4 ms·

Imperative + Functional == :-)

33 points by mcgwiz 14y ago

mardiros 14y ago
Server: Microsoft-IIS/6.0 == Service Unavailable
- weakwire 14y ago
  Functional not so functional?
mercurial 14y ago
This gentleman seems to be talking about a dialect of C# with more functional idioms and parallel-programming friendly features. However, a link describing the actual language itself would have been much more interesting. Briefly looking at the PDF, it doesn't seem to offer much more than rust (apart from .NET interop, which is certainly nothing to sneeze at, if it's your cup of tea), though the mention of a team at Microsoft using it as its primary language and having delivered "millions of lines of code" with it hints at a greater maturity.
- codewright 14y ago
  You appear to be an emissary from the Rust community. Let me make a request and hope you can forward it on to your dark masters in the land of Mozilla. Please stop using every single topic on HN to mention "Rust does it better". Bonus points: show off things actually made with Rust. I've let it slide up to this point because any attempt to "get serious" was always countered with the excuse, "we're not reeeeaaaaady". If you're going to keep proselytizing, you're going to be held to a higher standard. Also mentioning that Rust does it better without mentioning specifics is pretty much the opposite of useful. All you're doing is namedropping, not sharing knowledge or meaningful information. Summary of my preferred reactions to this comment: 1. Add more content to your comment so it's substantive rather than name-dropping 2. Ask the Rust community to stop trying to outdo the Haskell community at propagandizing (hi dons!) 3. Start building things in Rust and show HN. A brief blog post describing pros/cons/future improvements would be boss.
  - zoowar 14y ago
    I compiled rust tonight. It took longer to compile than it took me to read all the documentation. Strike one.
    - codewright 14y ago
      Congrats, you encountered my first complaint with Rust. Let me know how much fun the out-of-date documentation referencing syntax that no longer exists is. If you ask about it or seek a redress you'll be told, "we're not reeeaaaaaady". It's a state of affairs they've come to be quite comfortable with. You're supposed to use Rust unless you're a Mozilla employee or intern, afaict.
      - buster 14y ago
        Maybe you just shouldn't use it now when it's not ready. There is no reason to complain about missing documentation or other stuff if they didn't get to do it until now. That what "not ready" means. Keep calm and wait ;-)
      - graue 14y ago
        What do you expect? They aren't ready. I'm following the development of Rust with much interest. The design seems well-thought out, and occasional comments here by 'pcwalton, usually along the lines of “here's how we're trying to solve this problem in Rust”, are always worth a read. I wouldn't start a project in it, however, for at least another year unless my main goal was to experiment with the language. When I did play with it a little, I indeed ran into problems with out-of-date documentation, but someone in #rust IRC (I think it was 'brson) very helpfully answered my questions and pointed me to some newer docs. Perhaps your expectations are unrealistic. If you want a language you can use today for a real project, try Clojure or Go. I don't think you can fault the Rust project for being where they are now and communicating honestly.
    - Ygg2 14y ago
