|
rotor posted:Did you know? Progressive JPEG's avatar is actually a gif
|
#
?
Apr 25, 2022 23:37
|
|
|
|
| # ? Apr 28, 2024 04:48 |
|
|
im actually unironically considering whiteboarding out some architecture work for the backend behind that pager service to run card decks. oops i just reinvented JES again. neat. … how much volume and weight can a bike courier carry? what if im willing to pay an extra 5% , now how much can one carry?
|
|
#
?
Apr 26, 2022 00:14
|
|
|
11440 chars on 1“ stack of cards musl is like 60k lines of code , assume 80 char lines is good enough, thats roughly 60k cards ish and 143 cards os an inch so thats 420 inches of cards, 35 feet of cards. hm. hmmmmmmmm. how bout i pay the bike courier an extra 8%……
|
|
#
?
Apr 26, 2022 00:29
|
|
|
what
|
|
#
?
Apr 26, 2022 00:30
|
|
|
i have an actual pager server at work (with modems and everything) and as far as i can tell the architecture is "whatever"
|
|
#
?
Apr 26, 2022 00:31
|
|
|
nodejs isn't fundamentally bad, but npm is and kind of sours the whole ecosystem
|
|
#
?
Apr 26, 2022 00:34
|
|
|
Progressive JPEG posted:oh okay well the project isn't a web backend so if it doesn't have or use any kind of user interface you could make a case. most software does, though, and good luck selling that it shouldn't be the one everyone already has installed on every screen in their home/car/rocketship, has zero installation friction and no gatekeepers wanting a cut DaTroof posted:"but doctor," the patient sobbed, "i work for the paas provider!" fortunately it seems like a lot of paas providers these days just do complexity arbitrage on the other harder to use paas providers Lady Radia posted:so do you do anything except build crud apis or Progressive JPEG posted:if all you're doing is maintaining a crud app that serves 5 qps then using javascript is probably a fine way to save money since that kind of thing is pretty much a solved problem in any language, so it doesn't hurt to just use what the entry-level javascript-only devs that you exclusively hire already know i think that people's use of the word 'trivial' is an awesome shibboleth for determining if someone is a mid or senior engineer and something i look for when i interview applicants. mid levels will use 'trivial' dismissively. but trivial means obvious when it's wrong, cheap to debug and comprehensible by non-technicals. so a senior understands that when something is trivial it is a gift from the solution designer and will use it reverently. and a good solution designer understands that their job is to minimise complexity from the top down, to run to ground requirements that prevent software from being trivial. we finished a project last year that does $20-200 million a day of business at national scale, that turned a BU trying to make a profit to a BU trying to be a billion dollar company. there aren't many people who would describe that as trivial. but it was a full on JAMstack app on Next.js. the amount of things that javascript doesn't have a best in class library for is almost nothing and rapidly shrinking. because it has the most engineers. because it has the most companies. because it has the most engineers to recruit and libraries to use. it's not even the cheapest at this point, it's just being able to recruit full stop. as i said, i wish this wasn't the case, but as someone whose job includes making a business case internally and to clients, it is very difficult. you can't really put 'fuzzy happy feelings and a few cents of computer savings' against 'being able to hire both now and in 25 years' and 'having all the work already done for you for free six times'. the clients' job is to increase profits and reduce risk, right?
|
|
#
?
Apr 26, 2022 00:35
|
|
|
akadajet posted:nodejs isn't fundamentally bad, but npm is and kind of sours the whole ecosystem i think you will find that nodejs is actually very bad and npm is a symptom of that badness rather than a cause.
|
|
#
?
Apr 26, 2022 00:35
|
|
|
akadajet posted:nodejs isn't fundamentally bad, but npm is and kind of sours the whole ecosystem this is kinda where im at. i have issues with a some of nodejs generally but the absolute insanity of single-line dependencies is completely loving nuts.
|
|
#
?
