borar @borar.bsky.social · 18/07/2026Same story, repeated over the years. If it's something F# it's all laughter and jokes. When a similar feature is added to C# it is suddenly the best thing in the world. 030
Reposted by borarHaskell programming language @haskell.org · 02/07/2026Hyperbole 0.7 has been released! The HTMX-inspired framework for interactive web applications now supports multipart file uploads. discourse.haskell.org/t/ann-hyperb... #Haskelldiscourse.haskell.org[ANN] Hyperbole 0.7 - Multipart form file uploads, immediate triggershyperbole.live github hackage There were many changes and improvements, but the biggest new feature is Multipart Form file uploads. It works with the default settings using a secure configuration, bu... 0103
borar @borar.bsky.social · 23/06/2026If I understand correctly, you use Azure "something" (not very familiar, sorry) to serialize concurrent operations. Is it because your system already happens to use Azure for other things? Otherwise a more light-weight and local mechanism could be used as well? Like a mailbox processor. 100
Reposted by borargabby @fullmoon.id · 23/06/2026I wrote up a post that explains how type inference for anonymous records works, including row polymorphism The post is also a gentle introduction to type theory for newcomers: haskellforall.com/2026/06/reco...haskellforall.comRecord type inference for dummiesGentle tour of type inference for records 18413
borar @borar.bsky.social · 19/06/2026Short-circuiting is useful for a rule which can decide the final outcome "on its own". The CE syntax hides all the machinery and makes data dependency between rules easy syntactically. Coming back to your original remark, CPS and its short-circuiting is very much idiomatic and important in FP. 000
borar @borar.bsky.social · 19/06/2026One of my favorite applications of the continuation type is with domain modeling. It is useful to represent composable rules. For example, a set of rules which together decide whether to apply a discount based on various factors. All rules have the same type and can be combined nicely. #fsharp 110
borar @borar.bsky.social · 19/06/2026As for the CE, just like with other "computations as values", it makes it easy to use the type by adding some syntactic sugar. So the goal is certainly not to "lure" C# programmers in, or anyone else :) 100
borar @borar.bsky.social · 19/06/2026The continuation monad and continuation passing style are quite fundamental and important. For example, the main Giraffe HTTP handler type is almost this, but not quite, also supporting "early returns". There are multiple other applications. 100
borar @borar.bsky.social · 19/06/2026Indeed, the continuation monad is more general, especially the continuation monad transformer. My gist instantiates the transformer for Async because someone asked about early returns in Async some years ago :) 000
borar @borar.bsky.social · 19/06/2026Another approach for early returns in #fsharp, but specifically for Async computations: gist.github.com/Savelenko/5e...gist.github.comAsync with early return capability CE for F#Async with early return capability CE for F#. GitHub Gist: instantly share code, notes, and snippets. 252
borar @borar.bsky.social · 14/06/2026This is a sub-topic I am going to focus on soon. In my view it is not part of the domain/model. In terms of the Onion architecture, it lives in the Application layer. Use-cases own Fold-s and a Fold uses the building blocks provided by the model. The model does not know anything about records. 100
borar @borar.bsky.social · 09/06/2026Blog: Basic principles of record storing systema10.org/posts/basic-... #eventsourcingsystema10.orgBasic principles of record storing 110
borar @borar.bsky.social · 08/06/2026@ursenzler.bsky.social I was catching up on your ES blog articles and read about secondary events. I think this resonates with one of the reasons for the "record" terminology I discuss in my article above. 110
borar @borar.bsky.social · 08/06/2026A new blog article where I go against the de facto normal: "Event sourcing is record storing" #eventsourcing systema10.org/posts/event-...systema10.orgEvent sourcing is record storing 111
borar @borar.bsky.social · 08/12/2025Very similar experiences here, only longer and with 10x as many engineers. 010
borar @borar.bsky.social · 27/11/2025I think #eventsourcing is a bit of a misnomer resulting in much noise and confusion. I've always used the terms "record" and "journal" and propose "record storing" (RS) instead. 000
borar @borar.bsky.social · 25/11/2025Store (NumberConsumed n) record. When the next number is needed, fold the latest n from the journal, specifically from these records and/or a snapshot. Use the number n+1 and store (NumberConsumed (n+1)) record while detecting concurrent changes to the journal. Retry if a change is detected. 010
borar @borar.bsky.social · 25/11/2025F# has interfaces with static abstract members (IWSAMs, type-bound) which are one step closer to type classes than regular object-bound interfaces. Still not actual type classes though. 010
borar @borar.bsky.social · 28/08/2025What I meant is the term "railroad" is specifically conceived in the F# "world" so it's logical that one does not find it elsewhere. CEs are a concrete language feature which, although similar to do notation, is unique enough. So similarly normal to not see it elsewhere :) 120
