the MS-DOS market from a background in Unix can offer a strong enough performance on a 16-bit system, because the trade-off in performance by converting products from a 32-bit system is immense. For this reason, while he thinks that the Ingres Windows/4GL product is strong in the Unix world he doesn’t believe it will be a serious contender in the Windows 3.0 market. He concedes that if Unix was to become the workstation of choice with X Window and Motif graphical user interfaces then Gupta Technologies would have a hard time. However, from a strategy perspective Gupta is backing Windows and SQL for the next three years, meanwhile decisions will be made as to whether Unix and RISC architecture are the future.