How Borland Blew It All
A nostalgic yet critical history of Borland — from the revolutionary Turbo Pascal that changed programming forever, through the glory days of Delphi, to the strategic blunders and leadership failures that destroyed one of software's most iconic companies.
First Encounter with Turbo Pascal
I first saw Turbo Pascal 5.0 in 1989 and was stunned by how revolutionary the approach was. Before that, development required a multi-step process: writing code in an editor, compiling, linking, running the program. If errors appeared, you had to repeat the entire cycle.
Turbo Pascal offered an integrated environment where everything happened in a single window. A single keystroke compiled and ran the program; errors appeared right in the editor with the line number highlighted. What used to take a week of compilation now took two hours.
1981: The Founding of Borland
Three Danes — Niels Jensen, Ole Henriksen, and Mogens Glad — founded a firm called MIT for software development. After visiting a trade show in San Francisco, they moved the business to the United States.
There they met Philippe Kahn — a man with a mathematics background and musical experience who was initially in the country illegally. Kahn proposed the name Borland ("forest land" in Celtic) and became CEO for 12 years.
1983: Turbo Pascal
Kahn and the brilliant programmer Anders Hejlsberg created a fast integrated development environment. Compilation took no more than 5 seconds instead of the usual 5–30 minutes.
The product sold for $49.99 (versus $300 from competitors). In the first month, 3,000 copies were sold ($150,000). Two years later, Byte magazine reported 250,000 copies ($12.5 million). Eventually — 400,000 copies ($20 million).
Local banks suspected the company of fraud due to the unusually high transaction volumes.
1985–1990: Expansion
Compilers for other languages appeared: Turbo Basic, Turbo Assembler, Modula-2, Turbo Prolog, Turbo C.
In 1990–92, they added object-oriented programming and the Turbo Vision library for creating programs with windows and menus. This library became a textbook example of OOP for many programmers.
The company expanded in all directions: Eureka (differential equations), SideKick (office tools), Quattro Pro (spreadsheets), and acquired Paradox and dBase.
However, the attempt to merge disparate divisions (development languages and databases) led to financial problems and layoffs in the early 1990s.
1993: Windows 3.1
The arrival of Windows made DOS-oriented tools morally obsolete. A new environment for Windows application development was needed.
Object Vision (1992) let you assemble windows with a mouse, but it ran slowly and had limited functionality. The $495 price tag (later reduced to $250) drove users away.
Pascal programmers were lost: some switched to C or Visual Basic, some left the profession entirely.
1995: Delphi 1.0
In the same year that Philippe Kahn left the CEO position (due to disagreements with the board of directors), Delphi 1.0 was released at $49.95.
Delphi offered:
- An integrated Windows environment with RAD (Rapid Application Development)
- Visual form design by mouse
- Components written in Pascal, allowing customization
- Universal access to any database
- SQL support
The environment required only 8 MB of memory (Visual FoxPro required 32 MB) and ran efficiently.
1996: The Departure of Key Figures
Anders Hejlsberg left for Microsoft (receiving $3 million as a signing bonus). There, he went on to create C# and Visual Studio.
Delphi 2.0 was released in three editions (from $500 to $2,000), which hampered distribution in the West. Visual Basic was essentially free as part of MS Office.
Management declared that development tools were no longer a strategic priority. The focus shifted to application lifecycle management tools — which never gained popularity.
1996–2000: Stagnation
Delphi evolved linearly with minimal improvements. MIDAS and OLEEnterprise appeared but quickly vanished.
It felt like the team lacked the resources or the will for innovation.
The 2000s: Flailing
Kylix (2003) — a development environment for Linux — lasted only a year or two.
InterBase went open source, which led to the emergence of a more capable clone, Firebird. Borland itself returned to a paid version that lagged behind in features.
Bold for Delphi — one of the first ORM systems, built on UML models, but performed poorly.
ECO (for .NET) could quickly build a simple program but generated inefficient database queries.
When Microsoft released .NET, programmers migrated en masse to C# and Visual Studio (2005–2010).
Borland tried to adapt Delphi for .NET but couldn't offer convenient components. Instead, they recommended using Windows Forms from Visual Studio, later removing them and offering VCL.NET instead.
This behavior pointed to a deep leadership crisis.
Epilogue
In the late 2000s, the remains of Borland were sold to Embarcadero, which continues to sell Delphi today.
Philippe Kahn created several new companies. He sold one of them to Motorola for $325 million. He later moved into fitness wearables.
The core lesson: A company's success depends first and foremost on leadership's understanding of the strategic direction of development.