The ESBESB
-
@blakeyrat said:
@morbiuswilters said:
A lot? Binary protocols are fucking hell to extend, debug and implement.
Some dude at Microsoft's problem, not mine.
I know you work for a development platform stuck in 1976, but here at the "bleeding edge" we have a guy to do shit like that so we don't have to.
It'll be your problem as soon as there are 15 subtly different implementations of HTTP/2.0 that require 14 different meta tags on each page to work around bugs.
-
@Ben L. said:
It'll be your problem as soon as there are 15 subtly different implementations of HTTP/2.0 that require 14 different meta tags on each page to work around bugs.
Nope. Still won't.
-
@blakeyrat said:
Some dude at Microsoft's problem, not mine.
It's still your problem when Microsoft (and Apple, and Google, and Mozilla, etc..) fuck it up and you have to deal with the bugs. Or when somebody is like "Hey, your shit doesn't work in Firefox. Fix it, bitch!" and then you have to figure out why Firefox isn't picking up your headers correctly.
-
@blakeyrat said:
@Ben L. said:
It'll be your problem as soon as there are 15 subtly different implementations of HTTP/2.0 that require 14 different meta tags on each page to work around bugs.
Nope. Still won't.
Why not?
-
Two reasons:
1) This is an Internet protocol, so it will be in a "finished but not finished" state for approximately 47 years, giving Microsoft and web server makers plenty of time to sort all the issues out.
2) When it does finally get sorted out, I can be sure Microsoft's implementation will work because nobody would dare to make a web server where it didn't, for fear of losing all their users.
-
@blakeyrat said:
Two reasons:
1) This is an Internet protocol, so it will be in a "finished but not finished" state for approximately 47 years, giving
Microsoft andweb server makers plenty of time to sort all the issues out. Microsoft will release their implementation of HTTP/2.0 with Internet Explorer 92.1 service pack 3, ten years after the heat death of the universe.2) When it does finally get sorted out, I can be sure Microsoft's implementation will work but nobody will care because everything else will be using HTTP/5.1 and Internet Explorer will have exactly three users.
because nobody would dare to make a web server where it didn't, for fear of losing all their users.
-
@blakeyrat said:
2) When it does finally get sorted out, I can be sure Microsoft's implementation will work because nobody would dare to make a web server where it didn't, for fear of losing all their users.
Sure, it'll work, but it will require lots of workarounds and other assorted pain. And your attitude of "Well, I don't care because I don't have to deal with it" is pure bullshit. Even if you don't, you're just inflicting more pain on the people who do have to work with it, including that team at Microsoft (not "some dude"..) which is going to be cursing the IETF's name if this ends up a binary protocol.
If you truly don't care, then you should shut the fuck up and listen to people who do care because they know more than you and they actually have to deal with it.
-
@morbiuswilters said:
Sure, it'll work, but it will require lots of workarounds and other assorted pain.
My point is I don't care because I write software using libraries written by other people whose job it is to make it work. I never experience the pain, so I do not give a shit if it's painful or not. The same is true of all (rational) software developers. (Except the tiny portion who actually do have to implement the protocol, but hey, that's THEIR JOB to cope with it.)
If you're not the ONE GUY PER LANGUAGE implementing the standard library, and you have to care about this for even one iota of a second... then you fucked up somewhere.
@morbiuswilters said:
If you truly don't care, then you should shut the fuck up and listen to people who do care because they know more than you and they actually have to deal with it.
Right now I haven't heard any reason for it being inferior other than vague handwaving and weasel words. Unlike the rest of the IT community, I'm not afraid of change, and I'm not afraid of questioning assumptions of the past. (Assumptions like: protocols should be text-based.)
So bring it on.
-
@morbiuswilters said:
Next up: let's replace HTML with ASN.1!
Eeeeeeee…<nervous tic >
Not even in jest, please, not even in jest.
