From fredrik.nordvall-forsberg at strath.ac.uk Mon Jan 13 14:14:55 2025 From: fredrik.nordvall-forsberg at strath.ac.uk (Fredrik Nordvall Forsberg) Date: Mon, 13 Jan 2025 14:14:55 +0000 Subject: [msp-interest] New msp-interest mailing list address Message-ID: <1440e703-c8fd-4edd-aa72-bf6175a9a019@strath.ac.uk> Dear all, Our local IT support wants to shut down the CIS mailing list server, so we have finally managed to migrate the msp-interest list to the university mailing list server. The new address for posting a message is ? msp-interest at lists.strath.ac.uk (note the absence of "cis" in that domain name). Please update your address books! We should have moved all subscribers across to the new mailing list, and indeed, if you are receiving this message, you have been successfully migrated. Just in case, I will send a message to the old mailing list shortly, and then I imagine that CIS IT support would like to close it down sooner rather than later. As part of the move, I've requested for all permanent MSP staff to act as moderators, so approving new subscription requests should be quicker now. The msp-interest mailing list is dead! Long live the msp-interest mailing list! Best wishes, Fred From dilsat.yuksel at strath.ac.uk Mon Jan 13 14:55:03 2025 From: dilsat.yuksel at strath.ac.uk (Dilsat Yuksel) Date: Mon, 13 Jan 2025 14:55:03 +0000 Subject: [msp-interest] =?utf-8?q?__=5BMSP101=5D_April_Gon=C3=A7alves=3A_Type-Sensitive_?= =?utf-8?q?Algebraic_Macros_=28LT1414a=2C_1pm_Fri_17-Jan=29?= Message-ID: Hi everyone, On Friday, the MSP101 talk will be given by April Gon?alves (MSP). See the details of the talk below. Best, Dilsat ********************************************************************** Date: Friday, 17 January at 13:00 Room: LT1414a, Livingstone Tower Zoom link: https://strath.zoom.us/j/85449272187?pwd=0qNkQqybiaKgBpbEBRv38x11x4db5n.1 Speaker: April Gon?alves Title: Type-Sensitive Algebraic Macros Abstract: Despite recent advances made by Idris and Lean teams, metaprogramming in a typed language is still hard, frustrating and error-prone. In this short paper, we investigate a new view on macros via a type-sensitive algebraic theory for typechecker scripting for a more principled approach to type-directed macros. We show that our theory encodes typechecking and elaboration for ST?C, and from there, we build two other variations, Bidirectional ST?C and Search-based Type Inference, to showcase the versatility of our framework. Our results are implemented in Agda. ********************************************************************** -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.strath.ac.uk/archives/msp-interest/attachments/20250113/7dac1270/attachment.html From fredrik.nordvall-forsberg at strath.ac.uk Tue Jan 14 09:56:54 2025 From: fredrik.nordvall-forsberg at strath.ac.uk (Fredrik Nordvall Forsberg) Date: Tue, 14 Jan 2025 09:56:54 +0000 Subject: [msp-interest] Re: =?utf-8?q?=5BMSP101=5D_April_Gon=C3=A7alves=3A_Type-Sensitive?= =?utf-8?q?_Algebraic_Macros_=28LT1414a=2C_1pm_Fri_17-Jan=29?= In-Reply-To: References: Message-ID: Hi, As mentioned on Zulip, we will also have the 101 planning meeting for the upcoming semester after April's talk. Pre-planning is encouraged in the #UoS channel on Zulip (https://spls.zulipchat.com/#narrow/channel/227686-UoS/topic/MSP101.20planning ), especially if you cannot be there on the day! Cheers, Fred (who is not at all sending a message to debug a delivery issue for someone coming from the old mailing list) On 13/01/2025 14:55, Dilsat Yuksel wrote: > > Hi everyone, > > On Friday, the MSP101 talk will be given by?April Gon?alves (MSP). See > the details of the talk below. > > > Best, > Dilsat > > ********************************************************************** > > Date:?Friday, 17 January at 13:00 > > Room:?LT1414a, Livingstone Tower > > Zoom link: > https://strath.zoom.us/j/85449272187?pwd=0qNkQqybiaKgBpbEBRv38x11x4db5n.1 > > > Speaker:?April Gon?alves > > Title:?Type-Sensitive Algebraic Macros > > > Abstract:?Despite recent advances made by Idris and Lean teams, > metaprogramming in a typed language is still hard, frustrating and > error-prone. In this short paper, we investigate a new view on macros > via a type-sensitive algebraic theory for typechecker scripting for a > more principled approach to type-directed macros. We show that our > theory encodes typechecking and elaboration for ST?C, and from there, > we build two other variations, Bidirectional ST?C and Search-based > Type Inference, to showcase the versatility of our framework. Our > results are implemented in Agda. > > > ********************************************************************** > > > _______________________________________________ > msp-interest mailing list > msp-interest at lists.strath.ac.uk > http://lists.strath.ac.uk/mailman/listinfo/msp-interest -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.strath.ac.uk/archives/msp-interest/attachments/20250114/ec369fdf/attachment.html From dilsat.yuksel at strath.ac.uk Fri Jan 17 17:29:02 2025 From: dilsat.yuksel at strath.ac.uk (Dilsat Yuksel) Date: Fri, 17 Jan 2025 17:29:02 +0000 Subject: [msp-interest] [MSP101] Conor McBride: My Favourite Double Category (LT412, 3pm Mon 20-Jan) Message-ID: Hi everyone, On Monday, the MSP101 talk will be given by Conor McBride (MSP). See the details of the talk below. Best, Dilsat ********************************************************************** Date: Monday, 20 January at 15:00 Room: LT412, Livingstone Tower Zoom link: https://strath.zoom.us/j/85449272187?pwd=0qNkQqybiaKgBpbEBRv38x11x4db5n.1 Speaker: Conor McBride Title: My Favourite Double Category Abstract: I'll (re)introduce the "codeBruijn" representation of syntax with binding, in which terms are