Apr 26, 2022 00:38
|
|
|
rotor posted:this is kinda where im at. i have issues with a some of nodejs generally but the absolute insanity of single-line dependencies is completely loving nuts. that kind of rot could theoretically exist in any package system, but on node it's everywhere
|
|
#
?
Apr 26, 2022 00:40
|
|
|
i mean its gonna have to be regional, unless youre really into overnighting 35 feet of cards for real money and waiting 48 hours for job results. its better than we were getting in the 80s in the same building depending on if the ops hates you or not i guess. diffs might be easier, but then the ops center has to have storage for your 35 foot long card deck. and youll lose 8 chars per card for seq numbers. hmm. maybe we need some archival tool. oops invented tape again. neat. wait. paper tape? what densities did we get with paper tape??
|
|
#
?
Apr 26, 2022 00:42
|
|
|
80 chars in 14” of 1” tape. very slight area savings, but the stock is thinner so you can probably get more chars per cubic foot. would strongly depend on the stocks you were comparing tho. yeah that might work out better. ngl - kinda hurts that the pager is gonna necessitate a touch tone phone. oooh - set up a 6m receiver for cw to accept dispatch requests?? is there any 6m freq available for commercial use still????
|
|
#
?
Apr 26, 2022 00:57
|
|
|
chaosbreather posted:if it doesn't have or use any kind of user interface you could make a case. most software does, though, and good luck selling that it shouldn't be the one everyone already has installed on every screen in their home/car/rocketship, has zero installation friction and no gatekeepers wanting a cut you sound like that one spotify dude who constantly talked about how nodejs was the most productive way to program, just not in any measurable way.
|
|
#
?
Apr 26, 2022 01:23
|
|
|
Lady Radia posted:you sound like that one spotify dude who constantly talked about how nodejs was the most productive way to program, just not in any measurable way. hey thanks i really appreciate that or i'll get you gadget, whatever makes you happier
|
|
#
?
Apr 26, 2022 01:39
|
|
|
i unno what get me gadget means. is it good? that'd be nice 
|
|
#
?
Apr 26, 2022 01:39
|
|
|
https://www.youtube.com/watch?v=sZqVUjziexk
|
|
#
?
Apr 26, 2022 01:41
|
|
|
oh lol i get it now i hear part of what you're saying re: recruitment, but tbqh most of the problem i think with getting 'top talent' isn't the numbers, but more around actually paying the people who know poo poo well enough they accept. even when we were looking just for swift devs we'd have bites, they just avoided accepting (rightly so or whatevs). i dont think, imo, typescript avoids that as a full stack thing. but i understand there's lots and that matters. the good news is i dont deal with clients i work in house. gently caress yea
|
|
#
?
Apr 26, 2022 01:44
|
|
|
oh, youre using electronic storage? yeah, i used to use that too. i use paper tape now, and store data in 7 bit ascii, about 69 characters per foot. ( nice ) you really only grok your data when you see the individual bits, you know? *picks up wheelbarrow with 2MB of data in it and wheezes and puffs out the * hey. hey. hey. can you get the door for me? yeah. as far open as you can get it, no no stand on the other side and hold it from there, yeah *out the door*
|
|
#
?
Apr 26, 2022 02:03
|
|
|
Lady Radia posted:oh lol i get it now yeah for recruiting it's actually better to do something a bit weird and cool like elixir because you only get people who really know their poo poo and you can probably pay them much less but in terms of a strategic direction, numbers matter a lot to decision makers and it's genuinely concerning how few viable alternatives there are. .net is a non-starter everywhere that's been burned by microsoft before (p much everyone), and the numbers aren't there. and microsoft, creators of typescript, is making and more and more poo poo in react native for windows, like office and the new control panel in win11, so it's clear to decision makers that MS is doing to .net what they did to all their other poo poo java is in the headlines every other week for hilariously terrible security fuckups and everyone has been burned by a lovely java app rails is a pretty decent choice but getting a little stagnant php is hilariously expensive because no-one wants to program it cool languages they haven't heard of is completely off the table and anything that isn't javascript while you're making a web app (or native app capable when you're making a native app) means you have to have two codebases and two teams, which means that you have to deal with communication overhead, conflicting interests and politics between those two teams, which is by far the biggest risk point in any project technical or otherwise. i think wasm is going to change the landscape significantly, blowing it wide open for the first company to open source their good complex canvas-drawn ui framework, but until then not-node is a very very hard sell. even deno, aka, strictly better node, is a hard sell.
|
|
#
?
