Re: [Haskell-cafe] containers license issue
Mike Meyer m...@mired.org writes: Niklas Larsson metanik...@gmail.com wrote: 2012/12/15 Mike Meyer m...@mired.org: Only if Tanenbaum documented the internal behavior of Linux before it was written. Tannenbaum wrote Minix, the operating system that Linus used (and hacked on) before he did Linux. Minix contained lots of features that was reimplemented in Linux. Ah, Minix isn't documentation. And it has a radically different architecture The point is that Linux read the source code to Minix before implementing similar things - quite likely using the same algorithms, for instance. So if containers is a translation of FXT, then surely Linux is a translation of Minix. That makes a successful lawsuit unlikely The point of the point is that neither of these are translations of literary works, there is no precedence for considering them as such, and that reading somebody's work (whether literary or source code) before writing one's own does not imply that the 'somebody' will hold any rights to the subsequent work. Translations in software is what compilers do, not reimplementing specific algorithms. it'll never go to court, so there isn't an infringement. Wot? -k ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] containers license issue
Mike Meyer m...@mired.org writes: As it's commonly understood, reverse engineering doesn't involve looking at the code. I guess I should make it clear that I don't use it in the strict sense - I would call that clean-room reverse engineering. (I'm not sure which is the most commonly understood meaning, I think Wikipedia supports both interpretations) That's why it's called reverse engineering instead of copying. I mean the process of reimplementing specific functionality from another system, with or without knowledge of implementation details. I would use copying to mean verbatim cut-and-pasting, which is something else. In particular when copyright is concerned, I believe that verbatim copying in many cases will require a license to the original work, but merly examining the original work to make use of algorithms, tricks, and structures from it will not. I suspect there's nothing magical about source code, if I extract a component (algorithm or data structure, say) from a program and use it in my own program, I'm not convinced it matters if I extract it from object code, source code, or documentation - they're all copyrighted works, and could be interpreted as translations of the same work. I could be wrong about that, though. -k ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] containers license issue
Ketil Malde ke...@malde.org wrote: The point of the point is that neither of these are translations of literary works, there is no precedence for considering them as such, and that reading somebody's work (whether literary or source code) before writing one's own does not imply that the 'somebody' will hold any rights to the subsequent work. So IANAL, but I do have an amateur interest in copyright law. The debate over the word translation is completely irrelevant. The important point is whether it is a derived work. That phrase certainly includes more than mere translation. For example, it includes writing fiction that's set in the same fantasy universe or HHS the same characters as another author's works. It also includes making videos with someone else's music playing in the background. If you create a derived work, then the author of the original definitely has rights to it, regardless of whether it is a mere translation. That's also why the word derived in a comment was particularly Dacey to the legal staff and probably caused them to overreact in this case. The defense in the case of software is to say that the part that was copied was not a work of authorship in the sense that, say, a fiction character is. This is generally not a hard case to win, since courts see computer software as dominated by its practical function. But if you copied something that was clearly a matter of expression and not related to the function of the software, you could very well be creating a derived work over which the original author could assert control. That said, I agree that in this particular case it's very unlikely that the original author could have won an infringement case. I just balked a little at the statements about translation, which was really just an example. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
[Haskell-cafe] Cabal: Retrieve directory in which binaries get installed
Hello cafe, I'd like to retrieve the path in which binaries get installed when I `cabal install` a package. Normally, this would be ~/.cabal/bin on my Linux, but since I share my home folder on two architectures, in ~/.cabal/config I set the user-install prefix to ~/.cabal/$arch, so binaries get installed to ~/.cabal/$arch/bin. I have the feeling that it should be rather simple to retrieve the FilePath from the Cabal API, does anyone know how to do this? Thanks in advance, Björn ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Cabal: Retrieve directory in which binaries get installed
