Conversation

Notices

1. Infected Moomin (moon@shitposter.club)'s status on Tuesday, 28-Sep-2021 13:55:27 JST Infected Moomin

   in reply to

   • deprecated
   • cirnog
   @cirnog @deprecated for various reasons professionalism is not taken seriously in the computing fields, I think in part because anyone can experiment in it consequence-free until they build enough experience to cobble together a semi-working solution. The reason I brought up guilds was because even in the code of Hammurabi, if you were a house builder and your building fell down and killed somebody, you were put to death. You had professional responsibility that was higher than your employer.
   • バツ子（痛いの痛いの飛んでけ;; (shmibs@tomo.airen-no-jikken.icu)'s status on Tuesday, 28-Sep-2021 13:55:27 JST バツ子（痛いの痛いの飛んでけ;;

     in reply to

     • deprecated
     • cirnog
     @Moon @cirnog @deprecated bit of aside, but that Irving Finkel book, suggests the code was more a symbolic thing, like "carve it in stone and put it up for all to see" as a sort of "this is what we're going for, clean, fair, justice for everyone", and actual daily arbitration was more flexible and less absurdly punitive
   • Infected Moomin (moon@shitposter.club)'s status on Tuesday, 28-Sep-2021 13:55:28 JST Infected Moomin

     in reply to

     • deprecated
     • cirnog
     @cirnog @deprecated I almost completely agree with this post. I disagree slightly for two reasons. First computing isn't even a century old and professional guilds are thousands of years old so yes we are not as mature yet, and people care a lot less because most software is less critical than a house falling down and killing your family. Second, coding abstractions, code reuse and code modularity are foundational software engineering principles and we practice all of them now and didn't even 25 years ago. I admit that software library packaging has been almost as many steps back as forward and people keep making new systems without understanding what the fuck they're doing.
   • cirnog (cirnog@poa.st)'s status on Tuesday, 28-Sep-2021 13:55:28 JST cirnog

     in reply to

     • deprecated
     @Moon @deprecated I think guild are the wrong comparison, the industrial revolution is only ~250 years old so a better comparison is fields of engineering and subfields that developed in that time frame (chemical, electrical, petroleum, automotive). These fields have solid theoretical foundations due to the underlying physical theories they use, and they've continually developed by building on those foundations. Computing as a mathematical subject is relatively recent but computing as a subfield of electrical engineering has no excuse for it's sloppiness when it's built on top of electromagnetic theory.While it's not critical for a CRUD program to work properly, it is critical for control system software to keep machinery working, but there's no way to guarantee that boeing's outsourced pajeet code is any good because mission critical software developers are trained the same way as webdevelopers (self taught searching the web for hack solutions, maybe reading one or two books). The systematic acceptance of shitcode drives good code out of the market.And while it's great that we've currently got the software engineering principles you mentioned, there is nothing stopping the next generation of engineers from completely ignoring them because "we've got new tech and moore's law and don't need that anymore". No foundations means no permanent improvements.
   • Infected Moomin (moon@shitposter.club)'s status on Tuesday, 28-Sep-2021 13:55:29 JST Infected Moomin

     in reply to

     • deprecated
     • cirnog
     @cirnog @deprecated I think you guys are looking at the past with rose-tinted glasses.Stuff crashes, yeah there's code that has to handle it behind the scenes, that's life. The things crash now because remote systems sometimes go offline, things don't crash as much anymore because your entire office suite was a monolithic assembler program doing its own memory management.
   • cirnog (cirnog@poa.st)'s status on Tuesday, 28-Sep-2021 13:55:29 JST cirnog

     in reply to

     • deprecated
     @Moon @deprecated More fundamentally there hasn't ever been a time when software development was good engineering, it grew out of the idiosyncrasies of mathematicians and electrical engineers and developed so fast that developers never learned from their past mistakes and perpetually reinvented past solutions inside more and more layers of abstractions.Software has never had the shared foundational knowledge and universally agreed on best practices that other engineering disciplines have, so it's always been a mess.
   • Infected Moomin (moon@shitposter.club)'s status on Tuesday, 28-Sep-2021 13:55:30 JST Infected Moomin

     in reply to

     • deprecated
     @deprecated this is just curmudgeon stuff, there are backward and forward steps but there are more steps forward than backward.Nobody's stopping you from forming a company of superprogrammers that know the most and and don't need layers of abstraction to make things. You will just be outcompeted by a pajeet company that slapped together an MVP 10x faster than you and iterated until it worked servicably. That has nothing inherent to do with technology community, it's just how sales works.people also don't remember how shitty software used to be, everything used to have crippling bugs and now we have annoyances but people wouldn't tolerate a wordprocessor or OS that crashed literally every five minutes.
   • cirnog (cirnog@poa.st)'s status on Tuesday, 28-Sep-2021 13:55:30 JST cirnog

     in reply to

     • deprecated
     @Moon @deprecated You don't understand, most abstractions are just APIs to let different software and drivers talk with eachother. Windows has to be compatible with every printer made in the last 50 years so it needs endless amount of hacked together shitcode to work. Things are getting slower because there's a proliferation of garbage that you just have to write code around or nothing works.>everything used to have crippling bugs and now we have annoyances but people wouldn't tolerate a wordprocessor or OS that crashed literally every five minutes.We still have that but now it crashes on the API level either on your computer or on a server somewhere and silently retries several times, which is why your computer takes five minutes to load a web browser instead of just crashing.
   • deprecated (deprecated@poa.st)'s status on Tuesday, 28-Sep-2021 13:55:32 JST deprecated

     I have a somewhat unusual perspective in that I learned to program ~20 years ago but stopped programming for about a decade, before getting back into it 7-8 years ago

     almost all the new crap people are taught just gets in the way; it’s slow to develop with, often slow to run, and obfuscates what’s going on under the hood in ways that can make debugging very frustrating

     the industry standard is to use heavy abstractions for no benefit to programmers who are actually competent, and since I started programming again the hot new frontend abstraction (for example) has changed like four times while the actual web has been getting dramatically worse by the year

     it’s not progress, and I can’t even really call it regress because that doesn’t really cover it… it’s like we’ve fallen into a punji pit and aren’t even trying to get out because we’ve convinced ourselves it’s a good place to be