|
eschaton posted:that's because its creators were once smart but dropped out of the industry in the late 1970s and don't believe anything done since has any worth but I thought everything good in computer science was first done in the '60s and only now starting to trickle down to common langs checkmate????
|
# ? Apr 29, 2016 03:10 |
|
|
# ? May 25, 2024 12:12 |
|
Mahatma Goonsay posted:I just got to the point of learn you a haskell where they try and implement javascript truthiness and it made me think of this thread. https://github.com/Gabriel439/post-rfc/blob/master/sotu.md its really good at parsing & writing compilers, also pretty good at writing servers, and i love it for scripting instead of something like python. it's also used as a research language i guess which is why it gets some p esoteric features especially in the ghc extensions, but it isn't crazy like a real research language
|
# ? Apr 29, 2016 03:13 |
|
pokeyman posted:but I thought everything good in computer science was first done in the '60s and only now starting to trickle down to common langs most of the major advances in programming happened in the mid to late 1970s through the early 1980s at places like MIT AI and SAIL and Xerox PARC that was after the original C & UNIX creators stopped paying attention to the outside world there's a reason modern UNIX derives from the huge improvements added by UCB and Carnegie Mellon rather than the poo poo AT&T came up with (last good AT&T UNIX was v7)
|
# ? Apr 29, 2016 03:24 |
|
fart simpson posted:https://github.com/Gabriel439/post-rfc/blob/master/sotu.md you had me right up until that last clause Haskell is pretty crazy, and would be better written with more uniform syntax like one built out of s-expressions
|
# ? Apr 29, 2016 03:28 |
|
A large subset of Haskell's syntax is uniform and very good (anything dealing with function application), though it would've been better if (|>) and (>>) had been invented before ($) and (.)
|
# ? Apr 29, 2016 03:33 |
|
eschaton posted:there's a reason modern UNIX derives from the huge improvements added by UCB and Carnegie Mellon rather than the poo poo AT&T came up with (last good AT&T UNIX was v7) att had the grace and humility to know when they were licked. svr4 was the final admission that sun and ucb were right all along people who went on trolling about system v sucking in the 90s and 2000s were just weirdy beardies. realistically, the war ended in 1988
|
# ? Apr 29, 2016 03:57 |
|
Valeyard posted:does it "just work" and is actually useful? With "Build Automatically" and "Synchronize on Build" both turned on, updated class files get to my remote server in like two seconds after I hit save. Not bad.
|
# ? Apr 29, 2016 03:58 |
|
pokeyman posted:doesn't mean browsers all do the same/correct thing Yeah, this is the issue with failing gracefully, is that in the real world it's not graceful at all, SOMETHING depends on that HTML being correctly formed, CSS or Javascript. I wasn't part of the meeting that 'rejected' XHTML, but the fact that we can accept 'use strict' as a thing, but not a <doctype> for new sites that rejects broken html rather than rearranging it is basically nuts.
