» Net Code Options

[Insomniac]

Just out of curiosity, especially since I'm not much of a "Tech Guy" so I am unfamiliar with these sorts of things. Is it possible to alter Marathon's Net Code so that it is easier to host less laggy games? Often times when I host a game with my friend who lives in a nearby city, the game will lag a lot for him. This is strange considering that I have a decent connection and he lives close to me. This happens when I host for other people as well and at times I can even experience lag as the host. This also happens to other people as well when they host. They experience the same thing they tell me.

As a matter of fact, is it possible to host a game where no one experiences major lag?

[thermoplyae]

if this were possible, what possible reason could we have for not doing it already

[Dugit]

Ah, leave him alone. Every newbie makes mistakes like that. I did.

[Xilef]

There's a lot of things that need to be improved with networking, but it's far down the line and will take a while to do so is low on priority list.
The efficiency of quake world is desired, but with a server lobby, dedicated servers would be good, I'd certainly host one, but it's all a distant dream at the moment.

[Treellama]

You could have modern FPS networking, but you'd have to give up films and Lua net scripts. I've played many lag-free games, you just need everyone to have a good connection.

I will add that if people experience lag in every game you host, your connection is not as good as you think.

[thermoplyae]

leave him alone

i'll resist the fat sam image this time, but doo-git is not the minimum standard to which i hold pfhorums members

There's a lot of things that need to be improved with networking, but it's far down the line and will take a while to do so is low on priority list.
The efficiency of quake world is desired, but with a server lobby, dedicated servers would be good, I'd certainly host one, but it's all a distant dream at the moment.

i don't really understand this post. what needs improvement? i've never touched quake world, so what efficiency does it have that we don't? we does efficiency even mean in this context?

[Treellama]

i've never touched quake world, so what efficiency does it have that we don't?

QuakeWorld uses server authoritative dead reckoning. It can provide better latency hiding on terrible connections (it was designed to work on dialup) at the cost of requiring more server bandwidth (so dedicated servers were popular, although most new games coming out don't bother with them).

Basically, clients send location and trigger updates, and the server simply applies them when it gets them (or when it feels like it), and sends world state deltas back to the clients. This avoids having to wait on clients like Aleph One must (since it must replicate the input exactly in lock step to keep the game in sync) but breaks films and Lua scripts, and requires more bandwidth. It also has implications on cheating, depending on the latency hiding model used.

[thermoplyae]

i see. veering away from alephone, was there a reason for the parenthetical about being designed for dial-up? this model is roughly what i consider "modern FPS networking," but i guess i formed that opinion a little under a decade ago. are there models that somehow take advantage of the high bandwidth, middling latency, and persistent connections average users have nowadays?

[Treellama]

i see. veering away from alephone, was there a reason for the parenthetical about being designed for dial-up?

Broadband at that time was being on a University network, and 200+ ms ping times across the US were unheard of except on dial-up. It's probably the best model for modern consumer Internet connections, but only because we're, a decade later, only marginally better at reliability and latency than a dial-up modem. Internet providers have a duopoly, so not much incentive to improve; broadband routers still refuse to adopt common sense congestion control like TCP Vegas, so Joe AlephOnePlayer's siblings' bittorrents kill any games he joins; and Joe AkephOnePlayer is probably leeching Internrt off someone else anyway rather than paying for it, so it's not like he could pick a better ISP even if he knew how, so we're left answering questions like OP's.

[ellio7t]

Just out of curiosity, especially since I'm not much of a "Tech Guy" so I am unfamiliar with these sorts of things. Is it possible to alter Marathon's Net Code so that it is easier to host less laggy games? Often times when I host a game with my friend who lives in a nearby city, the game will lag a lot for him. This is strange considering that I have a decent connection and he lives close to me. This happens when I host for other people as well and at times I can even experience lag as the host. This also happens to other people as well when they host. They experience the same thing they tell me.

As a matter of fact, is it possible to host a game where no one experiences major lag?

direct server connection via ethernet.

[Dugit]

That'll do the job. I remember doing that back when I had an OS9.

[Insomniac]

Ah, leave him alone. Every newbie makes mistakes like that. I did.

What "Mistake" did I make exactly? Asking a question?

[thermoplyae]

asking a question that betrays a total vacuum of forethought, yes

[Dugit]

Everybody has done it.

