Post #2418342
2025-10-09 09:48 UTC
Replies (20)
-
@RosaCtrl@social.vivaldi.net 2025-11-04 07:51
So when I made the above poll I had something like this in mind. Like doing the actual work instead of reviewing it puts you in a different place. And this is also why you would like feedback from a peer https://sauropods.win/@futurebird/115475206279444149
-
@RosaCtrl@social.vivaldi.net 2026-04-03 13:20
RE: https://social.vivaldi.net/@RosaCtrl/116340960391939171 So this 🆙 has a name
-
@benc@mastodon.hawaga.org.uk 2025-10-09 09:51
@RosaCtrl@social.vivaldi.net this happens to me enough that I usually use toy code as the design process, compared to how badly all my paper-designed stuff works out. And I actually manage to throw it away. Usually.
-
@mahryekuh@hachyderm.io 2025-10-09 09:55
@RosaCtrl@social.vivaldi.net There are situations where I’ve refactored code multiple times before it loses its draft status. (With a liberal use of git force pushing.) It’s usually because I make a sort-of working version first, and do refinement of the code later, but it can also be because of the situation you are describing.
-
@emma_cogdev@fediscience.org 2025-10-09 09:56
@RosaCtrl@social.vivaldi.net I think my answer is "almost always" 😅
-
@helle@tacobelllabs.net 2025-10-09 10:03
@RosaCtrl@social.vivaldi.net oops, misvoted, irregularly, but it happens but if we are deep in algorithmic weeds this becomes quite common
-
@wim_v12e@scholar.social 2025-10-09 10:18
@RosaCtrl@social.vivaldi.net often, but usually it is for what I would call the "tactical" decisions rather than the "strategic" ones. I rarely have to change the overall architecture radically.
-
@Raffzahn@mastodon.bayern 2025-10-09 10:24
@RosaCtrl@social.vivaldi.net It's just natural. After all, when coding one is forced to finalize an idea. This includes working out all those little details one just ignored before. Solving them is making the difference between a nice idea and a working implementation - wich more often than not means partial or full redesign. :))
-
@dodecadron@datasci.social 2025-10-09 10:36
@RosaCtrl@social.vivaldi.net For me implementing stuff is often a way to get to understand the problem.
-
@tinylittleenormous@piaille.fr 2025-10-09 10:41
@RosaCtrl@social.vivaldi.net I even start coding without an idea, and let the code I see give me the idea/ plan.
-
@lhengstmengel@mastodon.nl 2025-10-09 12:08
@RosaCtrl@social.vivaldi.net oh yes. Often you have to walk the path before you realize it is the wrong path.
-
@bartavi@mastodon.nl 2025-10-09 14:18
@RosaCtrl@social.vivaldi.net I like how this is most likely a subtle critique at vibe coding.
-
@samir@mastodon.functional.computer 2025-10-09 20:04
@RosaCtrl@social.vivaldi.net Er, constantly? It helps when I’m in a TDD flow and writing a little test, making it pass, repeating, but even then I will usually either realise the implementation is flawed half-way through, or highlight multiple areas for refactoring during implementation and sort them out afterwards. That’s when the test/design/requirement itself isn’t fundamentally broken, of course.
-
@beandreams@friendhole.social 2025-10-09 21:16
@RosaCtrl@social.vivaldi.net I love how Alan Perlis put it in the intro to SICP: "Even though our programs are carefully handcrafted discrete collections of symbols, mosaics of interlocking functions, they continually evolve: we change them as our perception of the model deepens, enlarges, generalizes until the model ultimately attains a metastable place within still another model with which we struggle....
-
@4censord@unfug.social 2025-10-09 21:35
@RosaCtrl@social.vivaldi.net yes! its actually kinda how i usually structure my dev process i think about it a little, and then i start writing stubs everywhere. if i notice anything wrong, i then refine the plan only then i start writing actual code for most things
-
@jkachmar@hachyderm.io 2025-10-09 23:18
@RosaCtrl@social.vivaldi.net since no one else has linked it yet that i can see, i defer to Peter Naur’s essay “Programming as Theory Building” https://pages.cs.wisc.edu/~remzi/Naur.pdf
-
@flippac@types.pl 2025-10-10 11:14
@RosaCtrl@social.vivaldi.net The bigger the chunk of code the more likely this is to happen in my experience! (Otherwise you're more likely to get it done and then have the idea and go back, or at least for the original idea to work enough it's worth finishing first so you have something that works) Do I change my mind? All the time, sometimes about something very old.
-
@jjj@functional.cafe 2025-10-11 21:12
@RosaCtrl@social.vivaldi.net Often. And often there is not sufficient time or priority to redo it a better way, because that would require significantly more work and/or refactoring/redesign of existing code.
-
@romeu@mastodon.social 2025-10-15 11:55
@RosaCtrl@social.vivaldi.net it may be an effect of the specific method I use for coding but in my case very often , by design.
-
@janneke@todon.nl 2025-10-15 12:19
@RosaCtrl@social.vivaldi.net I believe this is what Fred Brooks (The mythical man month, 1975) called: Plan to throw one away. I wrote three implementations of a declarative music typesetter: MPP (MuTeX Pre Processor in Pascal, on DOS), MPP (same in C, on Unix), MPP (MusiXTeX Pre Processor in C++) before Han-Wen, who joined me working on the final MPP, wrote his first: MPP95 (a failure, of course;) before h/we started #GNU #LilyPond in 1997. LilyPond got a major rewrite, overhaul in the v1.3 series where we integrated #Guile and exposed all functionality to the user, giving them full freedom and control. Like every program should, IMHO.