jmoiron plays the blues

@jmoiron A task to make code "Cleaner" is a flag: it's almost always a no-impact medium-risk refactor with no measurable positive impact. Those cases where it's not should always describe the change and its benefits directly; "Clean" has very little value on its own.

@jmoiron Bangalter's "Mythologies" feels like the soundtrack to a lost animated version of the original NES Legend of Zelda instruction booklet.

@jmoiron Twitter was a place where every public figure's personal thoughts were accessible for engagement, with verified, trustworthy identification. People in the future will not understand why this was pissed away right as fact & fiction was being rapidly blurred by technology.

@jmoiron @copyconstruct move break and fast things

@jmoiron @elithrar I would never have felt confident enough to get into SLRs without DPreview. Really sad 😢

@jmoiron @glcst Yeah! In hindsight, the controversy around "birds are dinosaurs" does seem ridiculous when presented with these guys.

@jmoiron @glcst I've always found it crazy that the Australians fought and lost a war against a native bird species and that bird wasn't the Cassowary.

@jmoiron This is why the analogy is so frustrating. Many people use "YAML is k8s asm" to mean "you can build abstractions on top of k8s YAML", but that's not really true. You can build a static site generator, but you cannot change the way a checkbox works.

@jmoiron In the "YAML is k8s asm" analogy, asm is "an unwieldy layer you're not meant to use directly". But k8s YAML is not like asm, because k8s is not like a computer. YAML is like k8s HTML, k8s is a browser, and there is no JavaScript.

@jmoiron It's been really wild to see the "Sillicon Valley" brand completely collapse in my lifetime, from a symbol of inspiration and innovation in the 80s and 90s, to one of hubris and greed now.

@jmoiron @bryanl My approach changes slightly based on audience, but this is one of the main ones; I think of it as "socratic review". Questions invite conversation, and statements shut it.

@jmoiron Video game boxes did not survive my childhood, but I do have a bunch of SFC games w/ box that I bought in the early 2000s at Super Potato. https://t.co/5ZYAEXKoEJ

@jmoiron Generics use seems to fall into two main camps: faster type-safe data structures, and things I don't like. https://t.co/kFZGfzHYxp

@jmoiron A lot of people are completely playing themselves paying tons of cloud premium for shit they absolutely do not need.

@jmoiron "Get faster access to BingGPT by setting Edge as your default browser and downloading the Bing app." Microsoft is still Microsoft I guess.

@jmoiron If you can't see the slides for this talk on InfoQ, they can be found here: https://t.co/dpkZnHoNwu

@jmoiron And I do think we still have a field where academics rarely contend with the selection filters that govern this "natural world". Where they publish on obscure systems with no practical impact. Where they use words engineers know in ways engineers don't.

@jmoiron Regardless, the impact delay is a concern, and it's not limited to CS. Gregor Mendel described a set of laws for trait inheritance (known intuitively to farmers for millenia) in the 1860s, but his work went unknown for 40 years, delaying the modern study of genetics.

@jmoiron I have to admit, to me, the line has blurred. I've seen many impactful contributions from "academics" in the fields of performance engineering, databases, distributed systems, etc. Maybe this is a result of Gabriel's thesis, or maybe I only see the things I understand.

@jmoiron As I said before, I think these "dead" ideas are often rediscovered and put into practice. I think Gabriel might argue that these ideas are usually in papers written by engineers and not academics, but they are frozen out of being accepted as "scientific" literature.

@jmoiron Academics will sometimes lament that good ideas die on the vine, while crude implementations by practitioners gain traction despite their flaws. This is especially true, in my experience, in PLT.

@jmoiron I've always thought that there were selection filters not unlike natural selection. "Good" systems can leave no intellectual heirs, and "bad" systems can become so successful they demand study. To bring it back to Gabriel, this is the subject of one of his most famous essays.

@jmoiron The implications of systems as the natural world has more overlap with my areas of interest: if studying real world systems is important for computer science, what mechanisms determine which systems become successful?

@jmoiron In my experience as an engineer, I find these behaviours often _are_ forseen in the literature, but their predictions are too obscure. This may be due to the "incommensurability" Gabriel discusses in in his talk: academics publish in a language engineers do not understand.

@jmoiron As real computer systems become more complex, they often display new, unforseen emergent behaviours that become areas of scientific inquiry. I think we understand this: OSes, filesystems, and languages have been vibrant areas of study for a while.