3 ms·

She's pretty fast to dismiss Shen: > Shen is a splendid compromise, but I found myself wishing it came with Haskell's uncompromising purity and Standard ML's s

by nnq 14y ago

She's pretty fast to dismiss Shen: > Shen is a splendid compromise, but I found myself wishing it came with Haskell's uncompromising purity and Standard ML's straightforward type system. I guess Haskell/ML minded people have a reason to think this way, but I'd be more interested to know whether someone is doing anything interesting with Shen. Have you got any "production" code written in it and running to show off? EDIT: sorry! ...fixed the gender bug.

sanityinc 14y ago
> He's pretty fast to dismiss Shen: She's pretty fast to dismiss Shen.
pi18n 14y ago
I would dismiss Shen just for the weird license. What if I like it and want to use it in an embedded product or somewhere? I'd have to make sure any derivative work adheres to his spec even if it's only intended to run one library.
- hga 14y ago
  Bingo; I already have, after trying to convince the author to at minimum make the license short enough to be less ambiguous. He's absolutely sure it's not a problem no matter how many people tell him it is. This could also be a factor in the BODOL author's dismissal.
  - pi18n 14y ago
    Further down he writes the rationale is to prevent fragmentation of the community; from what I have seen it prevents developers who otherwise might from joining the community. It's not like e.g. Python or Haskell have a fragmentation problem despite permissible licenses. Scheme has fragmentation problems but that seems to be a symptom of simple implementation and lack of a common set of libraries. I think Shen is interesting, but if I used it I'd only be mining it for ideas to apply in other languages. It would make far more sense to me if he allowed derivatives iff they did not use the name.
    - hga 14y ago
      I don't disagree with the goal of preventing fragmentation, and I'd even be willing to play the flag game (i.e. have my version focus on SMP functionality in both senses of the latter, while observing the official spec with a different flag). It's the long winded, says the same things at least 5 (!) different times and ways, and therefore subject to many interpretations license, which is then subject to whims of the author or an external board (compare to Java, where in theory passing a suite of tests allows you to offer a version of Java(TM)). I'm not going to invest serious effort in something that might get killed simply for political reasons including different interpretations of the license; even if I trusted the good will of everyone involved, sooner or later the players will change.
hga 14y ago
Focus on "compromise", splendid or not. Shen is less functional than Clojure; the author has absolutely no interest in making SMP systems sing, last time I checked he's going down the actor message passing route. You could create a more functional version as long as it had a switch or default to the normal behavior and it was blessed by the powers that be (see comments on license), but you'd be "fighting the system" and all that.

by nnq 14y ago

sanityinc 14y ago

> He's pretty fast to dismiss Shen: She's pretty fast to dismiss Shen.

pi18n 14y ago

I would dismiss Shen just for the weird license. What if I like it and want to use it in an embedded product or somewhere? I'd have to make sure any derivative work adheres to his spec even if it's only intended to run one library.

hga 14y ago
Bingo; I already have, after trying to convince the author to at minimum make the license short enough to be less ambiguous. He's absolutely sure it's not a problem no matter how many people tell him it is. This could also be a factor in the BODOL author's dismissal.
- pi18n 14y ago
  Further down he writes the rationale is to prevent fragmentation of the community; from what I have seen it prevents developers who otherwise might from joining the community. It's not like e.g. Python or Haskell have a fragmentation problem despite permissible licenses. Scheme has fragmentation problems but that seems to be a symptom of simple implementation and lack of a common set of libraries. I think Shen is interesting, but if I used it I'd only be mining it for ideas to apply in other languages. It would make far more sense to me if he allowed derivatives iff they did not use the name.
  - hga 14y ago
    I don't disagree with the goal of preventing fragmentation, and I'd even be willing to play the flag game (i.e. have my version focus on SMP functionality in both senses of the latter, while observing the official spec with a different flag). It's the long winded, says the same things at least 5 (!) different times and ways, and therefore subject to many interpretations license, which is then subject to whims of the author or an external board (compare to Java, where in theory passing a suite of tests allows you to offer a version of Java(TM)). I'm not going to invest serious effort in something that might get killed simply for political reasons including different interpretations of the license; even if I trusted the good will of everyone involved, sooner or later the players will change.

hga 14y ago

Focus on "compromise", splendid or not. Shen is less functional than Clojure; the author has absolutely no interest in making SMP systems sing, last time I checked he's going down the actor message passing route. You could create a more functional version as long as it had a switch or default to the normal behavior and it was blessed by the powers that be (see comments on license), but you'd be "fighting the system" and all that.