26 Sep, 2010, Runter wrote in the 21st comment:
Votes: 0
In practice you can gag noise from specific muds.
27 Sep, 2010, Scandum wrote in the 22nd comment:
Votes: 0
I don't think decentralization is going to work due to the added complexity.

One option is for the MUD to provide a chat server address and have the client use Mud Master chat to connect.
27 Sep, 2010, Runter wrote in the 23rd comment:
Votes: 0
I too can post tragically useless comments that add nothing but misdirection to legitimate discussions.
27 Sep, 2010, quixadhal wrote in the 24th comment:
Votes: 0
It boils down to how much do you value control/security/authentication vs. how much you value simplicity.

Public keys (which are NOT all that complex guys), give you a reasonable idea that the person or server you are talking to is, in fact, the one you THINK you're talking to. By themselves, they don't prevent anyone from making new accounts/servers to work around bans, but with a bit of cooperation (yeah, LOLZ), they could help minimize it.

I've heard people here complain about bans being ineffectual because people can just go make new characters, or make new characters on different muds, to get around them. So, I offer one suggestion to help cut down on that. No like? No problem!

As for the peer-to-peer network itself. How much control do you want? A centralized network has control, but you have to hope the hubs doing the controlling share your sets of values. A decentralized network lacks control, because one person's abuse is another person's humor. If the abused decides to not only stop accepting packets, but to stop forwarding them as well, the network has to be more fully connected to avoid unintended fragmentation.

Peer or not, any given MUD can *always* choose to filter packets itself and ignore any it doesn't like. The question is routing. In the current scheme, all traffic goes to one point and is then rebroadcast to all other nodes (a star topology). As long as that one point is trusted and stable, it works fine. In a peer network, traffic is sent to your nearest neighbors, who are expected to forward it along to their neighbors. If one node rejects or fails to forward it, others are expected to do so to maintain connectivity. The problem is that chokepoints can develop where some nodes are only connected to a few neighbors. In that case, those chokepoint routes can effective block large segments of the overall network.

That is a problem one has to consider if you're thinking about a decentralized network. How do you minimize the effects of these chokepoints, without turning it into a fully connected graph?
27 Sep, 2010, David Haley wrote in the 25th comment:
Votes: 0
Quote
By themselves, they don't prevent anyone from making new accounts/servers to work around bans, but with a bit of cooperation (yeah, LOLZ), they could help minimize it.

I've heard people here complain about bans being ineffectual because people can just go make new characters, or make new characters on different muds, to get around them. So, I offer one suggestion to help cut down on that. No like? No problem!

Didn't you just argue that your solution can be worked around, with some "LOLZ"? I don't see what exactly you're proposing as a practical solution here.

Quote
In the current scheme, all traffic goes to one point and is then rebroadcast to all other nodes (a star topology).

Crat has already described that LPMuds i3 doesn't work this way; you can have several routers that agree to be peers to each other. MUDs are still connected to a single router, but the whole network is not a single-hub star.
27 Sep, 2010, Rudha wrote in the 26th comment:
Votes: 0
You know, this is going to come off as a somewhat crass point, but if it's 'just the admin' as some people suggest, why not just write one's own server and create their own network? :P

Granted: It would be nice if multiple IMC servers could connect and exchange network information of sorts, and forward messages etc, thereby allowing them to be routers.

Maya/Rudha
27 Sep, 2010, chrisd wrote in the 27th comment:
Votes: 0
XMPP
27 Sep, 2010, David Haley wrote in the 28th comment:
Votes: 0
Quote
You know, this is going to come off as a somewhat crass point, but if it's 'just the admin' as some people suggest, why not just write one's own server and create their own network? :P

That's exactly what already happened… that is why many people think all this talk of new protocols etc. is rather moot.

Quote
Granted: It would be nice if multiple IMC servers could connect and exchange network information of sorts, and forward messages etc, thereby allowing them to be routers.

That also has already happened. :smile:



See also: http://www.lpmuds.net/intermud.html
Logs available here: http://www.mudworld.net/imud/logs
27 Sep, 2010, Cratylus wrote in the 29th comment:
Votes: 0
quixadhal said:
…Public keys…


lol


Rudha said:
You know, this is going to come off as a somewhat crass point, but if it's 'just the admin' as some people suggest, why not just write one's own server and create their own network? :P

Granted: It would be nice if multiple IMC servers could connect and exchange network information of sorts, and forward messages etc, thereby allowing them to be routers.

Maya/Rudha


It actually works that way now!

So look, I totally dig avoiding a single point of failure. It's practically my middle name. And for that
reason, I'm generally disinclined to disagree with decentralization. This isn't just theoretical for me,
either. I was personal witness to the tragic decline and fall of the once mighty *gjs router.

/moment of silence

While I may-yet-be-convinced, I don't really see that a fully-distributed p2p scheme is worth the effort.
That was done, it's the ancestor of Intermud-3, and it sucked then, and I'm guessing it sucks now. You
can join a p2p intermud network *right now* if you want, and then you can be happy. Do a google
for Zebedee, get it installed on yer mud, and enjoy the freedom, homies.