borar @borar.bsky.social · 03/04/2025This is another formulation of my point. Unfortunately the "human factor" seems to be somehow a notion in the software industry. But it is nonsensical in other engineering disciplines. Rightly so, because it comes down to individualistic anti-intellectual stance and that learning is special. 100
borar @borar.bsky.social · 02/04/2025The "needs" are indeed relevant when we deal with different categories of applications: F# is not better for systems or embedded or game programming, Rust, C, etc. are better. But F# and C# are the same category so the "needs" are the same. 200
borar @borar.bsky.social · 02/04/2025While I understand the good intentions behind such reocurring statements, I am less careful and "politically" balanced: #fsharp is absolutely inherently better. The "needs" are irrelevant, objective properties are what matters in engineering. Just like precise power tools vs manual mediaval tools. 260
borar @borar.bsky.social · 13/03/2025I am guessing you mean something like "if/when the last F# programmer leaves"? This is very unlikely, either there are more remaining or others who can pick it up. People can learn things :) 100
borar @borar.bsky.social · 13/03/2025The right answer is indeed "you don't need to hire F# programmers". 110
borar @borar.bsky.social · 12/03/2025Even more so, given that the seminal paper on F# computation expressions just uses the monad and monoid terminology extensively, as it should. 000
borar @borar.bsky.social · 10/03/2025Functional programming and its idioms are the most fitting and sensible interpretation/realization of #eventsourcing. Fold multiple composable values of types provided by the layered model, alongside the use-case specific version, apply model logic, store a new record. "Aggregates" are a mistake. 020
Reposted by borarAnupam 🇮🇳《ミ》λ≡ @aj.social.id1.in.ap.brid.gy · 09/03/2025Why would people use other programming languages for frontend dev when #PureScript is just right there? #FunctionalProgramming 111
borar @borar.bsky.social · 10/03/2025I think I understood the popularity of "modern" as a term. It embodies the implicit and unconcious understanding that most of the (visible online) software industry is a web of ad-hoc, anti-fundamental knowledge, hype-driven, half-assed ideas which break easily. Then its on to the next "modern". 010
borar @borar.bsky.social · 27/02/2025Usually the "mini applications" are easy to recognize and define unambiguosly. Evolution of long-lived business applications follows this as well: a new (business) capability needs to be added, which clearly belongs to the application, but is stand-alone within the problem domain. 100
borar @borar.bsky.social · 27/02/2025Persistence is also, in general, shared. Some DB tables can be used by multiple components, some are specific to one. With #eventsourcing the journal and the record type is shared. 000
borar @borar.bsky.social · 27/02/2025Every vertical component has its own data access interfaces (remember, Onion), URL handlers, data access implementation, tests, etc. Vert. components can also share types and functions from "common modules" which are at the "base" of the vert. components. For example the common User type. 🔽 100
borar @borar.bsky.social · 27/02/2025All use-cases for stock management of scheduling are two other vertical components. Multiple vertical components can have their own models of the same domain notion, tailored at their purpose. A User type appears everywhere, but is more extensive/different in the user management vert. component. 🔽 100
borar @borar.bsky.social · 27/02/2025A vertical slice, which I usually call a vertical component, is a cake piece going through all layers of the Onion. It is a mini application on its own, in a way, focusing on a set of related use-cases. Not too fine-grained, for example all use-cases of user management is one vert. component. 🔽 200
borar @borar.bsky.social · 27/02/2025My 5c about this, as always somewhat heretical. "Bounded context" is vague and not really interesting/useful. Vertical slice architecture is not an architecture, but a principle *within* an architecture. In particular, in the Onion architecture (my default). #DDD 🔽 110
borar @borar.bsky.social · 21/02/2025Only if you are making a fat client and the back-end is merely a thin layer between it and the DB and other infra. 000
borar @borar.bsky.social · 17/02/2025I see. What is the difference with a function returning a Lazy and possibly performing side-effects? 000
borar @borar.bsky.social · 15/02/2025Once evaluated, Lazy always returns the same value without re-evaluation. Does this "caching" also happen here? 100
borar @borar.bsky.social · 15/02/2025Having said that, PureScript is like you are suggesting: F# 'match <exp> with' in PureScript is 'case <exp> of' and 'function' is simply 'case _ of'. 000
borar @borar.bsky.social · 15/02/2025A function in its nost basic representation, as learned at school, is a table with two columns. From this point of view, the 'function' keyword and what follows it makes total sense. 100
borar @borar.bsky.social · 11/02/2025Just to be sure, are you compiling in the release mode with tail call optimization turned on? Secondly, did you try making the nested function a top-level one? 100