|
# ? Apr 29, 2016 04:20 |
|
web "developers" also like javascript. its basically a lost cause
|
# ? Apr 29, 2016 04:21 |
|
eschaton posted:you had me right up until that last clause maybe but actual research languages can get even weirder
|
# ? Apr 29, 2016 04:42 |
|
eschaton posted:please, that's web "devs" the biggest source of problems was ads. putting ads on your page means letting the absolute bottom of the barrel inject arbitrary js (or back then flash) into your website. most ad networks just outright didn't support xhtml at all because they used document.write to inject their ads, and the only way you could actually guarantee that your site would work with ads was if you hand-curated them all, which is practical for approximately 0% of ad-supported websites
|
# ? Apr 29, 2016 05:06 |
|
pokeyman posted:ime this is specified in excruciating detail with test cases already written for you, so if you've genuinely found some unspecified case that's prolly a bug in the spec. it has thankfully been years since i've followed such things so maybe they gave up on this, but early on the goal was to produce a "total" specification that would precisely detail the error recovery such that every stream of bytes read from /dev/random would have an unambiguously correct way to parse it. this would still leave a lot of room for rendering inconsistencies, but it would nearly eliminate inconsistencies due to malformed markup
|
# ? Apr 29, 2016 05:12 |
|
Plorkyeran posted:the biggest source of problems was ads. putting ads on your page means letting the absolute bottom of the barrel inject arbitrary js (or back then flash) into your website. even more reason browser vendors should have just said HTML5 is XML and invalid documents don't render you aren't selling me on your position here ads = cancer
|
# ? Apr 29, 2016 05:12 |
|
it would be very easy to modify an adblocker plugin to just clear the entire page any time an ad is detected if that's what you really want most people prefer to use browsers that can actually render the web, and html5 came around before google successfully took away people's ability to not upgrade
|
# ? Apr 29, 2016 05:16 |
|
CPColin posted:With "Build Automatically" and "Synchronize on Build" both turned on, updated class files get to my remote server in like two seconds after I hit save. Not bad. drat, this the dream right here. From 15 minute turnaround to 2 seconds
|
# ? Apr 29, 2016 09:06 |
|
Terrible programmer status: deleted some useless code as part of some other changes. Everything stopped working and because it was my hobby project I had been lax about committing properly. Now, 3 days later I found the same SO page I'd got it from in the first place readded the actually not useless code, and added a comment pointing out how it's necessary
|
# ? Apr 29, 2016 09:53 |
|
there was a period of time when I served my website with the Content-Type: application/xhtml+xml HTTP header, which forces the browser to fail to render the page and give a parsing error if there's a mistake in your XHTML. I was able to do this because my markup is correct and my commenting system had a hand-written parser which would only accept a carefully-crafted subset of XHTML but there were a bunch of odd browsers which couldn't understand the header and failed to render it anyway what I'm saying is that you live in a world which has fallen short of my glorious perfection
|
# ? Apr 29, 2016 10:18 |
|
watched the jrebel video to the end https://zeroturnaround.com/software/jrebel/#video ~cringe~
|
# ? Apr 29, 2016 10:52 |
|
MALE SHOEGAZE posted:go lives at a weird level of abstraction make loving sure you use godep it won't make the language any better, but at least you won't have to rewrite everything in three months when your dependencies have skewed irreparably out of sync
|
# ? Apr 29, 2016 11:56 |
|
Maluco Marinero posted:Yeah, this is the issue with failing gracefully, is that in the real world it's not graceful at all, SOMETHING depends on that HTML being correctly formed, CSS or Javascript. I wasn't part of the meeting that 'rejected' XHTML, but the fact that we can accept 'use strict' as a thing, but not a <doctype> for new sites that rejects broken html rather than rearranging it is basically nuts. Speaking of HTML meetings, this video makes me mad: https://www.youtube.com/watch?v=Q4dYwEyjZcY
|
# ? Apr 29, 2016 14:59 |
|
HoboMan posted:Speaking of HTML meetings, this video makes me mad: whoa, how can a guy be so wrong?
|
# ? Apr 29, 2016 15:37 |
|
well he's old and white
|
# ? Apr 29, 2016 15:54 |
|
i know ruby and all that but i got bit by awful behavior that i need to complain aboutcode:
|
# ? Apr 29, 2016 17:44 |
|
eschaton posted:please, that's web "devs" what CAD do you mean? VLSI DRC is tuned to be so obnoxious as to require waivers. EE CAD will let you route acute angles that will cause silkscreening issues, or run a ground trace right under a crystal. ME CAD systems have 'designer modes' with unrealizable 2d NURBS, even after being banged into a non-intersecting geometry you still need a DFM review by a tooling expert. i think solidworks still balks at an extruded figure-8's midpoint. they're safety nets, and poor ones that still require human experts to cover the gaps
|
# ? Apr 29, 2016 18:58 |
|
what is the standard procedure for reviewing software design? I've got a class diagram, a sequence diagram, and a state chart that capture the idea well. i'd prefer to hook my laptop up to a projector and present the idea, but coworker wants a document. its a reasonable request, but i don't know what level of detail is needed, or what starting assumptions he may have about what the design is doing that need to be cleared up, or what concepts to focus on, you know, stuff that could be handled naturally by meeting in person.
