Ask HN: Why Did Pascal Fail?

Kernighan wrote that paper after writing "Software Tools In Pascal", and finding it considerably harder than it should have been to write that kind of thing in Pascal. (The predecessor book, "Software Tools", was written in RATFOR, and Pascal should have been miles better to write software in.)

In this book, they (Kernighan and Plaugher) tried to write code that would work in Pascal, not just in one dialect of it. That is, portability was essential to the code they were writing for the book. That may have lead to a greater emphasis on portability in Kernighan's paper than one might think warranted by the general situation.

For some reason, Pascal sat in my mind as an interpreted language. Maybe my whole arguments falls apart then.

Object Pascal is a hybrid language, having an even tighter relationship to Pascal than does C++ to C. You can easily write it without using OOP or classes (or newer things like generics), so that it closely resembles the ISO 1983 or 1990 Pascal.

Pascal is still widely used (categorized as a top 15 to 20 language). That it is not as popular as certain corporate backed C family languages (C# or Java), doesn't mean failure. Saying Pascal "failed", is very deceptive language or arguably having an agenda. That would be like saying Go, Rust, or Swift "failed" because they aren't as popular as C# or Java (with the big corporate money behind it). Keep in mind that Object Pascal is as or more popular than those languages.

Object Pascal was created and used by Apple in the 80s and 90s (used for OS and app development). They preferred Object Pascal over C (does that mean C "failed"). The reason Apple dropped Object Pascal years later for a combination of C++ and Objective-C was partly because they moved from the Motorola 68000 chips to IBM's PowerPC and partly because Symantec's (Think Pascal) and Metrowerks (CodeWarrior) Object Pascal products had beat Apple's Macintosh Programmer's Workshop Pascal compiler in the market. As part of a series of consolidation and strategic moves, Apple moved on to Project Builder IDE that was developed by NeXT (who they bought) and began preferring Objective-C (does that mean C++ "failed"). After the move to PowerPC, a lot of Mac developers were using Metrowerks Object Pascal compiler (CodeWarrior) to port and develop applications. Borland's Turbo Pascal joined the party for a time, but was more successful on Windows. Use of CodeWarrior's Pascal compiler for Macs continued until around 2000 or so.

Object Pascal was used in IDEs and compilers by many companies in the 80s and 90s, among them is Borland (then later to Embarcadero) which shipped Turbo Pascal and Delphi (both very successful well known products) and it was quite popular in the 90s and early 2000s. During this time period, Pascal and then Delphi (the name of the IDE using Object Pascal) was often rated between 1 to 10 in language popularity. Many would say that is success.

Arguably the reduction in use and popularity comes from mishandling by Borland, who got way too greedy, and started charging extravagant prices beyond anything normal people could afford. They abandoned their common user base for enterprise customers, along with getting into an all out corporate war with Microsoft (who appears to have been trying to get rid of competitors on their OS going against Visual Basic and used for app development). This battle concluded with Microsoft picking off Borland's top engineers (including Anders Hejlsberg, who created C#), and Borland being bought by Embarcadero. However, years later, things have "stabilized" with Embarcadero releasing a free Community Edition of Delphi. Along with Free Pascal/Lazarus (a compiler originally written with Turbo Pascal by the way in 1997) slowly becoming more well known, to include Oxygene (Rem Object's Pascal IDE) and PascalABC (numerous YouTube videos in Russian on it) also starting to appear more on the common radar.

Delphi (has academic licenses) and Turbo Pascal (which is freeware now) was and is still used for teaching students their first programming language in various schools, depending on country. PascalABC (open-source) has been taught in Russian schools since the early 2000s. Free Pascal/Lazarus (open-source) has also been popping up in various school systems around the world as well.

Object Pascal has many dialects. Most notable: Delphi (Embarcadero), Oxygene (Rem Objects), Smart Pascal (transpiles to JavaScript), Free Pascal/Lazarus (open-source), PascalABC (open-source and .NET), and DWScript (open-source).

It is also weird to put Pascal into a box, as if the language is not allowed to evolve/change. C and C++ is not the same language that it was 20 years ago. Doesn't make sense to have this expectation for Pascal.

In addition, you can't gloss over that the different dialects of Pascal addressed issues. For those that program with Pascal, these differences are often minor or considered a plus. You can choose the dialect that best suits your needs.

Many would say that it is unfair to bring up Kernighan's 40 year old paper (from 1981). Things in the Pascal world changed almost immediately after it was published.

Turbo Pascal came out in 1983, and addressed various issues. Then Apple created Object Pascal in 1986, addressing even more. Then came Delphi (1995) and Free Pascal/Lazarus (1997).

Take a look at the below link, Which addresses most points outlined in the paper, in terms of modern dialects of Pascal/Object Pascal.

https://wiki.freepascal.org/Why_Pascal_is_Not_My_Favorite_Pr...

Arguably, Kernighan created a "hit piece" paper for AT&T Bell Laboratories (they are on the document).

AT&T was pushing C and then C++, so were very much trying to take down Pascal. This is a document that was "weaponized" to be used against competition, and very strangely, it still being used as such when most of the points presented are no longer relevant.