Andrey Hitrin

Software engineer, amateur musician, amateur naturalist. Happy father of little boy

[Andrey Hitrin @ahitrin]

@seewanderin "Туше" - только и смогла ответить учительница

[Andrey Hitrin @ahitrin]

@ArtemRusinovich @TinyBetelgeuse Косплеить сожжение Просперо

[Minh Nhat Nguyen @menhguin]

@bcherny @YashGouravKar1 how much code could a claude code code if claude code could code claude

[Andrey Hitrin @ahitrin]

@tolkienzone Tom Bombadil won't buy it ;)

[Andrey Hitrin @ahitrin]

@Spanish_Kortez

[Andrey Hitrin @ahitrin]

@ol_fa_sol Мне помогает выписывание беспокоящих мыслей из головы на бумажечку или в комп (но только *обязательно* в то место, где их не забуду потом найти). Иначе голова будет крутить их без конца чисто чтобы не потерять. А когда выписал, то голову отпускает. Метода Максима Дорофеева

[Andrey Hitrin @ahitrin]

@VladVoice @lubish_khleb К сожалению, только у 1 из 20...

[Allen Holub. https://linkedIn.com/in/allenholub @allenholub]

Often, when I talk about improving development speed, I get a lot of mean-spirited comments lecturing me about how output (as compared to outcome) is unimportant and "feature factories" are dysfunctional (that one's true, at least). Speed IS important, though. If you produce the best product on the planet too slowly, your competitors will eat you alive. If you fix bugs at a glacial pace, you'll lose customers. Most importantly, to the C-levels, time is literally money, and no business can afford to waste money.

Speed is not a zero-sum binary. You can get speed _without_ sacrificing quality or customer focus, but that takes change, and the people who gripe about slowness are often unwilling to make the necessary changes. TDD reduces defect density by as much as 90%. Working in high-quality code improves development speed by as much as 7x. Mob/Ensemble programming can give you a 3x improvement. Getting rid of bottlenecks also yields major improvement. You get the idea.

None of the above impacts outcomes. Sneering at these changes ("We can't do that ivory-tower stuff here!") while chanting the output/outcome mantra is really just fear of change sticking up its ugly head. Change can be risky, but no business can function without risk. Sorry.

(NB: I'm still cross-posting to X, but I don't check it very often. It's better to post comments to LinkedIn: https://t.co/QLHLwrizw0.)

[Andrey Hitrin @ahitrin]

@dyrnoe_vliyanie Комиссар дал бы пятнадцать лет... в Имперской гвардии! Навыки уничтожения ксеносов, особенно подручными средствами, тут в цене #WH40K

[Andrey Hitrin @ahitrin]

@hayate_hideki @Alexis_Colby332 Зато через 10-20 миллионов лет из этих отходов образуются залежи нефти! Гуманоиды будущего останутся благодарны

[Andrey Hitrin @ahitrin]

@sebullllba Рано. Ложный всепланетный Император ещё не оформился официально, а до этого времени освободительного флота Воителя нам не видать

[Programmer Humor @PR0GRAMMERHUM0R]

fullStack https://t.co/nxZ2ydV1Ki

[Punt Road @punt_rd]

Bring on New Year’s Eve…!! Play Iron Maiden in to Judas Priest just at the right time!🤘👍

[Andrey Hitrin @ahitrin]

@RonyVernet IN HOC SIGNO VINCES

[Andrey Hitrin @ahitrin]

@uskolzaichik "Возвышение Хоруса" и далее по списку. Ересь, конечно, но 55 томов позволят закопаться на пару-другую лет

[Allen Holub. https://linkedIn.com/in/allenholub @allenholub]

The answer to "how can we reduce <some estimate>" is always "build something smaller." I've never seen an exception.

[Andrey Hitrin @ahitrin]

@scukko That's 2012, actually. Linus have started to working on his language and manners since that time

[Andrey Hitrin @ahitrin]

@sobermediocracy @Dissolute_Psy Именно. Не "я решил", а "решение было принято". Попытка ухода от ответственности, осознанная или подсознательная

[vicki @vboykis]

The single most important thing I have learned about software development over my career is that if you do not aggressively fight complexity, it will eat you alive.

[Andrey Hitrin @ahitrin]

@jasongorman Haha, easy! Just use `SELECT *` and build graph in memory.
Jokes aside, this straightforward solution worked for me in one of projects. But there I didn't work with large graphs. At max, few 100s of nodes and edges. Even after several years, it still works fine :)