Apr 26, 2022 02:04
|
|
|
MICROFICHE!! HELL YEAH! were gonna make this work baby.
|
|
#
?
Apr 26, 2022 02:04
|
|
|
why do you believe you need two separate front end and backend teams?
|
|
#
?
Apr 26, 2022 02:12
|
|
|
Lady Radia posted:why do you believe you need two separate front end and backend teams? why do decision makers believe you need two separate teams if you use two separate languages, you mean? because i certainly don't. they believe it because of risk/scaling and politics. it's more risky to need a thing that does two things than a thing that does one. it's harder to find a thing that does two things if you want a lot of things done. a thing that does two things is probably not as good at doing each of the things as a thing that does one thing. and if there are two types of things, that's a power vacuum some up-and-comer can carve out from their kingdom, then hold the project hostage with both of these are sensible concerns and reasonable arguments. of course they are respectively wrong and irrelevant – the best team is about 2-4 highly skilled engineers, which is our general pattern. but they can't do what we do otherwise they wouldn't need us. they want to manage, that's why they manage, and they want to manage a lot because then they get a lot of money and importance. and the only reason why we are able to do what we do, fielding a number of highly skilled engineers for a variety of clients and projects, is we specialise on typescript which is what is available in terms of talent and what clients demand. combine that with the fact that in typescript land you have multiple entire technologies that eliminate needing any kind of backend code entirely and it's a fait accompli
|
|
#
?
Apr 26, 2022 02:37
|
|
|
also Conway's Law, which is legit
|
|
#
?
Apr 26, 2022 02:40
|
|
|
chaosbreather posted:why do decision makers believe you need two separate teams if you use two separate languages, you mean? because i certainly don't. it really sounds like your problem here is being put in situations where you're told to accept typescript, vs. actually being one of the decision makers supporting it
|
|
#
?
Apr 26, 2022 02:44
|
|
|
Lady Radia posted:it really sounds like your problem here is being put in situations where you're told to accept typescript, vs. actually being one of the decision makers supporting it actually i'm put in situations where i have to sell typescript. which is, unhappily, the right choice for the clients. one of my problems is that that makes for a very homogenous day. one of my problems is that a homogenous industry is really bad for a lot of reasons. one of my problems is i am doing commercial software at all rather than sitting on a beach making weird unplayable games. i have a lot of problems. typescript is, as it goes, not the biggest, and i suggest anyone who is uncomfortable with it to get used to it or wait 3 years for the next thing to come and demolish it.
|
|
#
?
Apr 26, 2022 03:03
|
|
|
believing typescript is going away in 3 years feels antithetical to your core point doesnt it? not being argumentative but like, one of your pieces here was that it had teeth
|
|
#
?
Apr 26, 2022 03:05
|
|
|
i just use regular javascript idk why people bother with typescript
|
|
#
?
Apr 26, 2022 03:09
|
|
|
quote:the amount of things that javascript doesn't have a best in class library for is almost nothing and rapidly shrinking. because it has the most engineers. because it has the most companies. because it has the most engineers to recruit and libraries to use. it's not even the cheapest at this point, it's just being able to recruit full stop. meanwhile, opting for any niche tool or library at the cost of an already well-supported version is risky. the re-implementation will still lack the documentation, tutorials, upstream maintenance, and stack overflow community of the original. if you're doing something trivial that stays within the tiny bounds of what the niche library happens to cover, then this probably won't matter. but as your project grows, it will become the guinea pig that discovers the many corner cases that the niche version doesn't support yet. but aside from all of the other reasons that it affects the stability of the project and the productivity of the team, javascript specifically has the stink of operating in an ecosystem of penny-pinchers who are weirdly focused on reducing compensation no matter the cost, instead of investing the career development of their employees. this is a signal to prospective hires and probably explains a lot of the problems you're seeing. in other words if someone came to me and said that i should focus on programming language instead of the stuff that actually matters to project lifecycle, i'd tell em to get lost
|
|
#
?
