- Haskell-Cafe - Haskell.org

Are bottoms ever natural?
by (IIIT) Siddharth Bhat 20 Dec '17

20 Dec '17

I've been thinking about the issue of purity and speculation lately, and from what little I have read, it looks like the presence of bottom hiding inside a lazy value is a huge issue. How "natural" is it for bottoms to exist? If one were to change Haskell and declare that any haskell value can be speculated upon, what ramifications does this have? Is it totally broken? Is it "correct" but makes programming unpleasant? What's the catch? Thanks, Siddharth -- Sending this from my phone, please excuse any typos!

4 6

Re: [Haskell-cafe] Are bottoms ever natural?
by Alexey Muranov 20 Dec '17

20 Dec '17

Siddharth, are you aware that the question whether a given element belongs to the domain of a given computable function is algorithmically undecidable? Alexey. On Tue, 2017-12-19 at 16:51 +0000, Siddharth Bhat wrote: > > So, I have two opinions to share on this: > > Regarding the plane example, it is wrong to attempt to graph it on a > > plane, precisely because the domain is not the plane. > > I am confused by your assertion that it is impossible to avoid > > divergence in mathematics: we do not define division as a *partial* > > function. Rather we define it as a *total* function on a *restricted > > domain*. > > So, what one would need is the ability to create these restricted > > domains, which certain languages have as far as I am aware. > > At that point, now that we track our domains and codomains > correctly, > > all should work out? > > I would still like an answer to interesting transformations that are > > enabled by having purity and laziness and are not encumbered by the > > presence of bottoms. > > Cheers, > > Siddharth > >

1 0

Writing an interpreter for a language with mutability
by Jeremy Mikkola 20 Dec '17

20 Dec '17

I am quite certain I am not the first to try to do this, but my google-fu is failing me today. How does one go about interpreting a language with mutable objects in Haskell? The best approach I can think of is to represent the language's memory as a `Data.Map.Map RefID LanguageObject` where RefID is some type (probably Int) used as a reference. The LanguageObject structure might contain some values of type RefID to refer to other objects. Mutating an object involves simply replacing the value in the map at a given RefID. I don't like this approach for two reasons: 1. Map lookups aren't very efficient compared with actual references to the value. 2. I have to re-invent garbage collection, removing objects from the map when they no longer have incoming references. (Unlike simple interpreters for languages with immutable values, I can't rely on Haskell's garbage collector to do the work for me.) Is there a better approach? - Jeremy

6 6

Cybersecurity Posts (2) at the University of Sheffield
by Achim D. Brucker 20 Dec '17

20 Dec '17

(Apologies for duplicates - resending after I forgot to attach the body of the mail ...) Dear all, As part of an ambitious strategy for growth, the Department of Computer Science is seeking to appoint two Lecturers/Senior Lecturers/Readers with outstanding expertise in the field of cybersecurity. Computer and information security is a rapidly growing interest at Sheffield. In the Department of Computer Science we have recently established a Security of Advanced Systems group, led by Professor John Clark. There is related expertise in other departments such as Law, Mathematics and Statistics, and Politics. Furthermore, Sheffiield’s strength in engineering brings many opportunities for collaborative research in cybersecurity, in areas such as robotics and advanced manufacturing. To drive our research agenda in cybersecurity forward, we are now seeking to make two appointments at Lecturer, Senior Lecturer or Reader Level. Candidates will demonstrate an excellent research record in cybersecurity. Suitable areas of expertise include (but are not limited to): security policies, threat modelling, authentication, access control, malware and malware detection, network security, secure protocols, intrusion detection, and human factors and security. Candidates who work at the interface of artificial intelligence and cybersecurity are particularly encouraged to apply. In 2016 we were recognised for our work on diversity and equality with a Silver Athena Swan award. Interested candidates may find details of these (and other) posts at: https://www.sheffield.ac.uk/dcs/jobs/index Best, Achim -- Dr. Achim D. Brucker | Software Assurance & Security | University of Sheffield https://www.brucker.ch | https://logicalhacking.com/blog @adbrucker | @logicalhacking

1 0

Cybersecurity Posts (2) at the University of Sheffield
by Achim D. Brucker 20 Dec '17