      I too am not affiliated to Mozilla, unless the browser I'm using makes me an affiliated member. However judging from their discussion group, they know it's a problem and will probably try to reduce it to 15min(this is their goal, not their achieved min speed). Sadly, speed of compilation probably won't beat Go's which is 2min. I do understand your complaint, but dismissing an alpha language because it compilation is too long, or documentation isn't up to par is a bit of an overstretch. They haven't been optimizing for it and the syntax is shifting, so making documentation at this point is like building a house on a sand pit.
abolibibelot 14y ago
The paper is at http://research.microsoft.com/pubs/170528/msr-tr-2012-79.pdf http://research.microsoft.com/pubs/170528/msr-tr-2012-79.pdf (the redirection fails, but the url is in the query string...)
gtani 14y ago
another thread, blog ref'g this paper https://news.ycombinator.com/item?id=4871661 https://news.ycombinator.com/item?id=4871661
pcote 14y ago
Interesting goals but.... >> 1. Create a single language that incorporates the best of functional and imperative programming. Both offer distinct advantages, and we envisioned marrying the two. Imperative/Functional hybrid are nothing new. C#, Ruby, Python, Javascript, Scala, and Groovy all fit that bill. >> 2. Codify and statically enforce common shared-memory state patterns, such as immutability and isolation, with minimal runtime overhead (i.e., virtually none). "Statically enforcing variable immutability...." Isn't that just a fancy word for "constant"? >> 4. Do all of this while still offering industry-leading code quality and performance, rivaling that of systems-level C programs. Yet still with the safety implied by the above mentioned goals. He wants to allow for "pockets of imperative mutability tied together with functional tissue". That's not a bad thing but it's not exactly a "safe" feature if put in the wrong hands. Lots of people have written function objects with bad imperative code and side effects. Lots of people have fed them to their language's equivalent of the map function. That can cause bugs.
- vorg 14y ago
  > Imperative/Functional hybrid are nothing new. C#, Ruby, Python, Javascript, Scala, and Groovy all fit that bill. Some of those languages are more functional than others. The Scala creator claims it's 50% imperative 50% functional. Others of those, e.g. C# and Groovy, are closer to 10 or 20% functional.
  - pcote 14y ago
    It a certain degree, beauty is a better indicator than percentages when it comes to how functional a programming language is. Scheme scores high in that regard. Python does okay if you stay moderate in it's usage. Any amount of functional programming in Java tends to be downright ugly and unlikely to pass most code reviews.
- joeduffy 14y ago
  The strength of this system is that map is type checked to be provably side-effect free. The mapping function is free to internally allocate and mutate objects to get its job done, e.g. a RegexMatcher, but it may not mutate existing (non-isolated) state. This composes transitively, so even wrapping side-effects in abstractions and closures is insufficient to hide them.
mcgwiz 14y ago
This is a blog article from one of the researchers on a project to bring more immutability to C#, among other things. The paper the article references is well-summarized in a C# context here: http://www.infoq.com/news/2012/12/Immutable-CSharp http://www.infoq.com/news/2012/12/Immutable-CSharp

