3 ms·

I have been programming in Lisp for 50 years, all the way back to a Lisp 1.5 version. Every few hours someone decides to write lisp with a different syntax, on

by daly 7y ago

I have been programming in Lisp for 50 years, all the way back to a Lisp 1.5 version. Every few hours someone decides to write lisp with a different syntax, one that lacks parens. Yes, it can be done. No, it shouldn't be done.

brudgers 7y ago
The Racket community's interest in programming pedagogy appears to be focusing on extending the Racket ecosystem to better facilitate the creation of "infix" languages, e.g. Flatt's recent remarks on "Racket2." [1] The Racket community appears to be creating a world where students can make baby Python's using a Lisp instead of vice versa. To the degree that SICP is ground zero for Scheme dialects of Lisp extending the Racket ecosystem with an eye on making diverse languages is in the central current of Scheme's history of creating increasingly more complex interpreters...there's nothing wrong with syntactic sugar. I suspect the goal is to allow educators to express CS concepts in the context of Python without all the noise of Python's incidental complexity. [1]: https://news.ycombinator.com/item?id=20490423 https://news.ycombinator.com/item?id=20490423
- daly 7y ago
  If you don't like parens, don't use lisp. There are thousands of other languages discovering things like "lambda" as if it was a new idea. It is easy to create a domain-specific language on top of lisp if you want something with a specific focus. But trying to create a full language that hides lisp syntax is pointless. It's been tried SO many times and failed exactly that many times. There is so much to lose. You lose homoiconic syntax which usually means you lose macros. You can hack macros by translating back to lisp, but why? I can find hundreds of "lisp without parens", just do a google search. For example, there is Wisp. You can generate lisp "without parens" by a simple hack. Push all of the parens to the right side. (+ (* 2 3) 6) becomes (ignore the underscores. I don't know how to generate verbatim text in HN) _____________________________________( + _____________________________________( * 2 3 _____________________________________) 6 _____________________________________) Now if you "adjust your margins", all of the parens disappear off the page. Lisp is a language that everyone says "I don't get it and I don't like it" until.... they get it and then they LOVE it. That's when they learn the limitations of all the other languages.
  - brudgers 7y ago
    The package's author is near the center of the Racket community. https://www.asumu.xyz/ https://www.asumu.xyz/ It lives alongside the Racket DSL for Algol 60 https://docs.racket-lang.org/algol60/ https://docs.racket-lang.org/algol60/, Datalog https://docs.racket-lang.org/datalog/index.html https://docs.racket-lang.org/datalog/index.html, and ProfessorJ https://docs.racket-lang.org/profj/index.html https://docs.racket-lang.org/profj/index.html in the Racket ecosystem.

by daly 7y ago

brudgers 7y ago

The Racket community's interest in programming pedagogy appears to be focusing on extending the Racket ecosystem to better facilitate the creation of "infix" languages, e.g. Flatt's recent remarks on "Racket2." [1] The Racket community appears to be creating a world where students can make baby Python's using a Lisp instead of vice versa. To the degree that SICP is ground zero for Scheme dialects of Lisp extending the Racket ecosystem with an eye on making diverse languages is in the central current of Scheme's history of creating increasingly more complex interpreters...there's nothing wrong with syntactic sugar. I suspect the goal is to allow educators to express CS concepts in the context of Python without all the noise of Python's incidental complexity. [1]: https://news.ycombinator.com/item?id=20490423 https://news.ycombinator.com/item?id=20490423

daly 7y ago
If you don't like parens, don't use lisp. There are thousands of other languages discovering things like "lambda" as if it was a new idea. It is easy to create a domain-specific language on top of lisp if you want something with a specific focus. But trying to create a full language that hides lisp syntax is pointless. It's been tried SO many times and failed exactly that many times. There is so much to lose. You lose homoiconic syntax which usually means you lose macros. You can hack macros by translating back to lisp, but why? I can find hundreds of "lisp without parens", just do a google search. For example, there is Wisp. You can generate lisp "without parens" by a simple hack. Push all of the parens to the right side. (+ (* 2 3) 6) becomes (ignore the underscores. I don't know how to generate verbatim text in HN) _____________________________________( + _____________________________________( * 2 3 _____________________________________) 6 _____________________________________) Now if you "adjust your margins", all of the parens disappear off the page. Lisp is a language that everyone says "I don't get it and I don't like it" until.... they get it and then they LOVE it. That's when they learn the limitations of all the other languages.
- brudgers 7y ago
  The package's author is near the center of the Racket community. https://www.asumu.xyz/ https://www.asumu.xyz/ It lives alongside the Racket DSL for Algol 60 https://docs.racket-lang.org/algol60/ https://docs.racket-lang.org/algol60/, Datalog https://docs.racket-lang.org/datalog/index.html https://docs.racket-lang.org/datalog/index.html, and ProfessorJ https://docs.racket-lang.org/profj/index.html https://docs.racket-lang.org/profj/index.html in the Racket ecosystem.