Some feedback on
Rahul Muttineni @rahulmutt nov. 09 2016 03:40 Lead developper of ETA (the JVM Haskell compiler)
It’s a bit mind bending in that it’s like using a higher-level list monad, but it’s very, very cool. For beginning Haskellers, what would be really useful is a visualisation of what happens when you do various distributed/parallel stuff. It’s almost shocking how effortlessly you can run computations across threads/nodes.
The cool part is the composability in the distributed setting. You can make higher-order monadic functions that allow you to compose & reuse a long chain of distributed transactions via
teleport. Another benefit is that the transaction becomes first class and you can see exactly what’s going on in one place instead of distributing the logic across actors making the code equivalent to event callbacks, as you’ve stated.
What is Transient?
One of the dreams of software engineering is unrestricted composability.
This may be put in these terms:
ap2 two applications with arbitrary complexity, with all effects including multiple threads, asynchronous IO, indeterminism, events and perhaps, distributed computing.
Then the combinations:
- ap1 <|> ap2 -- Alternative expression - ap1 >>= \x -> ap2 -- monadic sequence - ap1 <> ap2 -- monoidal expression - (,) <$> ap1 <*> ap2 -- Applicative expression
are possible if the types match, and generate new applications that are composable as well.
Transient does exactly that.
<> express concurrency, the operator
<|> express parallelism and
>>= for sequencing of threads, distributed processes or web widgets. So even in the presence of these effects and others, everything is composable.
For this purpose transient is an extensible effects monad with all major effects and primitives for parallelism, events, asynchronous IO, early termination, non-determinism logging and distributed computing. Since it is possible to extend it with more effects without adding monad transformers, the composability is assured.
This program, will stream “hello world” from all the nodes connected if you enter “fire” in the console
main= keep $ initNode $ inputNodes <|> distribStream distribStream= do local $ option "fire" "fire" r <- clustered . local . choose $ repeat "hello world" localIO $ print r
Read the tutorial to know how to compile and invoke it.
This program will present a link in the browser and stream fibonnacci numbers to the browser when yo click it. (if you have Docker, you can run it straigh from the console; See this
main= keep . initNode $ webFib webFib= onBrowser $ do local . render $ wlink () (h1 "hello fibonacci numbers") r <- atRemote $ do r <- local . threads 1 . choose $ take 10 fibs localIO $ print r localIO $ threadDelay 1000000 return r local . render . rawHtml $ (h2 r) where fibs = 0 : 1 : zipWith (+) fibs (tail fibs) :: [Int]
This program combines both functionalities:
main= keep . initNode $ inputNodes <|> webFib <|> distribStream
The Wiki is more user oriented
My video sessions in livecoding.tv not intended as tutorials or presentations, but show some of the latest features running.
The articles are more technical:
- Philosophy, async, parallelism, thread control, events, Session state
- Backtracking and undoing IO transactions
- Non-deterministic list like processing, multithreading
- Distributed computing
- Publish-Subscribe variables
- Distributed streaming, map-reduce
These articles contain executable examples (not now, since the site no longer support the execution of haskell snippets).
Once composability in the large is possible, there are a infinite quantity of ideas that may be realized. There are short term and long term goals. An status of development is regularly published in .
Among the most crazy ones is the possibility of extending this framework to other languages and make them interoperable. treating whole packaged applications as components, and docking them as lego pieces in a new layer of the Operating system where the shell allows such kind of type safe docking. this composable docker allows all kinds of composability, while the current docker platform is just a form of degraded monoid that do not compute.
Wanna contribute? Make sure that you’ve read our contributor guidelines. We’d like to hear from you and your ideas, get in touch with other contributors through:
Once you learn something interesting, you can contribute to the wiki
You can also donate to the lead developer in order to make possible the dedication of more time to fullfil the potential advantages of true software composability across the whole stack.