There's getBinDir in the generated module Paths_pkgname (dist/build/autogen/Paths_pkgname.hs). This module is generated by cabal configure and can be imported in the modules of your package. Roman * Björn Peemöller b...@informatik.uni-kiel.de [2012-12-17 11:05:52+0100] Hello cafe, I'd like to retrieve the path in which binaries get installed when I `cabal install` a package. Normally, this would be ~/.cabal/bin on my Linux, but since I share my home folder on two architectures, in ~/.cabal/config I set the user-install prefix to ~/.cabal/$arch, so binaries get installed to ~/.cabal/$arch/bin. I have the feeling that it should be rather simple to retrieve the FilePath from the Cabal API, does anyone know how to do this? Thanks in advance, Björn ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
[Haskell-cafe] Navigating 'Strategic' programming babel
Clearly Haskell has great possibilities in the field of language-processing. And the nuisances associated with little actual computation buried under much data-structure navigation are well addressed by 'strategic-programming' systems. But now comes the rub -- there seem to be a lot of very similar systems. Any guidance on which/what/how to choose? My own current sketchy-patchy knowledge is as below. I would appreciate links/pointers to more substansive literature. First there was Meertens and his folks working on generic haskell Did that later become template haskell? That branched out into strafunski, stratego/xt. SYB is ___ not sure here: some literature suggests that its identical to strafunski. Some suggests that it is strafunski done more within the haskell language rather than in libraries. Then there's uniplate. How does it compare to SYB? Or is that a confused comparison? ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Navigating 'Strategic' programming babel
Hi Ravi, You might want to browse through Comparing Libraries for Generic Programming in Haskell: http://www.cs.uu.nl/research/techreps/repo/CS-2008/2008-010.pdf SYB and Uniplate are two widely used and well-maintained systems for strategic traversals over arbitrary datatypes. There are other options, too, but it depends on exactly what you want to do. Cheers, Pedro On Mon, Dec 17, 2012 at 11:30 AM, Ravi Sahni ganeshsahn...@gmail.comwrote: Clearly Haskell has great possibilities in the field of language-processing. And the nuisances associated with little actual computation buried under much data-structure navigation are well addressed by 'strategic-programming' systems. But now comes the rub -- there seem to be a lot of very similar systems. Any guidance on which/what/how to choose? My own current sketchy-patchy knowledge is as below. I would appreciate links/pointers to more substansive literature. First there was Meertens and his folks working on generic haskell Did that later become template haskell? That branched out into strafunski, stratego/xt. SYB is ___ not sure here: some literature suggests that its identical to strafunski. Some suggests that it is strafunski done more within the haskell language rather than in libraries. Then there's uniplate. How does it compare to SYB? Or is that a confused comparison? ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Navigating 'Strategic' programming babel
On Mon, Dec 17, 2012 at 4:06 PM, José Pedro Magalhães j...@cs.uu.nl wrote: Hi Ravi, You might want to browse through Comparing Libraries for Generic Programming in Haskell: http://www.cs.uu.nl/research/techreps/repo/CS-2008/2008-010.pdf SYB and Uniplate are two widely used and well-maintained systems for strategic traversals over arbitrary datatypes. There are other options, too, but it depends on exactly what you want to do. My interest is language processing for which stratego, strafunski, UUAG etc are specially tailor-made. Yes Ive read the 'comparing libraries' paper in which it says that strafunski and SYB approaches are so similar that strafunski is subsumed under SYB for the purposes of that paper. However it seems that the strafunski-ecosystem being integrated with aterm etc, likewise stratego, are more suited to full-scale language processing. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] [Haskell] ANN: Haskell bindings for the igraph C library
