4 ms·

Would be nice for them to give examples of what they consider "real progress". Because the pre-1996 list: > LISP, Algol, Basic, APL, Unix, C, Oracle, Smalltalk

Would be nice for them to give examples of what they consider "real progress". Because the pre-1996 list: > LISP, Algol, Basic, APL, Unix, C, Oracle, Smalltalk, Windows, C++, LabView, HyperCard, Mathematica, Haskell, WWW, Python, Mosaic, Java, JavaScript, Ruby, Flash, Postgress Seems just as "weak" as the list they give after 1966: > IntelliJ, Eclipse, ASP, Spring, Rails, Scala, AWS, Clojure, Heroku, V8, Go, React, Docker, Kubernetes, Wasm Basically, after LISP and Algol everything was just a small incremental change over them, some might even say a step back. From the way they phrase "progress", it seems the lineage would look like: Machine Code -> Assembly -> Fortran -> Lisp -> Algol That takes us back to 1958 then, not 1996. I could agree more if that was the argument. That since Lisp and Algol, nothing new in programming language really came out. And if making that argument, the hypothesis around why would be very different. Like it seems it might be more because Software just quickly got bootstrapped and then we did reach a sort maximum with it.

pjmlp 6y ago
Yes, it is also specially sad that the first almost memory safe systems language ESPOL, later replaced by NEWP still in use to this day in Unysis ClearPath MCP, dates back to 1961 and here we are trying to fix C, created 10 years later.
- jacquesm 6y ago
  I thought we gave up on trying to fix C?
  - pjmlp 6y ago
    Companies like Microsoft and Google still believe that while POSIX is around, C needs some fixing as rewriting everything is too expensive. "Checked C" https://www.microsoft.com/en-us/research/project/checked-c/ https://www.microsoft.com/en-us/research/project/checked-c/ https://www.youtube.com/watch?v=EuxAzvtX9CI https://www.youtube.com/watch?v=EuxAzvtX9CI "Improving Your Android App to Prevent Security Vulnerabilities" https://www.youtube.com/watch?v=zkoOD4hmiGE https://www.youtube.com/watch?v=zkoOD4hmiGE
bodhiandpysics1 6y ago
This is one of those canards that people who've never actually programmed in Algol or old lisps (or who forget what its like!) say. The thing to remember about Algol 60 is that it does not have nominal types! It only has (fixed) numerical and array types (it didn't even have records!). Nominal typing was invented by Nicholas Wirth and Tony Hoare for their Algol W, and then became real when Wirth released Pascal (it was also present in the Algol 68 spec, though Algol 68 didn't have real compilers until the mid 70s). That's not a small improvement over Algol 60... it was a massive conceptual change. Before 1978, and the lamdba papers, Lisp was a very different language. Lexical scoping (which technically was invented before in ML) was a massive change in how Lisp worked, because it made the functional style of programming possible.
- amw-zero 6y ago
  That's so interesting. It is crazy (and frightening) how quickly we forget. I can't imagine a world without record types. I was thinking about this the other day. When Djikstra was coming up with his graph algorithm, how the heck did he represent the graph as a data structure??

pjmlp 6y ago

Yes, it is also specially sad that the first almost memory safe systems language ESPOL, later replaced by NEWP still in use to this day in Unysis ClearPath MCP, dates back to 1961 and here we are trying to fix C, created 10 years later.

jacquesm 6y ago
I thought we gave up on trying to fix C?
- pjmlp 6y ago
  Companies like Microsoft and Google still believe that while POSIX is around, C needs some fixing as rewriting everything is too expensive. "Checked C" https://www.microsoft.com/en-us/research/project/checked-c/ https://www.microsoft.com/en-us/research/project/checked-c/ https://www.youtube.com/watch?v=EuxAzvtX9CI https://www.youtube.com/watch?v=EuxAzvtX9CI "Improving Your Android App to Prevent Security Vulnerabilities" https://www.youtube.com/watch?v=zkoOD4hmiGE https://www.youtube.com/watch?v=zkoOD4hmiGE

bodhiandpysics1 6y ago

This is one of those canards that people who've never actually programmed in Algol or old lisps (or who forget what its like!) say. The thing to remember about Algol 60 is that it does not have nominal types! It only has (fixed) numerical and array types (it didn't even have records!). Nominal typing was invented by Nicholas Wirth and Tony Hoare for their Algol W, and then became real when Wirth released Pascal (it was also present in the Algol 68 spec, though Algol 68 didn't have real compilers until the mid 70s). That's not a small improvement over Algol 60... it was a massive conceptual change. Before 1978, and the lamdba papers, Lisp was a very different language. Lexical scoping (which technically was invented before in ML) was a massive change in how Lisp worked, because it made the functional style of programming possible.

amw-zero 6y ago
That's so interesting. It is crazy (and frightening) how quickly we forget. I can't imagine a world without record types. I was thinking about this the other day. When Djikstra was coming up with his graph algorithm, how the heck did he represent the graph as a data structure??