S is for Snobol 4, Simula 67, Smalltalk, sed, SQL, SETL, Scratch and Spreadsheet.
Snobol 4 was the go to language for text generation and pattern matching before PERL/Python became widespread in the late 1990s (if your PERL/Python code is too slow try Snobol, compilers generating machine code have been written for it since the 1970s). It is a boutique language that entrances many people who encounter the Green book (your author fell under its spell as an undergraduate).
Simula 67 was the first object oriented language. However, 30 years went by before object-oriented programming could be said to have entered general use (some people would claim that this has still not happened). Yes, lots of developers were using C++ compilers in the late 1980s, but only because they wanted to look trendy and be able to list the language on their CV. During the late 1980s and early 1990s I was heavily involved with the static analysis of C using an in-house tool and found that a lot of C++ source could be handled by adding a few extensions to the tool, such as handling
new (rather than
malloc) and allowing
class to be used where
struct should have appeared. It was not until the mid-90s that parsing C++ source with an extended C parser became problematic on a regular basis (and then often only because of the use of iostreams).
Back in the day it was said that only Scandinavians understood OO programming. I think the real reason was that only Scandinavians understood the English used in the book written by the language designers: “Simula BEGIN” by Birtwistle, Dahl, Myrhaug and Nygaard (there is a rave review on Amazon, I wonder if the reviewer is Scandinavian).
Smalltalk is invariably the language of choice for object-oriented purists. I know of several development groups in industry that choose to use Smalltalk because they wanted to really learn how to do things the OO ‘way’ (Smalltalk is all objects, its users cannot do anything in a non-OO way); I don’t know what those subsequently responsible for maintaining of the code thought about this rationale.
The ‘launch’ of Smalltalk in the UK by Goldberg and crew happened before Apple’s Macintosh made Xerox PARC famous, and the talk was all about message passing being the future and everybody being able to modify any behavior of their operating environment (we were shown Smalltalk’s flexibility by a speaker who changed the definition of an operation that defined how the borders of windows were displayed; they agreed that this was potentially dangerous but that users soon learned by their mistakes). The launch of the Macintosh changed the whole PARC narrative.
Thinking back, the only thing that stopped the Smalltalk team inventing Open Source was the fact they thought it so obvious to ship the source of everything that it was not worth mentioning.
sed gets a mention as the most unlikely language used to implement a Turing machine.
SETL is a Cambrian explosion language which has had a small, but influential, user base.
Scratch suffers from being thought of as a language for children. If touch screens had been around when people first started inventing programming languages I’m sure things would be very different today, where most developers think that real programs are written in lines of text.
Spreadsheet languages have for a long time been a neglected topic from the perspective of tools to find faults and programming language research. Things are starting to change in industry and hopefully the work of keen academics will be appreciated in academia.
Things to read
The SNOBOL4 Programming Language by Ralph E. Griswold, J. F. Poage, and I. P. Polonsky.
Every language has its blind spot, the techniques that should never be written using that language. In the case of SQL it is tree data structures: “Joe Celko’s Trees and Hierarchies in SQL for Smarties” by Joe Celko.