|
# ? Apr 29, 2016 20:08 |
|
does anyone know of a thing that will check my HTML in a strict way? i can't find anything obviously wrong with it, but it's a huge file.
|
# ? Apr 29, 2016 21:23 |
|
HoboMan posted:does anyone know of a thing that will check my HTML in a strict way? i can't find anything obviously wrong with it, but it's a huge file. The W3C Markup Validator has always worked for me.
|
# ? Apr 29, 2016 21:34 |
|
HoboMan posted:does anyone know of a thing that will check my HTML in a strict way? i can't find anything obviously wrong with it, but it's a huge file. https://html5.validator.nu/
|
# ? Apr 29, 2016 21:34 |
|
Blinkz0rz posted:i know ruby and all that but i got bit by awful behavior that i need to complain about ruby is very much so a fail silently language code:
|
# ? Apr 29, 2016 21:39 |
what do i do in python when i want to read more stuff than i have ram? i.e. from 80 gb file i can read whatever amount i want, but what do i do when that amount happens to be 20gb and im on a pc with 16gb of ram (simplying it but w/e)
|
|
# ? Apr 29, 2016 21:40 |
|
kalstrams posted:what do i do in python when i want to read more stuff than i have ram? i.e. from 80 gb file i can read whatever amount i want, but what do i do when that amount happens to be 20gb and im on a pc with 16gb of ram (simplying it but w/e) you will just use a lot of virtual memory and it will go slower.
|
# ? Apr 29, 2016 21:48 |
Zaxxon posted:you will just use a lot of virtual memory and it will go slower.
|
|
# ? Apr 29, 2016 21:59 |
|
so i have a javascript best practices (lol) question: should i try to do everything with jquery objects rather than just directly using dom objects when i'm messing with the html or should i prefer using dom objects over jquery or does it not matter at all?
|
# ? Apr 29, 2016 22:01 |
|
kalstrams posted:nah, it told me to gently caress off, yanking a MemoryError out, thought that might be feature of numpy or matplotlib, not necessarily the barebones python 2.7.11 or you ran out of address space because you're using a 32 bit version of python
|
# ? Apr 29, 2016 22:01 |
b0lt posted:or you ran out of address space because you're using a 32 bit version of python
|
|
# ? Apr 29, 2016 22:02 |
|
HoboMan posted:so i have a javascript best practices (lol) question: should i try to do everything with jquery objects rather than just directly using dom objects when i'm messing with the html or should i prefer using dom objects over jquery or does it not matter at all? nothing matters hobo man
|
# ? Apr 29, 2016 22:14 |
|
HoboMan posted:so i have a javascript best practices (lol) question: should i try to do everything with jquery objects rather than just directly using dom objects when i'm messing with the html or should i prefer using dom objects over jquery or does it not matter at all? using jquery will make your life easier, and you will become good at using jquery specifically and not so much native javascript as long as you arent doing anything whacky then you probably wont need to worry about the performance hit of using jquery selectors everywhere
|
# ? Apr 29, 2016 22:29 |
|
javascript best practice is to not use it at all
|
# ? Apr 29, 2016 22:37 |
|
i figure jquery is good and has better defined behavior, but my biggest problem is with the .val() method. it don't feel good to use JavaScript code:
JavaScript code:
JavaScript code:
HoboMan fucked around with this message at 23:00 on Apr 29, 2016 |
# ? Apr 29, 2016 22:47 |
|
|
# ? May 25, 2024 12:12 |
|
Blinkz0rz posted:i know ruby and all that but i got bit by awful behavior that i need to complain about ohai.path not everything that has .com in it is a website
|
# ? Apr 29, 2016 22:48 |