Am 16.12.2012 20:24, schrieb Jason Dagit: How does this compare with fgl? http://hackage.haskell.org/package/fgl FGL is a pure Haskell library while our haskell-igraph package uses the foreign function interface to run all graph-related calculations in C the C library igraph (I haven't implemented any graph algorithms). The runtime performance with our igraph library should be the same as if you'd be using the native C library (if you ignore the small Haskell-C-Haskell overhead). It is also seems to be more of a higher level library. As user you don't have to worry about node-IDs/labels or whether your graph is static or not (in the FGL context). Using features like GADTs and type families/associated types it is possible to keep track of informations like whether or not your graph is directed/weighted or not, while in FGL all graphs are by default directed and unweighted. Consider for example edges :: Graph d a - [Edge d a] -- directed, unweighted graph g :: Graph D a -- undirected, weighted graph w :: Graph (Weighted U) a which leads to edges g :: [Edge D a] edges w :: [Edge (Weighted U) a] or even functions like toUndirected :: (IsDirected d, E (ToUndirected d) a) = Graph d a - Graph (ToUndirected d) a toDirected :: (IsUndirected u, E (ToDirected u) a) = Graph u a - Graph (ToDirected u) a which evaluate to toUndirected g :: Graph U a toDirected w :: Graph (Weighted D) a This is even revertable, and `toDirected . toUndirected == id` while the FGL function `undir` simply adds all missing edges and loses track of what the original/directed graph looked like. Maybe George has more details on why he wanted to use igraph instead of FGL. - Nils ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
[Haskell-cafe] HTF Quickcheck
Getting test failures, I believe, due to using conditional properties within quickckeck etc [TEST] RecFunSpec:map (test/RecFunSpec.hs:48) Gave up! Passed only 20 tests. *** Failed! (2ms) Is there a way for HTF/Quickcheck to specify these are not failed i.e. only fail if there is an actual error ? Thanks ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] HTF Quickcheck
Unfortunately, I don't know the answer to your question. However, if you don't find a solution, I suggest using SmallCheck instead of QuickCheck — it works better when you have many unsuitable cases. https://github.com/feuerbach/smallcheck/wiki/Comparison-with-QuickCheck As far as I know, SmallCheck is not supported by HTF, but it is supported by test-framework. Roman * gra...@fatlazycat.com gra...@fatlazycat.com [2012-12-17 15:50:32+] Getting test failures, I believe, due to using conditional properties within quickckeck etc [TEST] RecFunSpec:map (test/RecFunSpec.hs:48) Gave up! Passed only 20 tests. *** Failed! (2ms) Is there a way for HTF/Quickcheck to specify these are not failed i.e. only fail if there is an actual error ? Thanks ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] HTF Quickcheck
On Mon, Dec 17, 2012 at 03:50:32PM +, gra...@fatlazycat.com wrote: Getting test failures, I believe, due to using conditional properties within quickckeck etc [TEST] RecFunSpec:map (test/RecFunSpec.hs:48) Gave up! Passed only 20 tests. *** Failed! (2ms) Is there a way for HTF/Quickcheck to specify these are not failed i.e. only fail if there is an actual error ? I have no idea about HTF, but QuickCheck thinks those are failures, too. If I run into something like this, it indicates a bug in my arbitrary instance. Here I would just adapt it so that it generates valid test data (possibly with the help of a newtype wrap). Cheers, Simon ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] containers license issue
On 12/17/12, Ketil Malde ke...@malde.org wrote: I would use copying to mean verbatim cut-and-pasting, which is something else. I feel I should point out that, while that's currently a common definition of copying, it's not the legal definition. Copyright law predates the ability to mechanically create copies of a work. The closest you could come was the printing press, which started with what was at the time a handmade mirror image of the work you were going to print copies of. Back then, the only way to create a copy of a printed page (or anything else that wasn't created with the express purpose of being duplicated) was to read the original, and transcribe it into a copy. There was no cut-n-paste or copy command - just reading and creating another copy by hand. Changes made along the way that depended on the original - adding illustrations or illuminations, setting it to music, re-arranging it into a script for a play, or translating it to another language - were all considered derived works, and hence infringing. IANAL, but I've been studying copyright law since before the US signed the Berne treaty. Generally, copyright law in the US has been getting stricter, not more permissive - largely because the only people who cared were the large copyright holders. In general, new media brought under the purview of copyright is treated as strictly as the law can be interpreted. In the US, most radio stations pay a license fee in order to play music. Bars and restaurants (and even the Girl Scouts!) are required to pay a license fee for public performances if they play live or recorded music. If you buy media capable of recording music or video, part of the cost is used to compensate the copyright holders of the presumably infringing copies you're going to make onto them. If I took a picture of my living room, I'd technically be in violation of a slew of copyrights (images of the statuary on the mantelpiece, paintings hanging on the wall, etc.) but chances are that nobody would care unless I included images from a FOX television show. Personally, I think this is silly, but it's the law. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] containers license issue
