Enfin bref, cโest un membre de la TeamGo qui a fait ce retour je suppose ? ๐
04.03.2026 17:43
๐ 0
๐ 0
๐ฌ 1
๐ 0
Enfin bref, cโest un membre de la TeamGo qui a fait ce retour je suppose ? ๐
Ajoute ร รงa le tooling un peu faiblard dโun langage dynamiquement typรฉ comme Ruby, certaines sessions de dรฉbug se transforme en labyrinthe.
Je vois dโoรน la personne vient. Quand tu fais du โvraisโ OOP, tu suis le concept de late binding donc polymorphismes ร outrance. Ce qui veut dire que lire le code ne permet pas simplement de comprendre quelle partie est en rรฉalitรฉ exรฉcutรฉe.
whenever people talk about "shipping speed" they focus on 0 to 1
but what actually matters is how fast the team is still shipping a year into the project
this is why everyone seems like they "ship fast" but nothing actually seems to be getting done
Cลur sur Joseph โค๏ธ
What people are struggling with is less about the novel aspects of developing with AI, and more about the table stakes.
It's not the AI part that's so hard (turns out systems have never been that deterministic), it's the software part.
100x as much software, 100x as fast as before.
A screenshot of Picotron built-in Lua code editor.
Do I miss hand-crafting code?
Maybe. A little. Thatโs probably why Iโm playing with Picotron, writing Lua by hand in a code editor without even autocompletion.
But thatโs a hobby, not work. ๐
I might be missing a thing or two, but one thing is clear: โhand-crafting codeโ isnโt on the list.
Since Iโve introduced LLMs into my workflow, all these priorities have remained the same.
Delivery, maintainability, and stability metrics are still strongโsometimes even better.
...
- Relying on automated tests.
- Investing in an evolutionary architecture.
- Following DevOps and Platform Engineering practices.
- Using an efficient system to organize work (Lean, Agileโwhatever works; Iโll leave the debates to others).
Hereโs what actually worked for me:
- Bringing business experts and developers together (in the same room, Slack, Gather, etc.).
- Encouraging collaboration among developers (pairing, mob programming, or simply sharing knowledge).
- Maintaining fast feedback loops while programming.
...
Was โcrafting the perfect lineโ or โwriting all the code by handโ the most effective way to meet those expectations?
I donโt think so.
So what was expected?
- Delivering on time.
- Understanding the business.
- Ensuring the software platform is stable.
- Solving issues when they arise.
- Ensuring the software architecture doesnโt block new features.
- Occasionally, improving the developer tooling.
I have mixed feelings about this.
Kelseyโs stance was also mine for a long time.
However, Iโve realized something lately:
Crafting the perfect line of code or finding the ideal variable name was never expected of me.
Not by my employer. Not by my tech lead. Not by my peers.
Dac, merci pour les infos !
A titre perso, je m'en tiens ร ce bouquin: gameprogrammingpatterns.com
Quand je me rends compte qu'un bout de code va devenir difficile ร utiliser, je check ce livre.
รa me semble un bon compromis.
Aprรจs je bidouille juste, je ne fais pas de vrais prod. ๐
Ma comprรฉhension est que la plupart des jeux sont des projets "fire and forget". On les dรฉveloppes, on les sorts et รฉventuellement on les portes sur une autre plateforme. Fin.
Le projet est fini avant d'atteindre le stade oรน ces patterns apporte leurs ROI.
Pour avoir bossรฉ sur des codes bases de plus de 10 ans avec certaines qui utilisaient ces pratiques et d'autres non, celles avec design pattern and co รฉtait quand mรชme les plus agrรฉables ร maintenir. Pour autant, je me verrais pas essayer de transposer toutes ces pratiques sur du gamedev.
Je n'ai toujours pas compris d'oรน sort cette lute constante entre clean code (and co) vs les game dev. ๐ค
Clean Code, Clean Arch, Design Patterns, etc. ce sont des concepts qui sont nรฉes dans le monde de l'informatique de gestion, pour faire en sorte que des logiciels soient maintenable 10 ans aprรจs
Moi qui mรฉlange un yaourt La Laitiรจre avec du lait et qui jette tout รงa dans une yaourtiรจre Moulinex.
Bilbo looking at his phone top on bottom is ChatGPT After all, why not? Why shouldn't I keep it? You're absolutely right โ you found it, it's been with you a long while, and it's only natural to feel fond of something that's served you so well, especially when someone like Gandalf suddenly seems to want it for himself.
Okay folks, looks like I switched over to the enemy.
Claude Code is nice.
A bit, Iโm mostly working with Picotron so the environment is self-contained and doesnโt provide any AI stuff. Sometimes I ask a question about my code to Claude Web to have a quick insight and some food for thought.
New homemade pizza attempt ๐
Coucou les #gamedevs
On tente un projet un peu fou avec @ty.games : faire un รฉditeur coopรฉratif breton.
On en parle lร :
www.linkedin.com/posts/damien...
Cโest mon prochain objectif. Jโai un mini PC qui attend tranquillement dans un placard. ๐
Same.
Je paye pour Fastmail, 1Password et Cozy Cloud.
Mรชme mes potes dev ne comprennent pas. ๐
Chronicles IV: Ebonheim is a classic dungeon-crawling CRPG with a non-procedural world, nuanced character building, and brutal, turn-based tactical combat.
Wishlist today at store.steampowered.com/app/3919080/...
#indiegames #sologamedev #ultima #roguelike #gamedev
First attempt (New York Style)
My wife got me a pizza oven for Christmas. Now, I need to practice. ๐จโ๐ณ ๐
A modern looking cyclops runs gleefully towards splore.p64
#picotron 0.2.2 is up on lexaloffle and humble: picotron.net
- Splore (ctrl-1 to toggle fullscreen!)
- Native Window Exports
- Search Across Files
- File chooser() / Generalised Drag & Drop
- Updated Manual with more sections & examples
Changelog: www.lexaloffle.com/dl/docs/pico...
This reflective acrylic sheet shows how 1980s illustrators were able to trace real-life objects on the computer. Once it was aligned, it reflected whatever was in front of it over the monitor's display, allowing artists to trace it in a tool like MacPaint or MS Paint. Simple, low tech and effective.