20 Dec '17

(Apologies for duplicates) -- Dr. Achim D. Brucker | Software Assurance & Security | University of Sheffield https://www.brucker.ch | https://logicalhacking.com/blog @adbrucker | @logicalhacking

1 0

coercing D a to D b
by Evan Laforge 19 Dec '17

19 Dec '17

Given: data D a = A a | B Int Char dmapNoCoerce :: (a -> b) -> D a -> D b dmapNoCoerce f (A a) = A (f a) dmapNoCoerce _ (B i c) = B i c I have to reconstruct a B change it from D a to D b. But at the lower level, couldn't this be implemented as a type cast? What prevents such an optimization? I can write this as dmapCoerce :: (a -> b) -> D a -> D b dmapCoerce f (A a) = A (f a) dmapCoerce _ b@(B {}) = Unsafe.Coerce.unsafeCoerce b >From the core, it looks like dmapCoerce indeed has a cast with no allocation, while dmapNoCoerce allocates a new B. It seems to work, but is it safe? Is there a more principled way to do it? I can't convince Data.Coerce to cooperate, presumably because 'a' and 'b' are not coercible themselves, and it doesn't believe me if I try to tell it the type role is phantom.

3 4

error terminated list
by Evan Laforge 19 Dec '17

19 Dec '17

I've been using this lately to represent a process which can fail but still produce data: data UntilFail err a = a :+ UntilFail err a | Done | Fail err I know there's been discussion of this type before, but I don't remember what it was called and I can't find the right search terms to turn it up. Ring any bells? I have some awkward names for it, but maybe someone else thought of something more elegant. I have some ad-hoc functions for it: mapUntilFail :: (a -> Either err b) -> UntilFail err a -> UntilFail err b concatMapUntilFail :: (a -> UntilFail err b) -> UntilFail err a -> UntilFail err b -- | Like 'concatMapUntilFail', but consume a variable number of results. -- -- A more precise type would end with @Done [a]@. processUntilFail :: (a -> [a] -> (UntilFail err b, [a])) -> [a] -> UntilFail err b Maybe someone who has thought more deeply about it has come up with more elegant abstractions.

1 0

Haskell IDE Engine logo poll
by Damien Flament 19 Dec '17

19 Dec '17

Hi ! The final poll is available featuring all logos submitted until last Friday. Please vote for your favourite*s* here: http://freeonlinesurveys.com/p/Cw1FRJ2G?qid=1017541 Thanks. Damien.

1 0

The Definition of Haskell
by Todd Wilson 19 Dec '17

19 Dec '17

Why doesn't Haskell (or at least "core" Haskell) have a formal definition and safety result the way Standard ML does [1], [2]? Is it just that no one has gotten around to producing them, or because there are fundamental difficulties that stand in the way of completing such a project? Todd Wilson [1] Robin Milner, Mads Tofte, Robert Harper, and David MacQueen. The Definition of Standard ML (Revised). MIT Press, 1997. [2] Daniel K. Lee , Karl Crary , Robert Harper, "Mechanizing the Metatheory of Standard ML", http://www.cs.cmu.edu/~dklee/tslf/

8 9

Managing package release workflow
by Clinton Mead 18 Dec '17

18 Dec '17

Hi All I've already got a few packages on hackage <https://hackage.haskell.org/user/clinton>, but I'd like to upload a few more, but I'm finding the workflow management particularly excessive. Here's what I'd like to do roughly: 1. Choose a number of stack snapshots I'd like my package to successfully build and test against. 2. Automatically apply appropriate version bounds to dependencies in my cabal file based on those snapshots. 3. When these changes are pushed to Github, attempt to compile them using Travis CI, against all the stack snapshots chosen. 4. If we do a push to a particular branch or with particular tag, this signifies we're attempting a release. Attempt to compile using Travis CI, perform the tests. If the compile and tests are successful, push a separate commit, adjust the "source-repository" field in github to point to that release commit (not to master) and upload to hackage. I'm currently doing all of this manually but it's increasingly a bit of a pain. Is there tooling to help me automate this? Thanks, Clinton

2 1