Ketil Malde ke...@malde.org wrote: Mike Meyer m...@mired.org writes: Niklas Larsson metanik...@gmail.com wrote: 2012/12/15 Mike Meyer m...@mired.org: Only if Tanenbaum documented the internal behavior of Linux before it was written. Tannenbaum wrote Minix, the operating system that Linus used (and hacked on) before he did Linux. Minix contained lots of features that was reimplemented in Linux. Ah, Minix isn't documentation. And it has a radically different architecture The point is that Linux read the source code to Minix before implementing similar things - quite likely using the same algorithms, for instance. So if containers is a translation of FXT, then surely Linux is a translation of Minix. Since they are both C, the concept of translation doesn't enter into that case at all. I never said that containers was a translation of FXT. I said that translations are considered to be derived work requiring permission to copy. This is deeply embedded in copyright law, and the GPL *depends* on it working for software. Otherwise, a distribution of GPL' d software translated to another language (say as asm from the compiler, or a linkable object, or even an executable) wouldn't be a derivative work and could be distributed without having to obey the license. The point is that containers being in haskell isn't a defense against copyright violation. It doesn't mean that it *is* a translation, merely that it might be. That the author of containers said it was derived from FXT opens the possibility that his version is actually GPL'ed, so using it opens up the possibility of a lawsuit. It may not stand up in court, but the lawyer objecting is trying to avoid just that possibility. That makes a successful lawsuit unlikely The point of the point is that neither of these are translations of literary works, there is no precedence for considering them as such, Actually, there *is* a precedent for considering them such. The clean-room implementations of the IBM PC bios were done because a judge ruled that translating from asm to binary was an infringing copy. If you have a precedent that translating from one programming language to another is *not* the creation of a derived work, I'd be interested in hearing about it. Without such a precedent, then a case can be made that such a translation is a derived work (and you're the only person I've ever heard claim otherwise) which opens up the possibility of a lawsuit, which is what the problem is. and that reading somebody's work (whether literary or source code) before writing one's own does not imply that the 'somebody' will hold any rights to the subsequent work. This is correct. Reading a copyrighted work does not prove that some subsequent creation is a copy of that work. It does, however, make it *possible* that such a work is a copy and needs the appropriate permissions. Translations in software is what compilers do, not reimplementing specific algorithms. Translation in a copyright case is a term of law, and has *nothing* to do with the behavior of compilers (other than the precedent that what a compiler does is considered a translation for copyright purposes). Reimplementing an algorithm may or may not be a copyright violation. Depends on whether or not the reimplementation involved creation of a work derived (in the legal sense) from the original. Access to the original is required for that to happen. The two being in the same language is *not* required for that to happen. Access to the original and working in the same language is *not* sufficient for that to happen. it'll never go to court, so there isn't an infringement. Tannenbaum isn't going to sue Linus. So either he doesn't believe there is an infringement (I suspect this is likely, as he has had access the source and would probably have at least said something if he thought there was an infringement) or doesn't care enough to go to court. Until a judge tells Linus he's infringing, he isn't. -- Sent from my Android tablet with K-9 Mail. Please excuse my swyping. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] containers license issue
Ketil Malde ke...@malde.org wrote: In particular when copyright is concerned, I believe that verbatim copying in many cases will require a license to the original work, but merly examining the original work to make use of algorithms, tricks, and structures from it will not. If you don't actually copy any of the text in the latter case, that would be correct. But there's an incredible amount of grey between those two extremes of black and white, and it's possible that you've unintentionally recreated significant bits of the original. The Oracle/Google lawsuit was all about those shades of grey - some of the API's in Dalvik were implemented by people who had read the Java sources. Oracle claimed as much as possible was derivative, Google that none of it was. The judge ruled that some uses were infringing and some uses were not. This was a technically literate judge - he ruled that one of the cases was non-infringing because he could trivially implement the function in Java himself. The lawyer who pointed out the possible infringement here isn't really worried about losing such a lawsuit - there are lots of ways to deal with that short of actually releasing any sources they consider proprietary. They want to avoid the lawsuit *at all*, as that will almost certainly be more expensive than losing it. At least, that's what I hear from clients who ask me not to include GPL'ed software. mike -- Sent from my Android tablet with K-9 Mail. Please excuse my swyping. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Navigating 'Strategic' programming babel