[EwoudCP]

The guy asked a question, already saying that he didn't know the ins and outs of the issue. This is not a mistake. This is a question.

If I ask my friend what time it is, is that a mistake?
If someone I know asks me something about computers, is that a mistake?
No, it is only a mistake when someone tells me about computers when they don't know what they're talking about.

There's a reason people ask questions, and in this case, it is fully justified, considering that this is not a simple google-able matter.

This post has been edited by EwoudCP: Dec 30 2009, 03:34 PM

[Dugit]

The guy asked a question, already saying that he didn't know the ins and outs of the issue. This is not a mistake. This is a question.

If I ask my friend what time it is, is that a mistake?
If someone I know asks me something about computers, is that a mistake?
No, it is only a mistake when someone tells me about computers when they don't know what they're talking about.

There's a reason people ask questions, and in this case, it is fully justified, considering that this is not a simple google-able matter.

Hmm... yes, but Thermo did have a point in that, although it was a genuinely innocent question, for which I understand and empathize with, he is right that if it was possible, it would have been done before. I'm not angry or even mildy miffed at him (Insomniac); it's a simple mistake to make. For such a topic as computers or the time, which needs constant reassertion of the information, this topic is based on a more or less stationary objective, and thus, if this was such a question, it must surely be a repeat.

Vale,
Dugit

[EwoudCP]

Hmm... yes, but Thermo did have a point in that, although it was a genuinely innocent question, for which I understand and empathize with, he is right that if it was possible, it would have been done before. I'm not angry or even mildy miffed at him (Insomniac); it's a simple mistake to make. For such a topic as computers or the time, which needs constant reassertion of the information, this topic is based on a more or less stationary objective, and thus, if this was such a question, it must surely be a repeat.

Vale,
Dugit

Searching on these Pfhorums about netcode doesn't come up with useful results.

But true, if it could've been done, it would have. But it is only really clear in retrospect.
Marathon seems like a somewhat simple game environment, and the assumption that its netcode could be enhanced to match other, newer and often more complex games is not unreasonable.
The idea that the netcode is not optimal is not strange: I experience more lag playing Marathon with people than I do with many newer games - and I'm not the only one.
Considering these factors, it is only logical to come to the conclusion that the netcode might not be as optimal as it could be.

But then again, there are few Aleph One games to choose from, while when I fire up TF2, many servers, including low-latency ones, present themselves. That, and A1 has a very different basis for its netplay than modern games, as was said before.

[Dugit]

Searching on these Pfhorums about netcode doesn't come up with useful results.

But true, if it could've been done, it would have. But it is only really clear in retrospect.
Marathon seems like a somewhat simple game environment, and the assumption that its netcode could be enhanced to match other, newer and often more complex games is not unreasonable.
The idea that the netcode is not optimal is not strange: I experience more lag playing Marathon with people than I do with many newer games - and I'm not the only one.
Considering these factors, it is only logical to come to the conclusion that the netcode might not be as optimal as it could be.

But then again, there are few Aleph One games to choose from, while when I fire up TF2, many servers, including low-latency ones, present themselves. That, and A1 has a very different basis for its netplay than modern games, as was said before.

Yes, it is only clear in retrospect, but if you've been here for longer, you'd probably think in that sense before you ask a question, to avoid such attacks.
Anyway, if someone wants a game with less lag, as is aforementioned, it is not the game, it is the connections taking part in the game. One bad internet connection stirs the entire game awry with lag.

[thermoplyae]

you don't have to actually search anywhere to come to this conclusion -- that was my point. if you'd imagined yourself as the recipient of the question "you know that product you guys handle? why don't you make it generally better?" it would have been obvious straight away that this is either an inherently terrible question or at least that it's phrased improperly. xilef at least took a step toward a useful rephrasing by comparing the situation in alephone to that of other publicly developed games (presumably because he's not so technically-inclined as to be able to formulate his question in the abstract -- and that's fine!). you should note that his post generated not only a reply from a developer but also a brief discussion about networking in multiplayer games in general.

yes, everyone's guilty of it. that doesn't mean you shouldn't be harshly scolded.

[Dugit]

yes, everyone's guilty of it. that doesn't mean you shouldn't be harshly scolded.

No. He shouldn't be harshly scolded for an innocent mistake.

This post has been edited by Dugit: Dec 30 2009, 07:59 PM