... were important initial requirements. The functional aspects of Erlang were introduced ... 1988. Perhaps noteworthy is that "functional programming" was not really mentioned ..., or fail to consider real-world failure scenarios in heterogeneous environments.... on topic (concurrency in the functional world), we've found that ... to do with CSP than functional programming. FP contributes by removing ...

..., 2:18 pm, Arne Hanssen <arne.hansenNOS...@runbox.com> wrote: ... As an outsider I am wondering how concurrency is treated in the functional world. Is it a major research field, do people think about it when programming and do the languages offers concurrent features. Some research: the 2007 ICFP ...

...As an outsider I am wondering how concurrency is treated in the functional world. Is it a major research field...the ability to write in a purely functional style is hugely beneficial. And what ..."the-multi-core-trend" have on functional programming and programming languages?. Unlike many other languages, functional languages will survive the transition to multicore. --...

One popular claim found in litterateur about functional programming is that programs written in this class of ... theorem. As an outsider I am wondering how concurrency is treated in the functional world. Is it a major research field, do people think about it when programming and ...lets call it "the-multi-core-trend" have on functional programming and programming languages?. Cheers -arne

... optionally link to parent process - respect user-defined "modified timing" parameters - set priority level and some of the mandatory work: - register meta-data for debugging, such as initial function - initialize data structures to hold links, monitors, process dictionary So, it's all a tradeoff between debugging support, utility and speed. From a programmer's point of view, I ...

...: let! x = ... Presumably x returns a future that will be forced only when x is actually inspected? Exactly, yes. The code can still be purely functional. Any interaction between processes/threads/whatever is completely abstracted away. Right. Similar constructs can be found in e.g. GHC and Alice. Yes. But the reason these kinds of computations can ...

... the user-level process. This can be done on an ad hoc basis in the application but it is obviously simpler for the application programmer if it is handled by the language runtime. 3. have the compiler generate checks on function entry to see if too much real or virtual time has been used and if so, re-schedule. I believe at least one implementation of Erlang uses 3.

... so that it can't be said that the test is rigged in favour of message passing. But I'd look for one rigged in favor of shared memory, so it's better to look for real-world cases. By all means offer one rigged in favour of shared memory, I'd like to see it. applications you are intersted in are such that the amount of CPU you consume processing the data is of the ...

...This is all important, because at the level of hardware, even functional languages do a lot of mutation. Whenever you use a copying...with its performance model intact! 3. Message-passing with purely functional data and a shared heap. Message passing is just as... this as a fundamental problem. Even languages that are purely functional at the user level may face this problem if they need...

...--do you need a message pair for each operation? Optimally you structure your program so that the combination of several operations can be done by the remote process with a single call. For further flexibility, you can send to the remote process a function that operates on its data structure. In distributed systems this requires code marshalling, but that's not unheard of. Lauri