Strafunski is now rather out of date - it was developed before Cabal and used a custom install depending whether or not you wanted to use the DriFt preprocessor. Andy Gill has a modern re-implementation of Strafuski on Hackage called KURE. Aside from SYB, Neil Mitchell's Uniplate is popular and generally well documented. On 17 December 2012 11:12, Ravi Sahni ganeshsahn...@gmail.com wrote: However it seems that the strafunski-ecosystem being integrated with aterm etc, likewise stratego, are more suited to full-scale language processing. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] containers license issue
So I heard back from softwarefreedom.org, and they're looking for a representative from haskell.org to talk to them, as they want to avoid conflict-of-interests with other clients. Does anyone with any official status want to talk to real lawyers about this issue, then let the list know of anything interesting that was said? Let me know. - Clark On Mon, Dec 17, 2012 at 10:45 AM, Mike Meyer m...@mired.org wrote: Ketil Malde ke...@malde.org wrote: In particular when copyright is concerned, I believe that verbatim copying in many cases will require a license to the original work, but merly examining the original work to make use of algorithms, tricks, and structures from it will not. If you don't actually copy any of the text in the latter case, that would be correct. But there's an incredible amount of grey between those two extremes of black and white, and it's possible that you've unintentionally recreated significant bits of the original. The Oracle/Google lawsuit was all about those shades of grey - some of the API's in Dalvik were implemented by people who had read the Java sources. Oracle claimed as much as possible was derivative, Google that none of it was. The judge ruled that some uses were infringing and some uses were not. This was a technically literate judge - he ruled that one of the cases was non-infringing because he could trivially implement the function in Java himself. The lawyer who pointed out the possible infringement here isn't really worried about losing such a lawsuit - there are lots of ways to deal with that short of actually releasing any sources they consider proprietary. They want to avoid the lawsuit *at all*, as that will almost certainly be more expensive than losing it. At least, that's what I hear from clients who ask me not to include GPL'ed software. mike -- Sent from my Android tablet with K-9 Mail. Please excuse my swyping. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
[Haskell-cafe] haskell-like scripting language
I gave Shelly a try. Pretty cool - using it for some of the scripts on my system. Has me wondering though: is anyone working on creating a actual Haskell-like scripting language and engine? Shelly is cool, as I said, but I imagine it would be more valuable to have another language that is actually separate from Haskell, with an interpreter that is more lightweight and changes much less frequently (than GHC). Something that could be nearly as portable as Bash or Perl. -- frigidcode.com signature.asc Description: OpenPGP digital signature ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Navigating 'Strategic' programming babel
I won't compare and contrast all these, but I want to point out that there is a nicer version of uniplate in the lens package. On Dec 17, 2012 5:31 AM, Ravi Sahni ganeshsahn...@gmail.com wrote: Clearly Haskell has great possibilities in the field of language-processing. And the nuisances associated with little actual computation buried under much data-structure navigation are well addressed by 'strategic-programming' systems. But now comes the rub -- there seem to be a lot of very similar systems. Any guidance on which/what/how to choose? My own current sketchy-patchy knowledge is as below. I would appreciate links/pointers to more substansive literature. First there was Meertens and his folks working on generic haskell Did that later become template haskell? That branched out into strafunski, stratego/xt. SYB is ___ not sure here: some literature suggests that its identical to strafunski. Some suggests that it is strafunski done more within the haskell language rather than in libraries. Then there's uniplate. How does it compare to SYB? Or is that a confused comparison? ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] haskell-like scripting language