Formally speaking, it would be actually SGML and XML that they'd be replacing. HTML would be implemented using an imperial crapton of OIDs in some “namespace” that would rapidly become horribly familiar. We'd have the horrible mess that is X.509 except much much worse, and there's more than enough “binary is great and efficient and easy” weenies about for them to try this sort of thing on and we'd get stuck with the shit caused by it. Even reading the various OWL/RDF serializations — themselves only for people who think that XQuery is too nice — is nicer than this.
-
@blakeyrat said:
If you're not the ONE GUY PER LANGUAGE implementing the standard library
You wouldn't believe how often it seems that, of all the guys to pick to implement the HTTP client library, some language has got stuck with the wrong person. Catastrophic bugs, mysterious failures and… whimsical design choices seem to be the norm.I tell a lie, you would believe it. You're here.
-
@blakeyrat said:
My point is I don't care because I write software using libraries written by other people whose job it is to make it work. I never experience the pain, so I do not give a shit if it's painful or not. The same is true of all (rational) software developers. (Except the tiny portion who actually do have to implement the protocol, but hey, that's THEIR JOB to cope with it.)
You are a goddamn moron. Hey, here's an idea: the W3C should drop Javascript and require all browsers to use Go for client-side scripting. Not my problem, so why should I care, right?
@blakeyrat said:
If you're not the ONE GUY PER LANGUAGE implementing the standard library, and you have to care about this for even one iota of a second... then you fucked up somewhere.
Once again: you are retarded. Have you ever worked in the fucking software industry? Do you think a single guy implements a new technology in most languages? Do you think the IE team has a single guy who works on HTML5 and nobody else at all touches it? I have to debug this shit all the fucking time. And, yeah, it's frequently involving Microsoft browsers or Microsoft servers, you ignorant piece of shit. Why don't you shut your fucking mouth when the adults are talking? If we start discussing button alignment or the color of fonts, we'll ask your opinion.
@blakeyrat said:
(Assumptions like: protocols should be text-based.)
There are plenty of binary protocols out there, and they tend towards the shit-tastic. Look, I know you're delusional enough to think you're smarter than the entire rest of the software industry, but why do you think virtually everyone has moved away from storing or transmitting structured data in binary formats?
-
@dkf said:
@morbiuswilters said:
Next up: let's replace HTML with ASN.1!
Eeeeeeee…<nervous tic >
Not even in jest, please, not even in jest.
Fucking idiots like Blakeyrat would probably be on board because, hey, he's not afraid of change!
@dkf said:
Formally speaking, it would be actually SGML and XML that they'd be replacing.
It already exists. And I see no reason why they'd need to implement all of SGML; just HTML would be fine.
-
@morbiuswilters said:
You are a goddamn moron. Hey, here's an idea: the W3C should drop Javascript and require all browsers to use Go for client-side scripting. Not my problem, so why should I care, right?
How is that even remotely similar to what we're talking about?
For the record, the answer is: because that would break millions of websites. Changing the way HTTP works breaks maybe 30 products, max. Assuming it breaks them at all, which it wouldn't.
@morbiuswilters said:
Have you ever worked in the fucking software industry?
Yes.
@morbiuswilters said:
Do you think a single guy implements a new technology in most languages?
Now you're being a pedantic dickweed? Fuck.
HEY PEDANTIC DICKWEEDS: I DIDN'T MEAN LITERALLY ONE PERSON JESUS CHRIST PLEASE DIE NOW
@morbiuswilters said:
And, yeah, it's frequently involving Microsoft browsers or Microsoft servers, you ignorant piece of shit.
Ok, praytell, what is wrong with Microsoft's HTTP implementation?
@morbiuswilters said:
Look, I know you're delusional enough to think you're smarter than the entire rest of the software industry,
I don't think that. I'm just not afraid of change or questioning assumptions, like you so obviously are.
@morbiuswilters said:
but why do you think virtually everyone has moved away from storing or transmitting structured data in binary formats?
"Moved away from"?
I can think of exactly one case of this: Microsoft switching the Office file formats to be XML-based.
All other systems I can think of either still store binary formats, or never stored binary formats in the first place.
-
@blakeyrat said:
How is that even remotely similar to what we're talking about?