Apr 26, 2022 03:19
|
|
|
Lady Radia posted:believing typescript is going away in 3 years feels antithetical to your core point doesnt it? my core point is 'there is no better business case than typescript for project owners' that won't always be true but the fact it looks always true is a reason the business case is as good as it is. they'll be a new thing in 3 and that new thing will be dominant in 5ish but they will be in another job by then because they made the right decision with typescript.
|
|
#
?
Apr 26, 2022 03:20
|
|
|
the recruiting argument makes sense out at the tail end where, for instance, hiring a team of perl programmers is like actually really hard. but js vs java vs python vs whatever? its a nonissue.
|
|
#
?
Apr 26, 2022 03:21
|
|
|
rotor posted:the recruiting argument makes sense out at the tail end where, for instance, hiring a team of perl programmers is like actually really hard. in the business case it matters a lot. big number going up is a massive difference on paper to slightly less big number plateauing or going down. i can crush someone's dreams into pieces with slightly less big number going down on a google slide
|
|
#
?
Apr 26, 2022 03:28
|
|
|
chaosbreather posted:in the business case it matters a lot. big number going up is a massive difference on paper to slightly less big number plateauing or going down. i can crush someone's dreams into pieces with slightly less big number going down on a google slide thats super
|
|
#
?
Apr 26, 2022 03:29
|
|
|
i mean you're saying that it's easy to snowjob people into going with typescript. that isn't a compelling argument
|
|
#
?
Apr 26, 2022 03:32
|
|
|
chaosbreather posted:yeah for recruiting it's actually better to do something a bit weird and cool like elixir because you only get people who really know their poo poo and you can probably pay them much less ime experience you'll actually pay them more, but you'll also get people who know their poo poo chaosbreather posted:in the business case it matters a lot. big number going up is a massive difference on paper to slightly less big number plateauing or going down. i can crush someone's dreams into pieces with slightly less big number going down on a google slide stop being weird
|
|
#
?
Apr 26, 2022 03:42
|
|
|
thank you for the good chat and perspective. not sure i agree but i think i understand the rich folks perspective there
|
|
#
?
Apr 26, 2022 03:43
|
|
|
Achmed Jones posted:i mean you're saying that it's easy to snowjob people into going with typescript. that isn't a compelling argument an argument for what? what is it that you think i am arguing? i am saying it is bad how difficult it is to form a business case to use anything other than typescript.
|
|
#
?
Apr 26, 2022 03:43
|
|
|
chaosbreather posted:an argument for what? what is it that you think i am arguing? i am saying it is bad how difficult it is to form a business case to use anything other than typescript. hahahahahahaha i actually like typescript, and that statement is hilarious to me
|
|
#
?
Apr 26, 2022 03:44
|
|
|
DaTroof posted:hahahahahahaha me too and me too
|
|
#
?
Apr 26, 2022 03:45
|
|
|
chaosbreather posted:me too and me too how so the latter? are you saying you can't find jobs that aren't full-stack js/ts? because i guarantee that's a personal problem and not an industry-wide fact
|
|
#
?
Apr 26, 2022 03:47
|
|
|
|
| # ? Apr 28, 2024 04:48 |
|
|
DaTroof posted:how so the latter? are you saying you can't find jobs that aren't full-stack js/ts? because i guarantee that's a personal problem and not an industry-wide fact finding a job? what i'm saying it's really really funny that it's really hard to make a business case for using something other than typescript to project owners
|
|
#
?
Apr 26, 2022 03:49
|
|