Shelly is cool, as I said, but I imagine it would be more valuable to have another language that is actually separate from Haskell, with an interpreter that is more lightweight and changes much less frequently (than GHC). Something that could be nearly as portable as Bash or Perl. Hugs? It has doesn't change frequently pretty much nailed. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] HTF Quickcheck
On Mon, Dec 17, 2012 at 06:04:15PM +0200, Roman Cheplyaka wrote: Unfortunately, I don't know the answer to your question. However, if you don't find a solution, I suggest using SmallCheck instead of QuickCheck — it works better when you have many unsuitable cases. https://github.com/feuerbach/smallcheck/wiki/Comparison-with-QuickCheck As far as I know, SmallCheck is not supported by HTF, but it is supported by test-framework. test-framework also sometimes thinks a Gave up! is not a fail, which I think is a bug. But I gave up on trying to get that fixed [1], e.g. this (admittedly contrived) property is still a pass with test-framework: defaultMain [testProperty foo $ \x - x == 23 == True] Cheers, Simon [1] https://github.com/batterseapower/test-framework/issues/16 ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] containers license issue
On 12/17/2012 01:47 PM, Clark Gaebel wrote: Does anyone ... want to talk to real lawyers ... No. This is absurd. If anyone cares, email the original author and explain the situation. Ask if he's cool with the reimplemented version. Chances are he'll say yeah, and we've just solved the problem like adults. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
[Haskell-cafe] category design approach for inconvenient concepts
Recently I read this article I happened across, about the categorical design pattern: http://www.haskellforall.com/2012/08/the-category-design-pattern.html Barely understood it, of course, but it was a rather intriguing concept. So now I'm looking at all my programming problems trying to make types that can be composed and that seem to fit the idea of a category. However, what I'm wondering about is ideas that can be composed but that don't seem to fit the idea of category, because they don't obey the associativity law. To give a specific example (pseudo code like, without any idea here of implementation or proper syntax): Say you created a type called Component (C for short), the idea being to compose Components out of other Components. Every C has zero or more connectors on it. Two Cs can be connected to form a new C using some kind of composition operator (say, .), provided there are enough connectors (one on each). Presumably you would need a Fail constructor of some kind to represent the situation when there is not enough connectors. Say you had a C (coupler) with two connectors, a C (thing) with one connector, and a C (gadget) with one connector. So you could have... (coupler . thing) . gadget Because the coupler and the thing would combine to create a component with one spare connector. This would then combine with the gadget to make the final component. However, if you did... coupler . (thing . gadget) Then thing and gadget combine to make a component with no spare connectors. And so the new component and the coupler then fail to combine. Associativity law broken. So, can I adjust my idea to fit the category concept? Or is it just not applicable here? Or am I just misunderstanding the whole concept? -- frigidcode.com signature.asc Description: OpenPGP digital signature ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] category design approach for inconvenient concepts
On 18/12/2012, at 3:45 PM, Christopher Howard wrote: Recently I read this article I happened across, about the categorical design pattern: http://www.haskellforall.com/2012/08/the-category-design-pattern.html It's basically the very old idea that an Abstract Data Type should be a nice algebra: things that look as though they ought to fit together should just work, and rearrangements of things ought not to change the semantics in surprising ways (i.e., Don't Be SQL). However, what I'm wondering about is ideas that can be composed but that don't seem to fit the idea of category, because they don't obey the associativity law. Say you created a type called Component (C for short), the idea being to compose Components out of other Components. Every C has zero or more connectors on it. Two Cs can be connected to form a new C using some kind of composition operator (say, .), provided there are enough connectors (one on each). Categories contain two things: objects and arrows that connect objects. Some important things about arrows: - for any object x there must be an identity idx : x - x - any two compatible arrows must compose in one and only one way: f : x - y g : y - z = g . f : x - z - composition must be associative (f . g) . h = f . (g . h) when the arrows fit together. Of course for any category there is a dual category, so what I'm about to say doesn't really make sense, but there's sense in it somewhere: the things you are trying to hook together with your . operator seem to me more like objects than arrows, and it does not seem as if they hook together in one and only one way, so it's not so much _associativity_ being broken, as the idea of . being a composition in the first place. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] category design approach for inconvenient concepts