33 points by mcgwiz 14y ago

mardiros 14y ago

Server: Microsoft-IIS/6.0 == Service Unavailable

weakwire 14y ago
Functional not so functional?

mercurial 14y ago

This gentleman seems to be talking about a dialect of C# with more functional idioms and parallel-programming friendly features. However, a link describing the actual language itself would have been much more interesting. Briefly looking at the PDF, it doesn't seem to offer much more than rust (apart from .NET interop, which is certainly nothing to sneeze at, if it's your cup of tea), though the mention of a team at Microsoft using it as its primary language and having delivered "millions of lines of code" with it hints at a greater maturity.

codewright 14y ago
You appear to be an emissary from the Rust community. Let me make a request and hope you can forward it on to your dark masters in the land of Mozilla. Please stop using every single topic on HN to mention "Rust does it better". Bonus points: show off things actually made with Rust. I've let it slide up to this point because any attempt to "get serious" was always countered with the excuse, "we're not reeeeaaaaady". If you're going to keep proselytizing, you're going to be held to a higher standard. Also mentioning that Rust does it better without mentioning specifics is pretty much the opposite of useful. All you're doing is namedropping, not sharing knowledge or meaningful information. Summary of my preferred reactions to this comment: 1. Add more content to your comment so it's substantive rather than name-dropping 2. Ask the Rust community to stop trying to outdo the Haskell community at propagandizing (hi dons!) 3. Start building things in Rust and show HN. A brief blog post describing pros/cons/future improvements would be boss.
- zoowar 14y ago
  I compiled rust tonight. It took longer to compile than it took me to read all the documentation. Strike one.
  - codewright 14y ago
    Congrats, you encountered my first complaint with Rust. Let me know how much fun the out-of-date documentation referencing syntax that no longer exists is. If you ask about it or seek a redress you'll be told, "we're not reeeaaaaaady". It's a state of affairs they've come to be quite comfortable with. You're supposed to use Rust unless you're a Mozilla employee or intern, afaict.
    - buster 14y ago
      Maybe you just shouldn't use it now when it's not ready. There is no reason to complain about missing documentation or other stuff if they didn't get to do it until now. That what "not ready" means. Keep calm and wait ;-)
    - graue 14y ago
      What do you expect? They aren't ready. I'm following the development of Rust with much interest. The design seems well-thought out, and occasional comments here by 'pcwalton, usually along the lines of “here's how we're trying to solve this problem in Rust”, are always worth a read. I wouldn't start a project in it, however, for at least another year unless my main goal was to experiment with the language. When I did play with it a little, I indeed ran into problems with out-of-date documentation, but someone in #rust IRC (I think it was 'brson) very helpfully answered my questions and pointed me to some newer docs. Perhaps your expectations are unrealistic. If you want a language you can use today for a real project, try Clojure or Go. I don't think you can fault the Rust project for being where they are now and communicating honestly.
  - Ygg2 14y ago
    I too am not affiliated to Mozilla, unless the browser I'm using makes me an affiliated member. However judging from their discussion group, they know it's a problem and will probably try to reduce it to 15min(this is their goal, not their achieved min speed). Sadly, speed of compilation probably won't beat Go's which is 2min. I do understand your complaint, but dismissing an alpha language because it compilation is too long, or documentation isn't up to par is a bit of an overstretch. They haven't been optimizing for it and the syntax is shifting, so making documentation at this point is like building a house on a sand pit.

abolibibelot 14y ago

The paper is at http://research.microsoft.com/pubs/170528/msr-tr-2012-79.pdf http://research.microsoft.com/pubs/170528/msr-tr-2012-79.pdf (the redirection fails, but the url is in the query string...)

gtani 14y ago

another thread, blog ref'g this paper https://news.ycombinator.com/item?id=4871661 https://news.ycombinator.com/item?id=4871661

pcote 14y ago

Interesting goals but.... >> 1. Create a single language that incorporates the best of functional and imperative programming. Both offer distinct advantages, and we envisioned marrying the two. Imperative/Functional hybrid are nothing new. C#, Ruby, Python, Javascript, Scala, and Groovy all fit that bill. >> 2. Codify and statically enforce common shared-memory state patterns, such as immutability and isolation, with minimal runtime overhead (i.e., virtually none). "Statically enforcing variable immutability...." Isn't that just a fancy word for "constant"? >> 4. Do all of this while still offering industry-leading code quality and performance, rivaling that of systems-level C programs. Yet still with the safety implied by the above mentioned goals. He wants to allow for "pockets of imperative mutability tied together with functional tissue". That's not a bad thing but it's not exactly a "safe" feature if put in the wrong hands. Lots of people have written function objects with bad imperative code and side effects. Lots of people have fed them to their language's equivalent of the map function. That can cause bugs.

vorg 14y ago
> Imperative/Functional hybrid are nothing new. C#, Ruby, Python, Javascript, Scala, and Groovy all fit that bill. Some of those languages are more functional than others. The Scala creator claims it's 50% imperative 50% functional. Others of those, e.g. C# and Groovy, are closer to 10 or 20% functional.
- pcote 14y ago
  It a certain degree, beauty is a better indicator than percentages when it comes to how functional a programming language is. Scheme scores high in that regard. Python does okay if you stay moderate in it's usage. Any amount of functional programming in Java tends to be downright ugly and unlikely to pass most code reviews.
joeduffy 14y ago
The strength of this system is that map is type checked to be provably side-effect free. The mapping function is free to internally allocate and mutate objects to get its job done, e.g. a RegexMatcher, but it may not mutate existing (non-isolated) state. This composes transitively, so even wrapping side-effects in abstractions and closures is insufficient to hide them.

mcgwiz 14y ago

This is a blog article from one of the researchers on a project to bring more immutability to C#, among other things. The paper the article references is well-summarized in a C# context here: http://www.infoq.com/news/2012/12/Immutable-CSharp http://www.infoq.com/news/2012/12/Immutable-CSharp