|
Sapozhnik posted:imagine if this was the only system call interface that existed and you wouldn't be too far off
|
# ? Mar 30, 2019 19:20 |
|
|
# ? Jun 6, 2024 21:15 |
|
some academics were nice enough to write up a paper on why fork is garbage: https://www.microsoft.com/en-us/research/publication/a-fork-in-the-road/
|
# ? Apr 10, 2019 19:06 |
|
yeah we've known all of this for years and yet some people still stubbornly refuse to admit that fork was a bad idea
|
# ? Apr 10, 2019 19:13 |
|
fork was a bad idea but threads are also a bad idea
|
# ? Apr 10, 2019 19:29 |
|
I always thought I just didn't understand how fork really worked because my understanding of it -- it just splits one process into two processes with exact same state and access to the same memory? Neither of them knows whether it's the original without checking PIDs?? -- was so obviously deranged.
|
# ? Apr 10, 2019 19:31 |
|
both get copies of the memory space. and you know whether you are the parent or the child based on the return value of fork().
|
# ? Apr 10, 2019 19:32 |
|
fork was fine for limited uses but the fork+exec combination instead of having a real process spawning api was dumb
|
# ? Apr 10, 2019 19:49 |
|
in other news i have a post about graphics / web platform development https://blog.mecheye.net/2019/04/6-years-of-noclip-website/
|
# ? Apr 10, 2019 20:21 |
|
Suspicious Dish posted:both get copies of the memory space. and you know whether you are the parent or the child based on the return value of fork(). Like I said, deranged.
|
# ? Apr 10, 2019 20:25 |
|
Suspicious Dish posted:in other news i have a post about graphics / web platform development https://blog.mecheye.net/2019/04/6-years-of-noclip-website/ wow that is a really neat website for the game art. i hope you keep working on it also it looks like you missed a word in the opengl section where you say "you still need to which ideas are bad and why"
|
# ? Apr 10, 2019 20:35 |
|
Suspicious Dish posted:in other news i have a post about graphics / web platform development https://blog.mecheye.net/2019/04/6-years-of-noclip-website/ oh neat i saw that site around and didn't know it was you
|
# ? Apr 10, 2019 22:26 |
|
Suspicious Dish posted:in other news i have a post about graphics / web platform development https://blog.mecheye.net/2019/04/6-years-of-noclip-website/ i could not agree more on the paragraph about refactoring, to the point where it is in spirit very applicable outside of hobby projects as well.
|
# ? Apr 11, 2019 09:47 |
|
i disagree with the part on refactoring coming up with better organization or abstractions is half the fun of programming on hobby fun projects (and thats the source of its issues on non-hobby projects)
|
# ? Apr 11, 2019 14:57 |
|
Suspicious Dish posted:in other news i have a post about graphics / web platform development https://blog.mecheye.net/2019/04/6-years-of-noclip-website/ Oh, btw, i think you've got a typo here: quote:If you want to contribute, please join the join the Discord!
|
# ? Apr 11, 2019 15:11 |
|
Refactor a small amount of a dealio and you have to basically test it again This is actually peachy on well tested (meaning automated and maybe randomized testing) projects. Very few well tested hobby projects
|
# ? Apr 11, 2019 15:13 |
|
Meanwhile, I have a long list of things I'd love to refactor but having actual users + API stability guarantees means that I can't
|
# ? Apr 11, 2019 15:37 |
|
bob dobbs is dead posted:Refactor a small amount of a dealio and you have to basically test it again the few times i have been really angry with a colleague have been when people decided to "clean up" or refactor some piece of ancient code in the middle of the system with a release looming. invariably breaking something somehow, and bestowing no advantage as the code was untouched for half a decade because it already did the right thing, and it doesn't matter how clean it is when no one ever touches it anyway.
|
# ? Apr 11, 2019 15:41 |
|
Yeah, the attractiveness of well refactored code is that its easier to test. But refactoring actually tested code makes it untested. Threw the baby out with the bathwater Limiting factor to the actual value of code is in its being tested, most of the time
|
# ? Apr 11, 2019 15:45 |
|
Thermopyle posted:i disagree with the part on refactoring My problem with refactors is that they seem like such a nice idea at first, but you always run into The Kludge, and then you go "well, um, hmm", and you think about it for a long time and you're really not that much happier from when you started. These sorts of high-level problem solving ideas are fun to tinker with in the abstract, but when rubber hits the road you always make concessions. But I'm someone that much prefers to Just Write The drat Code than build a configurable meta-system. If code can be shared, share it. Otherwise, don't. And if it's halfway to a fit, it probably makes more sense to write it again than force yourself to share it. Repeat yourself when it would be less effort than not. Thermopyle posted:Oh, btw, i think you've got a typo here: Lutha Mahtin posted:also it looks like you missed a word in the opengl section where you say "you still need to which ideas are bad and why" thank you; fixed
|
# ? Apr 11, 2019 16:38 |
|
Cybernetic Vermin posted:it already did the right thing, and it doesn't matter how clean it is when no one ever touches it anyway. sometimes the reason is bad, poorly documented/ explained, or no longer required but somebody typed it and got it into a product refactoring locally is IMO usually the best way to understand legacy code submitting huge refactors at the end of a project is loving stupid though, as in any nontrivial code base there will be issues that no amount of automated testing will catch very often these are actually unintentional bug fixes that break expectations...
