TG: Programming: Where Delphi Went Off the Tracks

Home - Programming - Where Delphi Went Off the Tracks

In college, I was complaining to a good friend of mine, who worked in the PC lab, that I wished I could find a code compiler that didn't cost hundreds of dollars. He introduced me to Turbo Pascal 3.0. And that was the moment that launched a long and highly fulfilling career.

I built database systems, integrations, a BBS, and a number of other systems that made me an appreciable amount of money with that compiler. From there I jumped to Borland Pascal 7, building shipping systems. Then it was on to Delphi 1 porting the shipping system from DOS to Windows. Using Delphi 2, 4, and 5, I built case management, auction, fiduciary, customer, inventory, and sales software. Using Delphi 5, 7, XE, XE2, XE3, and XE5, I worked on enterprise point of sales systems, licensing, customizations, rolling REST and RPC frameworks by hand, etc.

So yeah. I'm a fan. And I didn't use but a fraction of what Delphi could do.

During that last run, Borland became Inprise, which reverted to Borland, which became Embarcadero. And each time it morphed, it lost something. When Borland first came out with Delphi 1.0, it rocked the coding world. Based on technology from the Zinc Library, it was like no other programming experience I'd ever had. It took Windows programming and made it easier than DOS programming. Then Delphi 2.0 came out and Borland was producing 32-bit software. And it was AMAZING.

But then something happened. I'm part of the camp that blames the initial problem on Anders Hejlsberg and others getting poached by Microsoft. But other companies recover from that kind of speedbump. Not Borland. Development slowed to a crawl. Quality suffered. Microsoft debugged and expanded its .NET tools, while Borland seemed to just fall behind. It took a ridiculously long time to get Unicode and 64-bit support. Delphi.NET didn't take off, and Delphi in general has lost market share. In any list of modern languages, it's not even at the bottom. A recent UC Davis study listed 17 languages, even such fringe languages as Haskell and Clojure, but no mention of Delphi or Object Pascal. Given that Delphi is a development environment rather than just a language, why does Eclipse beat it for use as an IDE when Delphi has more base functionality and potential for expansion?

I'll tell you why: marketing.

Borland\Inprise\Embarcadero's marketing strategy sucks. They should be working hard to build a development community. They should be subsidizing open source developers that use their product to build new and innovative systems. They should promote the experts, tout the successes, and it should HUGELY visible. But instead, being a Delphi developer is like being a COBOL developer on a Data General system -- people picture an old guy with a gut, a white beard, and suspenders; a dinosaur. Or they assume that, if you're not ready to be put out to pasture, you're from Russia, India, or China and you're using a pirated copy of Delphi. Because what self-respecting programmer uses (disgusted sneer) Delphi?

I personally don't think it's too late for Delphi -- it's just too darned good, too darned easy to use -- but their marketing department needs to be replaced, they need an evangelist, and they need to get serious about making an IDE for more than just Object Pascal.

Update: 2/9/2017
I'm not wrong about Delphi needing a better marketing department and an effective evangelist, but today I might have to find a decent recipe for crow. The factors that are finally pulling Delphi out of a very long spell of doldrums are 1) it's increasingly effective cross-platform development capabilities, and 2) FreePascal. At this point I'm personally aware of several developers using Delphi's IDE to build apps for iOS and Android as well as Windows. And that's just pure, unadulterated, bad-ass. And then there's FreePascal, the Object Pascal compiler for Linux which actually has nothing in common with Delphi outside of Object Pascal and the imperfect but impressive ability to compile Delphi third party libraries with little to no modification.

And that's probably why, on the TIOBE index, "Delphi/Object Pascal" ranks #9 for languages in use today, up from #10 last year. This is huge. I can remember when, just a few years ago, Delphi #20 on TIOBE and not listed at all almost anywhere else. It was just plain embarrassing to admit to anyone that knew anything about programming that I was using Pascal.

To be clear, TIOBE is based on net searches as an indication of interest and usage. IEEE lists it at #28 below such winners a LISP and VHDL and just barely beating out Prolog and Clojure. Almost every other developer-based poll out there still doesn't list Delphi or Pascal. It's so bad that, when quoting TIOBE, other sites will literally omit Delphi just because.... well... It's *Delphi*.... But at least it's not entirely dead. There are signs of life yet...