On November 1988, Byte Magazine published two separate editions; first, the standard monthly issue, focused on the newly introduced NeXT computer; and the “Fifth Annual Extra All-IBM Issue” focused on the IBM PC, both of which are thankfully available on the Internet Archive at the time of this writing. Both of these magazines feature the same expensive advertising on pages 2 to 5, showcasing products from a company called Borland.
The name “Borland” should not, a priori, ring a bell in any Millennial, let alone any Gen Z reading this article. But for a lot of software developers self-identifying as Gen X (like this author) or as Boomer, “Borland” immediately evokes memories of a time long gone; a company that could have been more, but which consciously decided to crash and burn.
And there is nobody to blame about this state of affairs other than Borland themselves. After rising to absolute stardom from 1982 to 1991, seemingly able to kick the mighty Microsoft from its own throne, a series of mishaps and bad decisions brought the company to its knees… and then to oblivion.
Markets are a bitch, and then you die.
It is not the role of this article to tell the story of Borland. There is plenty of material around, starting with a quite complete Wikipedia entry definitely worth checking out. (TL;DR: they were the kings of developer tools.)
Instead, we are going to focus on three major chapters of the Borland story, that can serve as a vantage point for other software businesses that would be interested in leaving hubris aside for a moment.
Compilers And IDEs
The advertising featured in the pages of Byte tell the story of a series of products collectively sharing the “Turbo” moniker. That single word defined a whole category of software development tools: Turbo Assembler, Turbo C, Turbo Debugger… and of course, the one and only, the star of all developer tools in the 1980s, the mighty Turbo Pascal.
In simple terms, Borland was the IntelliJ of the 1980s. Their IDEs were the “State of the Art” to which all other software vendors aspired to. There was not a single product from another vendor, not even Microsoft, who could rival the usability, speed, and price point of a Borland product.
And let us be honest; even today, firing up Turbo Pascal on a virtual machine running FreeDOS is quite an experience. The IDE is simple, fast, effective, efficient, and provides pretty much anything you would need to write and debug code more or less efficiently on a TUI, or Text-based User Interface.
Let us speak for a moment of that iconic TUI with a blue background: the library enabling it was called Turbo Vision (of course) and it consisted of a fully-fledged library for “windowed” DOS applications, similar in spirit to what many TUI libraries offer nowadays. Needless to say, it was wildly popular.
Even better, there is a modern port for C++ and it tells the story of what happened in 1997, when Borland released the C++ version of the library on their (kids, do not laugh) FTP site. These days you can find the project code on SourceForge, and this has inspired someone to create a modern port of it, itself used by the turbo editor application. For the adventurous souls amongst you, the modern iteration of tvision compiles off-the-box on the recently released Fedora 43, but remember to dnf install ncurses-compat-libs first.
Oh, and of course, there is a Pascal version of Turbo Vision bundled for free with Free Pascal, aptly called “Free Vision”. But the legacy of Turbo Pascal is wider than just a library or a single programming language, and we have talked about it extensively in a previous edition of this magazine.
Borland was also amongst the first companies that bet on C++ when few did, and at least during the 1990s, provided the market with the most advanced C++ compiler of its time. To make things even more interesting, this C++ compiler was available for free at a time when good C++ compilers cost in the hundreds of dollars for a single-user license. I know because I literally taught myself C++ using this brilliant compiler, around 2000.
Bjarne Stroustrup took notice and repeatedly mentioned Borland at each HOPL conference talk about the history of C++:
Borland, the largest single C++ compiler supplier, publicly stated that it had shipped 500,000 compilers by October 1991.
(“A History of C++: 1979–1991.” In The Second ACM SIGPLAN Conference on History of Programming Languages, 271–97. Cambridge Massachusetts USA: ACM, 1993. https://doi.org/10.1145/154766.155375.)
The first Borland compiler with “rudimentary template support” shipped November 20, 1991, quickly followed by version 3.1 and the much more robust version 4.0 in November 1993.