On the other hand, just as there is a place in this world for wikis, there is also a place in this
world for forums, and having a network for mud communication that has enforced rules of conduct
appears to be the preferred way of doing things. Why? Who cares? I'm not a psychologist. All I know
is that if you want to maximize intermud participation, a network with admins is the way.

So, I'm all for avoiding single point of failure, and to that end, I've coded a peering server system
that allows for federation. If one admin goes berserk, he can be avoided with little damage.

I think that for the purposes of something like intermud communication, a whole-hog p2p system
with crypto bells and whistles is not only overkill, but, pace Runter, basically something with
low likelihood of wide adoption. May the gods of mudding have mercy on me for agreeing with
Scandum on something.

We already have something with wide adoption. I'm not seeing much point in coding something
that already exists which has a lower chance of adoption than the thing we can just work to fix,
rather than drop.

-Crat
http://lpmuds.net
27 Sep, 2010, quixadhal wrote in the 30th comment:
Votes: 0
David Haley said:
Didn't you just argue that your solution can be worked around, with some "LOLZ"? I don't see what exactly you're proposing as a practical solution here.


Why do you hate me, David? Not everyone can be as annoyingly perfect as you'd like them to be. Nor does a solution being imperfect make it NOT A SOLUTION. It can be worked around, but it's better than NOTHING, which is what we have now.

But, hey… nobody has ever taken MUD security seriously in the past, so why start now? With telnet out of the default desktops of the majority of the public, it's not like anyone will distrub the dozen or so of us who are still here. :)
27 Sep, 2010, David Haley wrote in the 31st comment:
Votes: 0
Quote
Nor does a solution being imperfect make it NOT A SOLUTION.

Sure. But make it imperfect enough, and one wonders why it's even being considered as a 'solution'. In other words, the question is: what problem are you trying to solve and why do we care about solving it given all the trouble the 'solution' will cause us?

Something is not always better than nothing, simply because it costs time and effort (now and later) to get 'something' where that something might not be actually giving us much benefit (now and later). In other words, the total cost of getting something might actually exceed the total cost of doing nothing about it.
27 Sep, 2010, Scandum wrote in the 32nd comment:
Votes: 0
Cratylus said:
We already have something with wide adoption. I'm not seeing much point in coding something
that already exists which has a lower chance of adoption than the thing we can just work to fix,
rather than drop.

The most wide adoption is MMCP, which is a decentralized friend to friend protocol. I wonder how hard it would be for a mud server to function as a MMCP node or server. It's also an option to extend the protocol if needed. The main advantage is that MUDs can use MSSP to suggest their chat server of choice, making implementation very easy. It's also an option to run mcl, tintin++, or another MMCP capable console client as a permanent service.
27 Sep, 2010, David Haley wrote in the 33rd comment:
Votes: 0
Can we please stop going down rabbit holes?
27 Sep, 2010, quixadhal wrote in the 34th comment:
Votes: 0
David Haley said:
Can we please stop going down rabbit holes?

My brother's cousin's friend's father's mistress's dealer's kid said that rabbits like to hoard shiny things, and so there might be GOLD at the bottom of their warrens!!!! GOLD, I tell ya!!!!
28 Sep, 2010, Runter wrote in the 35th comment:
Votes: 0
I propose we solve the problem with a grid of hash pointers.
28 Sep, 2010, Rudha wrote in the 36th comment:
Votes: 0
Runter said:
I propose we solve the problem with a grid of hash pointers.


Are you suggesting that hash pointers migrate?

Maya/Rudha
28 Sep, 2010, quixadhal wrote in the 37th comment:
Votes: 0
African or European pointers?
28 Sep, 2010, Scandum wrote in the 38th comment:
Votes: 0
Rudha said:
Are you suggesting that hash pointers migrate?

They're trying to bully me, their references are lame personal attacks aimed at my person that are nonsense to newcomers. These kids are all from the imc2 network, and chatting with each other has clearly instilled some kind of group identity in them. I think they feel pretty threatened by me because they always come after me with at least four persons.

If you read between the lines I think their subconscious message is that they don't like my idea of a network that makes MUDs (I guess in an essence their MUDs), as well as their beloved IMC2, completely irrelevant.

They've been laying low as of late, but IMC2 is ran by the MudBytes owners, so I guess they feel it's safe to be assholes in a topic where they feel I'm attacking IMC2 and by virtue, the MudBytes admin.
28 Sep, 2010, Rudha wrote in the 39th comment:
Votes: 0
Wait, … what? Just … what?

Maya/Rudha
28 Sep, 2010, David Haley wrote in the 40th comment:
Votes: 0
To be clear, all of our involvement here has been with respect to the LPMuds intermud network; the MB admins might or might not do anything similar. Although obviously it would be preferable if there were coordination, our plans will proceed with or without MB's involvement/blessing.
20.0/48