If what you're saying is that you'll accept a solution that does half telopt, then I'm not sure you've quite registered the overwhelming opinion against telopt for this articulated here and elsewhere.
The main concern has been that many muds can't readily implement a telnet handler. My compromise would solve that. I've heard no notable rational argument against doing this in telnet, in fact several people have acknowledged that ideally telnet is the correct way to do this.
So MSSP-REQUEST - yes, that can be done.
variable:value\r\n - no, that is not an option because it has no added value and is very divisive.
18 Mar, 2009, David Haley wrote in the 102nd comment:
Votes: 0
Scandum said:
in fact several people have acknowledged that ideally telnet is the correct way to do this.
Why?
Scandum said:
variable:value\r\n - no, that is not an option because it has no added value and is very divisive.
Divisiveness isn't a problem until a protocol has actually solidified. We're trying to build consensus here; invariably that means people will have different ideas. Pointing fingers at people and blaming them of being "divisive" for suggesting alternatives is not productive at all and only makes it sound like you're not interested in anything other than what you already think.
18 Mar, 2009, Cratylus wrote in the 103rd comment:
Votes: 0
Scandum said:
Cratylus said:
If what you're saying is that you'll accept a solution that does half telopt, then I'm not sure you've quite registered the overwhelming opinion against telopt for this articulated here and elsewhere.
The main concern has been that many muds can't readily implement a telnet handler. My compromise would solve that. I've heard no notable rational argument against doing this in telnet, in fact several people have acknowledged that ideally telnet is the correct way to do this.
So MSSP-REQUEST - yes, that can be done.
variable:value\r\n - no, that is not an option because it has no added value and is very divisive.
Well…you didn't answer my question.
Plaintext query sent to the mud like a normal login: yes Plaintext response printed to the client like normal game text: yes
Right?
I mean, otherwise you'd be saying you're ok with me only eating half the poison apple.
A plain text response would exclude the usage of telnet. So for both command line and telnet to work along side each other the data would have to be send as IAC SB MSSP <data> IAC SE.
in fact several people have acknowledged that ideally telnet is the correct way to do this.
Why?
More important, where?
18 Mar, 2009, Cratylus wrote in the 106th comment:
Votes: 0
Scandum said:
A plain text response would exclude the usage of telnet. So for both command line and telnet to work along side each other the data would have to be send as IAC SB MSSP <data> IAC SE.
I'm not sure why you'd think that using telopt for half of this would change anything.
Please unconfuse me.
Did you think that the point is to make you unhappy, and that you being half-unhappy is adequate compromise?
I'm not sure why you'd think that using telopt for half of this would change anything.
It would allow implementation without messing with the mud lib, correct?
And your point seems to be to completely invalidate the usage of telopts, which I don't quite understand.
18 Mar, 2009, elanthis wrote in the 108th comment:
Votes: 0
If you go with a plain-text protocol, you really do need to have the server advertise support first, e.g. by sending an MSSP-SUPPORTED\r\n line (probably before the banner and/or screen clear so users don't see it). Scandum hit the nail on the head about it being annoying for some MUDs to get a crawler that connects, starts trying to "login" with MSSP-REQUEST, and then disconnects. Some MUDs log that kind of activity for security or debugging purposes… and if they don't support MSSP, they're not going to have code in place to ignore the MSSP-REQUEST "login."
For what it's worth, my reasoning against using TELNET subnegotiation was based on the assumption that since most servers and clients can't actually handle it worth crap (including Scandum's own code) that using it would be dangerous and reduce the chance that anyone uses it. However, it is clear that several crawlers already do support it without problems, so I admit to being wrong. My one and only real complaint with MSSP at this point is the use-last-value semantics, which I still feel are counter-productive and potentially dangerous. It could also just be implemented as a ZMP command and save the need for a whole second telopt code and subrequest handler, but that's a very very minor point not worth arguing since Scandum clearly disagrees.
18 Mar, 2009, Cratylus wrote in the 109th comment:
Votes: 0
Scandum said:
It would allow implementation without messing with the mud lib, correct?
I have to admit to you that this is getting a little scary. I'm not sure now you've been here for the entire discussion.
Scandum said:
And your point seems to be to completely invalidate the usage of telopts, which I don't quite understand.
I think you must use telopts where they are appropriate. I do not think that for the purpose stated weeks ago in your original post, telopts are the right tool.