|
# ? Apr 11, 2019 17:34 |
|
Remulak posted:refactoring locally is IMO usually the best way to understand legacy code yeah, if I’m faced with some weirdass complicated code that does strange things for seemingly no reason, my usual approach is to just try to refactor it into something sensible. 99% of the time I end up just throwing away the result of this because all I achieve is making the code bad in different ways, but that’s okay because the point was to discover why the code is bad, not to improve it
|
# ? Apr 12, 2019 01:43 |
|
Plorkyeran posted:the point was to discover why the code is bad that’s like saying the point is to discover why the fire is hot
|
# ? Apr 12, 2019 03:24 |
|
Thermopyle posted:i disagree with the part on refactoring I disagree with the part about trying to cram applications into HTML and CSS via JavaScript like, SD’s stuff is super impressive, but I should be able to visit a web site on one of my computers older than they are and have it work well in other words, modulo HTTPS cipher/cert support, if a site doesn’t work reasonably or perform acceptably in Netscape 2 with JavaScript disabled on an early-1990s Mac, it’s not the loving web
|
# ? Apr 12, 2019 06:11 |
|
Yospos is the only place i've ever heard the opinion "web uis are bad" All I remember from my early career was the same lamentations but it was java. So many shifty java apps that half worked
|
# ? Apr 12, 2019 10:37 |
|
amd you had to download and install them with access to your whole filesystem, what a nightmare
|
# ? Apr 12, 2019 12:20 |
|
rt4 posted:amd you had to download and install them with access to your whole filesystem, what a nightmare as in http://www.fart.app/ had access to your whole system if you wanted to use their in browser java app?
|
# ? Apr 12, 2019 12:27 |
|
no i wasn't really thinking of applets since they rarely appeared in the wild
|
# ? Apr 12, 2019 12:51 |
|
the only applet that ever mattered was runescape
|
# ? Apr 12, 2019 14:53 |
|
Suspicious Dish posted:My problem with refactors is that they seem like such a nice idea at first, but you always run into The Kludge, and then you go "well, um, hmm", and you think about it for a long time and you're really not that much happier from when you started. The road to hell is paved with, "Why can't we just...."
|
# ? Apr 12, 2019 14:58 |
|
suspicious dish's stuff is a great showcase of the promise of modern web tech (the x11 tutorials are stupidly brilliant, to the point where i think i hurt his feelings previously commenting that it was a huge loss to humanity that he didn't spend that effort on something a touch more actually relevant). that said both stuff like gmail and huge electron apps are absolute trash and i hate them. it is not really down to some very deep principle, best to leave it pretty case-by-case. rich programmability blended with running text in something that still approximates a page is what i still hope wins out though, even for stuff like email where gmail and its ilk have just bastardized desktop concepts into a browser window.
|
# ? Apr 12, 2019 15:38 |
|
in that sense i hope webassembly will take off. it's still web tech so full of idiotic crap, but it's basically finally allowing us to have 'desktop applications' that just happen to run in a new OS (the browser) with a new window manager (the browser).
|
# ? Apr 12, 2019 15:50 |
|
the way razor components does things where it computes dom changes server side and then pushes the diff to the client is very good. no javascript "development" required.
|
# ? Apr 12, 2019 15:58 |
|
all you need is an always-on low-latency internet connection!
|
# ? Apr 12, 2019 16:01 |
|
ErIog posted:The road to hell is paved with, "Why can't we just...."
|
# ? Apr 12, 2019 16:20 |
|
Cybernetic Vermin posted:suspicious dish's stuff is a great showcase of the promise of modern web tech (the x11 tutorials are stupidly brilliant, to the point where i think i hurt his feelings previously commenting that it was a huge loss to humanity that he didn't spend that effort on something a touch more actually relevant). nah, don't worry about that. I just gave up that ghost when I stopped caring about Linux. I'm very disillusioned with the trajectory of the software industry, and the original draft for that blog post was way more caustic. The things I care about don't seem to be the things many other developers care about. I'm happy to keep making toys and things off in my own little bubble for the rest of life. If other people like them, good! I'm sure my talents could be used to serve ads or whatever else does but I don't want to do that.
|
# ? Apr 12, 2019 17:03 |
|
Suspicious Dish posted:nah, don't worry about that. I just gave up that ghost when I stopped caring about Linux. The "software industry" has always sucked, imo
|
# ? Apr 12, 2019 22:07 |
|
Sagacity posted:in that sense i hope webassembly will take off. it's still web tech so full of idiotic crap, but it's basically finally allowing us to have 'desktop applications' that just happen to run in a new OS (the browser) with a new window manager (the browser). so ... it’s finally letting us have something we already had, but slower? hurrah does it fix widgets? no? then it’s pointless. the problem isn’t javascript, it’s trying to base ui on dom and css that were designed for documents not uis. they’re “fixing” the one part that already had good workarounds like typescript, and leaving the bad parts of web guis in place
|
# ? Apr 12, 2019 23:36 |
|
it's fine they'll just draw a new ui with decent widgets into a canvas
|
# ? Apr 12, 2019 23:42 |
|
Soricidus posted:so ... it’s finally letting us have something we already had, but slower? hurrah my father doesn't understand how to update software but he sure as poo poo understands how to restart chrome if it asks him to, and get automatic updates of google docs or whatever and with regards to css, there are specs coming up that will essentially allow you to render to a canvas in a more or less sane way please note that this is all webdev so i assume they will still manage to mess it up, reinvent six wheels and have to rewrite the whole stack five times, but that would count as "cautiously optimistic"
|
# ? Apr 13, 2019 08:07 |
|
|
# ? Jun 6, 2024 21:15 |
|
Sagacity posted:this may come as a shock but not having to install things is huge for most people app stores also solve this problem, with the added bonus of giving you apps with a better user experience, more likely to work without a constant internet connection, and not liable to disappear at a moments notice with no clear migration path to anything else because you don’t have any control over your own data like yeah installing and updating apps used to be a huge drag on like windows xp but we’ve moved on from that now
|
# ? Apr 13, 2019 08:45 |