For the record, the answer is: because that would break millions of websites. Changing the way HTTP works breaks maybe 30 products, max. Assuming it breaks them at all, which it wouldn't.So apparently the stupidity of a move is judged based on how many people would have to change their behavior? Guess that means Win8 is a failure.. now who's afraid of change?
Of course, that's not even relevant since W3C moving to Go wouldn't break shit--browsers would still support Javascript, but would also have to support Go. And it would be a disaster to develop in, because of problems inherent to Go, not because people would have to learn something new or change.
@blakeyrat said:
Ok, praytell, what is wrong with Microsoft's HTTP implementation?
Here's one: clicky. That took about 3s of Googling. And I can guarantee you there are more.
@blakeyrat said:
"Moved away from"?
I can think of exactly one case of this: Microsoft switching the Office file formats to be XML-based.
All other systems I can think of either still store binary formats, or never stored binary formats in the first place.
Are you trying to demonstrate how dumb you can be? Is that the theme for this thread? I never said "Hey, products are converting long-established formats to text-based" because that's a rather expensive and difficult undertaking. (Of course, Microsoft did it for Office, and I doubt anyone there is missing the old binary formats.)
However, when it comes to new development, practically every project starts with text-based formats. Why do you think XML and JSON are so ubiquitous. Fuck, this entire conversation started because Minecraft is moving away from a binary protocol for their chat messages.
Hey, why don't we start returning binary-encoded data from our AJAX calls??
-
@morbiuswilters said:
So apparently the stupidity of a move is judged based on how many people would have to change their behavior?
That's not what I said.
@morbiuswilters said:
Here's one: clicky. That took about 3s of Googling. And I can guarantee you there are more.
Oh well the one problem in an obsolete browser 6 years ago-- YOU'VE CONVINCED ME!
@morbiuswilters said:
Are you trying to demonstrate how dumb you can be?
No; I'm trying to think of cases where a program has switched from a binary format to a text-based one.
@morbiuswilters said:
Hey, why don't we start returning binary-encoded data from our AJAX calls??
Sure. Why not.
Hey Morbs, calm the fuck down. Here's a list of things I haven't said:
1) HTTP should switch to being binary-based.Get it? You're arguing against something I never said. Stop it. It's fucking annoying as hell, and you're fucking abrasive as hell.
All I said is I don't necessarily see that move as being a negative one. If you have a GOOD REASON for it being a negative move, then by all means: let me know what it is. You haven't done that. All you've done is given vague scenarios unrelated to reality. Which is why I assume the actual reason is: "I, MorbiusWilters, am deathly afraid of change." If that's not the reason, and I'm incorrect, then why not explain to me what the actual reason is? Because usually people who give generic vague warnings full of weasel-words are just afraid of change.
But don't sit there and yell and bitch at me as if I said, "HEY EVERYBODY HTTP SHOULD BE BINARY LOLOL" because I never fucking said that and fuck you.
In fact why don't you just go away.
-
@morbiuswilters said:
@immibis said:
(Somewhat related, just when you thought Minecraft couldn't get any worse, it now encodes chat messages in JSON)
That doesn't seem so bad. Text-based protocols are a lot easier to work with than binary ones
JSON is overkill...
The main advantage of JSON I can see is that if you see it on a packet sniffer you can see the text [i]and formatting[/i] of the message, instead of just the text. Oh wait, you can't, because it's encrypted.
Maybe that, being a standard format, you can just plug in an existing library? Nope, it would take about as long to serialize/deserialize messages to a simple binary protocol as it would to JSON.
Maybe that it's compact? Nope, there's already a Minecraft-specific binary format similar in structure to JSON. And anything that handles these chat messages, including Minecraft itself, is likely to already support that format.
At least it has a length field. Otherwise you'd have to parse it just to find the start of the next message.
Edit: WTF happened to my line breaks
[mod - fixed formatting - PJH]
-
@immibis said:
WTF happened to my line breaks
You should've included a line length parameter in each line. Or used a human-readable in-band signaling technique like this: “<p>”
-
@immibis said:
Edit: WTF happened to my line breaks
I have to manually add <br> tags to my posts. I don't deny that this is bullshit.