intrinsically indexed over their support (the variables they are involved with) rather than their scope. These supports form the objects of a category of order-preserving injections ("thinnings" to MSP locals and "my triangle" to Blaise Pascal), whose dual ("selections") plays a crucial role in managing the restrictions of the variable support in specific substructures. Meanwhile, the notion of "simultaneous substitution" tightens to a relevant structure where every variable in the source support has an image, and every variable in the target support is used by at least one of those images. The action of substitution (which includes substitution composition) respects support precisely. It is reliant on the way a selection from a substitution's source support variables retains only some of their images and is itself relevant only once we have identified the corresponding selection from the target support of only those variables used in the retained images. We acquire squares with selections for vertical edges and relevant substitutions for horizontal edges which compose in both dimensions, forming a double category. (For anyone who saw, or for that matter, gave Phil Wadler's talk at the last SPLS, codeBruijn shifting happens at the root of a term, obviating the complex relationship between renaming and substitution which happens only because de Bruijn shifting happens at the leaves.) (For dependently typed programmers, more generally, there may be transferable lessons in managing coherence.) ********************************************************************** -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.strath.ac.uk/archives/msp-interest/attachments/20250117/0e47bb00/attachment.html From dilsat.yuksel at strath.ac.uk Fri Jan 24 09:49:11 2025 From: dilsat.yuksel at strath.ac.uk (Dilsat Yuksel) Date: Fri, 24 Jan 2025 09:49:11 +0000 Subject: [msp-interest] =?iso-8859-1?q?__=5BMSP101=5D_Andr=E9_Videla=3A_Pipelining_and_d?= =?iso-8859-1?q?ependent_types_=28LT412=2C_3pm_Mon_27-Jan=29?= Message-ID: Hi all, On Monday, the MSP101 talk will be given by Andr? Videla (MSP). See the details of the talk below. Best, Dilsat ********************************************************************** Date: Monday, 27 January at 15:00 Room: LT412, Livingstone Tower Zoom link: https://strath.zoom.us/j/85449272187?pwd=0qNkQqybiaKgBpbEBRv38x11x4db5n.1 Speaker: Andr? Videla Title: Pipelining and dependent types Abstract: Pipelining is a crude CPU optimisation to speed up execution of sequential programs. By allowing the manipulation of types as values, dependent types enable new abstractions to manipulate large-scale of software based on sequential computation inspired by CPU pipelining. I will show how to define pipelines, sequential programs that cannot easily be parallelised using a list of types as a specification. Given this list of type, the implementation of a pipeline is given by the functions between each type. This pipeline abstraction can be generalised to categories such that the pipeline is made up of objects and its implementation is given by the morphisms between them. This enable the use of kleisli morphisms for effectful programs. The pipeline can be further abstracted over graded categories enabling automatic composition of errors. Finally, I will show how to use dependent pipelines to allow composition of programs with type-dependency. An architecture particularly useful for single-pass compilers with well-scoped and well-typed implementations. ********************************************************************** -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.strath.ac.uk/archives/msp-interest/attachments/20250124/6b89fc13/attachment.html From dilsat.yuksel at strath.ac.uk Thu Jan 30 16:15:59 2025 From: dilsat.yuksel at strath.ac.uk (Dilsat Yuksel) Date: Thu, 30 Jan 2025 16:15:59 +0000 Subject: [msp-interest] [MSP101] Jules Hedges: We solved dependent optics! (LT209, 3pm Mon 3-Feb) Message-ID: Hi all, On Monday, the MSP101 talk will be given by Jules Hedges (MSP, Cybercat Inst.). See the details of the talk below. Best, Dilsat ********************************************************************** Date: Monday, 03 February at 15:00 Room: LT209, Livingstone Tower Zoom link: https://strath.zoom.us/j/85449272187?pwd=0qNkQqybiaKgBpbEBRv38x11x4db5n.1 Speaker: Jules Hedges Title: We solved dependent optics! Abstract: There are two major generalisations of lenses. First there are optics, which require almost nothing of their base category and give you something in return. And then there are dependent lenses (aka container morphisms, aka natural transformations of polynomials), which require a lot of their base category but give you even more in return. One day several years ago I innocently twote the question of how to unify these two constructions, which is motivated for mathematical, computational and sociological reasons. The problem ended up occupying us (joint work with Dylan Braithwaite, Matteo Capucci, Bruno Gavranovi?, Eigil Rischel and Andr? Videla) for several years, and its difficulty became a meme. I will explain the answer that finally satisfied us. This involves first constructing a category of "dependent adaptors" and then freely adjoining a particular family of monoidal costates using a technique we call "weighted coparametrisation" that we reinvented. The definition began life as a prototype in Idris using QTT features, and was then translated back into category theory using what might or might not be a novel semantics of polymorphic dependent type theory. ********************************************************************** -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.strath.ac.uk/archives/msp-interest/attachments/20250130/b60a2583/attachment.html