If you go with a plain-text protocol, you really do need to have the server advertise support first, e.g. by sending an MSSP-SUPPORTED\r\n line (probably before the banner and/or screen clear so users don't see it). Scandum hit the nail on the head about it being annoying for some MUDs to get a crawler that connects, starts trying to "login" with MSSP-REQUEST, and then disconnects. Some MUDs log that kind of activity for security or debugging purposes… and if they don't support MSSP, they're not going to have code in place to ignore the MSSP-REQUEST "login."
Wait. Do you guys really want to put these crawlers similarly to how google scans the web? If so, may I ask how? Trying to telnet to random ports at random IPs in search for muds? Baaaah. I thought that's more of a convenience, where you register your mud with a mssp listing site once, and it refreshes the data automatically… and in that case, you wouldn't registe if you wouldn't support it, right?
18 Mar, 2009, David Haley wrote in the 111th comment:
Votes: 0
wrkq said:
Wait. Do you guys really want to put these crawlers similarly to how google scans the web? If so, may I ask how? Trying to telnet to random ports at random IPs in search for muds? Baaaah. I thought that's more of a convenience, where you register your mud with a mssp listing site once, and it refreshes the data automatically… and in that case, you wouldn't registe if you wouldn't support it, right?
Yes, this solves the "annoying login" problem. Besides, MUDs would still be seeing connection attempts even with telnet subnegs… it's just that those wouldn't trigger a "bad password" message or something.
I think you must use telopts where they are appropriate. I do not think that for the purpose stated weeks ago in your original post, telopts are the right tool.
Then I suggest you find a good name for your protocol and stop calling it MSSP, that's just gonna confuse everyone and it makes it clear you're not talking about MSSP but something else entirely. MSSP as currently defined is beta since it has long passed the point where things are just theoretical, and you're clearly not willing to work on a compromise.
18 Mar, 2009, Cratylus wrote in the 113th comment:
Votes: 0
Scandum said:
MSSP as currently defined is beta since it has long passed the point where things are just theoretical, and you're clearly not willing to work on a compromise.
So let me get this straight. Because you refused to listen to people for long enough, you can now claim that changes you don't like are no longer possible?
Your post on TMC makes it sound like I'd be mistaken to think of this as *your* protocol. Sure sounds like it is to me, though.
-Crat
18 Mar, 2009, David Haley wrote in the 114th comment:
Votes: 0
It would tend to agree that it's not entirely fair play to put out an RFC, wait just a few days, then go write some software and hold it up as a golden standard and accuse other people of being unwilling to compromise.
There are already two crawlers and several muds supporting MSSP, so there.
18 Mar, 2009, Cratylus wrote in the 116th comment:
Votes: 0
Scandum said:
There are already two crawlers and several muds supporting MSSP, so there.
So then the protocol is complete? Or did your rush it? RFC closed? Open? What?
Surely you see what it looks like from the outside.
-Crat
18 Mar, 2009, David Haley wrote in the 117th comment:
Votes: 0
I can whip up a protocol, stick it into a few MUDs, submit some patches, get Crat to stick it in his MUD too, and convince MB to have some toy example using it. Does that make it official…?
There are already two crawlers and several muds supporting MSSP, so there.
While I am one of the MUDs, and codebases, supporting this as currently written: it's not a final product. Many things are rushed to market prior to being finalized – WiFi standards, anybody? I remember using a few A or N standard pieces of gear before things were agreed upon 100%, yet some manufacturers went ahead and put hardware to market anyhow.
18 Mar, 2009, Cratylus wrote in the 119th comment:
Votes: 0
Scandum said:
Then I suggest you find a good name for your protocol
Unless Scandum comes out and says that this really is his protocol and I can go get stuffed, I'm going to operate under the good faith presumption that this is a community effort and I will continue to provide my input.
Please understand Scandum, I am not trying to be difficult.
Well, the current discussion isn't going anywhere. There's no serious constructive debate going on about the variables or the actual telnet negotiations. Given how transparent and unobtrusive the telnet implementation is I don't see any reason for mudlist crawlers not to implement it.
This is yet another thread that has turned into off topic trolling and rambling, so I'll have to drop out of the debate and make a very specific announcement thread because dealing with this crap is a total waste of my time and energy.
does half telopt, then I'm not sure you've quite registered the
overwhelming opinion against telopt for this articulated
here and elsewhere.
The main concern has been that many muds can't readily implement a telnet handler. My compromise would solve that. I've heard no notable rational argument against doing this in telnet, in fact several people have acknowledged that ideally telnet is the correct way to do this.
So MSSP-REQUEST - yes, that can be done.
variable:value\r\n - no, that is not an option because it has no added value and is very divisive.