In a word: Fragmentation. Delphi and TurboPascal took a lot of the users and set them on their own paths. Then there’s the official ISO standard (more like a couple) vs what the original author considers the standard. The mainframe crowd and the PC crowd didn’t agree what was what. Nobody really wants to learn 5 different languages to get things done. One true unified language being standard helps a lot. Instead of making Pascal better, the offshoots became competitors to the original.

A teaching language

Pascal started its life as a programming language for teaching. It was carefully crafted to get people started and gently tugg them onto the right way: structured programming. And the language as such if compared to “Your-father’s-C” is much gentler to the novice than C. I think its actually quite amazing that Pascal caught on as it did. I think it is worthwhile to have a look at Algol 68 http://en.wikipedia.org/wiki/ALGOL_68 to get an idea how other languages at the time looked like that were considered to be advanced.

The original Pascal had to be extended in order to “scale” for greater projects (i.e. no notion of modules/units, etc.). Implementations extended Pascal to acommodate the needs of application development. As mentioned: by Philip this lead to fragmentation.

While fragmentation happend to C as well, the real issue with Pascal is, that it was never unified in the same way as C has been. Niklaus Wirth, the creator of Pascal started to develop successor languages (Modula, Oberon) which was probably not the best marketing choice.

Still with Delphi, Pascal had retained a lot of popularity throughout the years, especially as a better alternative to Visual Basic 6. Over the time however support faded and now I think one can only speculate. I think that people considered it on a “professional” scale somewhere between C (and emerging C++) and Visual Basic. With Java and .NET emerging the whole programming language landscape shifted. Lots of projects were probably migrated to Java and the .NET platform eventually and the community lost its drive. That might be the real issue, once lots people “leave the ship” it gets hard for a platform to stay relevant.

Pascal was not so bad after all as some people remember it. Pascal compilers were really fast back then, because Wirth had done a really fine job at grammar definition, it is easily parseable.

Aging with grace

In the long run Pascal has aged with grace before disappearing from stage. It does look outdated from today’s perspective but overall it has been a success story.

A lot of comparisons about its success highlight things that it lacks in comparison to C, however I feel like that with new high level languages emerging, it was actually too close to C, and the world was not big enough for both of them. Today, there is not viable alternative to C/C++ while scripting languages are created by the dozens.

On the Mac, at least, C, C++, and especially OOP happened to Pascal. The original Macintosh Toolbox API was specified using Pascal interfaces and calling conventions, and Pascal was the language of choice for early Mac programmers. C arrived on the platform fairly early, and Pascal and C coexisted for a number of years while programmers on each side argued about which language was better. Compiler vendors (Apple, Think/Symantec, and later metrowerks) provided compilers for Pascal, C, and C++, often in a single product. At the same time, the benefits of object oriented programming were becoming obvious. Apple’s OOP framework was MacApp, written in Object Pascal, and this was surely the framework used to build the first versions of Photoshop (and many other important products). As C++ gained popularity and other OO frameworks appeared written in C++ (Think Class Library from Symantec, followed later by PowerPlant from metrowerks) Apple gave in and rewrote MacApp in C++. That was pretty much the end of Pascal in regular use on the Mac, save for maintaining older code.