(“Evolving a Language in and for the Real World: C++ 1991-2006.” HOPL III: Proceedings of the Third ACM SIGPLAN Conference on History of Programming Languages, 2007, 59. https://doi.org/10.1145/1238844.1238848.)
Early this year some of us have celebrated the 30 years of Borland Delphi, the Visual Basic killer that could have been, if it were not for the inability of a whole company to pivot away as soon as possible from desktop applications to web programming.
The Office Suite That Almost Was
As we said above, Borland tried to compete on the same ring as Microsoft, and it failed; mostly because Microsoft itself redecorated the ring while Borland was busy cheering at the crowd, and by the time it realized what had happened, it was too late.
Borland almost had the “office suite” that could have killed Microsoft’s own (they certainly tried their best). Let us enumerate: it had WordPerfect (think Microsoft Word), Quattro Pro (think Excel), Paradox (think Access), and… I think that is it. I do not know if there was an equivalent for PowerPoint; if there was not, maybe that was the killing bullet: could you conceive not having presentations slide software on your office suite?
Borland even wanted to provide an alternative UI to Windows 3.1 called “Dashboard for Windows”, but without success (again, I have to say I used it, and I was fond of it). At the time this article hits your RSS reader, Quattro Pro and Paradox are still available as part of a thing called WordPerfect Office but in the age of LibreOffice, Google Docs, and AI copilots, I seriously wonder if they are selling any licenses at all today.
Some say that the demise of the Borland Office (or whatever its name was) started the day Borland bought Ashton-Tate, the makers of dBASE, the leading database application for PCs running DOS… but we will never know for sure.
Social Contracts
Borland, just like Lotus, belongs to a time when tech millionaires and billionaires actually thought they could improve the state of the world in meaningful ways. This is no longer the case, but the Wikipedia page of one of Borland’s founders, Philippe Kahn, makes this point abundantly clear:
Under Kahn’s direction, Borland became the first software company to offer domestic partners full benefits and a pioneer for gay rights in Silicon Valley. Kahn was a key speaker at the pivotal gay rights conference on the Apple campus on October 19, 1993.
Borland was also famously involved in the legality of software patents and intellectual property… taking a very simple approach for end users, while openly fighting other corporations on the same topic:
Borland’s approach towards software piracy and intellectual property (IP) included its “Borland no-nonsense license agreement”; allowing the developer/user to utilize its products “just like a book”. The user was allowed to make multiple copies of a program, as long as it was the only copy in use at any point in time.
A no-nonsense license agreement. Think about that.
Now the question is: how was Borland able to pull such good quality software? It all boils down to people and culture, as usual. Jim Coplien wrote a paper in 1993 called “Borland Software Craftsmanship: A New Look at Process, Quality and Productivity” where he describes his experience spending time with the Quattro Pro development team whose processes are, according to Coplien, “off the charts”:
The Borland Quattro Pro for Windows (QPW) development is one of the most remarkable organizations, processes, and development cultures we have encountered in the AT&T Bell Laboratories Pasteur process research project. The project assimilated requirements, completed design and implementation of 1 million lines of code, and completed testing in 31 months. Coding was done by no more than eight people at a time, which means that individual coding productivity was higher than 1000 lines of code per staff-week.
Despite such brilliance, despite the popularity of Delphi, and thanks to the changing market conditions of the 1990s, Borland’s name slowly faded away from the minds of developers. Incomprehensibly enough, and adding insult to injury, the company was renamed as “Inprise” in 1998, later again as Borland in 2000, then confusingly enough referred to as “Embarcadero” and then as “CodeGear”, because if you want to alienate your customers, nothing beats renaming yourself to death.
Let this article be a memorial for those white haired developers amongst us who have fond memories of the good old times of Borland and all the Turbo things.
Cover photo by Charlotte Coneybeer on Unsplash.