Christopher Howard christopher.how...@frigidcode.com wrote: Say you created a type called Component (C for short), the idea being to compose Components out of other Components. Every C has zero or more connectors on it. Two Cs can be connected to form a new C using some kind of composition operator (say, .), provided there are enough connectors (one on each). Presumably you would need a Fail constructor of some kind to represent the situation when there is not enough connectors. Say you had a C (coupler) with two connectors, a C (thing) with one connector, and a C (gadget) with one connector. So you could have... (coupler . thing) . gadget Because the coupler and the thing would combine to create a component with one spare connector. This would then combine with the gadget to make the final component. However, if you did... coupler . (thing . gadget) Then thing and gadget combine to make a component with no spare connectors. And so the new component and the coupler then fail to combine. Associativity law broken. So, can I adjust my idea to fit the category concept? Or is it just not applicable here? Or am I just misunderstanding the whole concept? You are not misunderstanding the concept. You just need to learn how to apply it in this situation. Pluggable components with multiple slots are still modelled by regular categories. Let's say that Comp is such a component category (I'm using Haskell syntax): Comp :: * - * - * To model a component that takes two inputs you just write a component that takes one (!) tuple of inputs: myComp :: Comp (X, Y) Z For partial application of your component you need a bit more than the basic category pattern. Most categories form a family of applicative functors: instance Applicative (Comp a) Then partial application is really just this: partialMyComp :: X - Comp Y Z partialMyComp x = myComp . fmap (\y - (x, y)) id Every category that forms such a family of applicative functors is a profunctor (see the 'profunctors' package by Edward Kmett): instance Profunctor Comp That makes expressing partialMyComp slightly more pleasing: partialMyComp x = lmap (\y - (x, y)) myComp Finally to save even more keystrokes enable the TupleSections extension: partialMyComp x = lmap (x,) myComp You can get similar description through the Arrow interface: instance Arrow Comp partialMyComp x = myComp . arr (x,) However, often the applicative interface is much more pleasing. Now to the theory: What does an applicative functor give you? It basically extends a categorical concept by the ability to combine multiple morphisms to a single one. Given two morphisms, c1 :: Comp A B c2 :: Comp A C and a function f :: B - C - D an applicative functor gives you a well-defined way to combine c1 and c2 into another morphism of type c3 :: Comp A D An applicative functor provides about the same theoretical soundness as a category in that such a combination of multiple Comp morphisms is itself always a Comp morphism, and that the combination operator itself cannot introduce unwanted effects because of the Applicative laws: pure f * x = fmap f x f * pure x = fmap ($ x) f pure f * pure x = pure (f x) In other words, all effects are introduced by primitive components. I call them atoms. Greets, Ertugrul -- Not to be or to be and (not to be or to be and (not to be or to be and (not to be or to be and ... that is the list monad. signature.asc Description: PGP signature ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] HTF Quickcheck
So what use are conditional properties meant to be used for ?? If you just want 2 integers that are not equal, it would seem a lot simpler to do this as a conditional rather than constructing some pair in an arbitrary instance ?!? Thanks On Mon, Dec 17, 2012, at 11:20 PM, Simon Hengel wrote: On Mon, Dec 17, 2012 at 06:04:15PM +0200, Roman Cheplyaka wrote: Unfortunately, I don't know the answer to your question. However, if you don't find a solution, I suggest using SmallCheck instead of QuickCheck — it works better when you have many unsuitable cases. https://github.com/feuerbach/smallcheck/wiki/Comparison-with-QuickCheck As far as I know, SmallCheck is not supported by HTF, but it is supported by test-framework. test-framework also sometimes thinks a Gave up! is not a fail, which I think is a bug. But I gave up on trying to get that fixed [1], e.g. this (admittedly contrived) property is still a pass with test-framework: defaultMain [testProperty foo $ \x - x == 23 == True] Cheers, Simon [1] https://github.com/batterseapower/test-framework/issues/16 ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
