| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11
Comments:
<0> oh, you have a stolen copy too, wannerful
<1> yeah, probably. It's an authentic disc, but I wouldn't be surprised if they ripped it off from work or something
<1> I should have gotten a copy of visual studio for free when I was enrolled in a comp sci cl*** a few semesters ago.
<0> vc2005 is more recent (released in November) and a better compiler
<1> ahh, yeah, i'll go with that then
<1> thanks for the help
<0> "That's what we're here for" 6 Victor A. Wagner Jr.
<2> http://www.local6.com/news/7542555/detail.html
<2> idiots
<2> We know he's disturbing people in the theater
<2> But, that's not going to stop us from going to the theater
<2> we couldn't possibly rent movies.
<3> He couldn't?
<2> yes, clearly they could
<2> that's why they're idiots
<3> Yes, clearly he could
<3> That's why his mom's an idiot
<3> Doesn't seem like there's any easy answer there, if he's really being disruptive.
<3> Maybe the mom could have worked out some arrangement with the theater ahead of time
<0> yeah, sometimes seeing it on the "big screen" makes a huge difference
<3> I sympathize with the kid, but this was the mother's problem to solve, and she didn't solve it.
<0> correct
<4> ahh, just got out of an interview
<4> that was a great one
<4> first guy to really have me try to solve the problems I didn't know
<4> and I think I did really well with all
<0> how do you think you did
<3> What were some of the problems
<0> nm
<3> And congratulations
<4> there were a couple of spots i felt like i was just spinning in circles
<5> You failed, rdragon, I was the interviewer
<4> haha
<4> well they are doing a multiplayer game, so a lot of it was coming up with ways to optimize stuff sent over the network
<4> and how to sync time
<4> and interpolating positions of characters
<0> sync is a really huge problem
<5> Nah, sync is easy
<4> nah, it's not that bad
<5> Hell, I'm giving a conference about part of that these 3 days
<4> i mean, getting it precise is hard ,but 'close enough' is good enough for games
<4> the client sends a time request, the server responds with the time, and the client adds half the roundtrip time to the time reported back, and ***umes that's the 'current' time on the server
<5> That's ugly
<5> You gotta keep stats and such
<4> *shrug* it's simple
<5> Simple doesn't always mean good ;)
<4> I think it's pretty good, it's certainly not terrible or bad
<4> the client->server time is usually longer than the server->client, so maybe you can throw in some bias
<5> Why would client -> server be longer
<4> because usually the client upstream is slower than it's downstream, and the server could potentially hold on to that request for awhile before responding to it (if it's busy doing other stuff)
<5> Bandwidth and ping is not the same
<4> i understand that
<5> And if a server is "busy doing other stuff" it should probably be redesigned
<0> don't you still have the problem of two moving players, one shoots at the other and the two systems disagree on whether and intervening object "blocks" the shot? (rounding errors, timing, whatever)
<6> Client also has fewer connections than server
<5> Which two systems?
<0> oh, the server is making ALL the decisions?
<4> well there's two types of games going on here... one is peer to peer, the other is with a central server(which would make all decisions, which is also the type of system i've been talking about)
<6> I think it'll work well if the client sends 2 or 3 requests after recorded time intervals.
<0> ok, I thought motion etc was handled by the clients (to not overload the central server)
<6> The difference in reception time of the server will give a better roundtrip estimate.
<4> nope
<4> the client uses prediction, though
<4> the server updates the clients every so often
<6> Or the server could send its own round trip estimate.
<4> dead reckoning is a commonly used technique - the server uses the client interpolation algorithms to figure out where the client thinks the object is at, compares it to where it knows it's really at, and if they differ more than some threshold, it updates
<4> there's a bunch of other ways to handle it - being farther away from the object means you get less updates
<4> sending relative position updates instead of absolute
<4> etc
<6> Whats the nature of the object?
<6> A 3D mesh in the game for example?
<4> IRR^ - a character that can walk around
<4> 2d or 3d is a minor detail
<6> I was just cooking an example. Confirming rather.
<6> Thats interesting actually, because the client-side perception of the quality of synchronisation may be different from the actual quality of synchronisation.
<4> yeah, it's accepted that each client has a different inaccurate view
<6> So you could focus on timely updates of objects the user is expecting to focus upon, and worry less about the others.
<4> im not really familiar with how peer-to-peer games handle it - I suppose someone is designated arbiter, or maybe several people handle different stuff
<6> Its tricky, especially when there is no guarantee about the QoS of the client ISP.
<4> yup
<4> well
<4> if they have a crappy ISP, their gameplay is going to be crappier than others
<4> if you can't get data through, there's no way around that
<4> best you can do is design it so the data is most likely to get through
<6> One thing you can do is to develop a proximity map between different clients.
<4> eh?
<6> That way a nearby client can act as a transient update source.
<4> bah
<6> As long as its a viable update source.
<4> if something is far away, there's no need for frequent updates
<6> Its client/server + peer to peer
<4> that's the whole idea - send less updates about stuff far away
<6> Yeah but thats not the problem I am addressing.
<4> what problem are you addressing?
<6> I am addressing the problem of where to send the update from so you can be sure that its timely.
<6> where from / how to ... both
<4> i don't think that's the problem
<6> Well gtg then
<4> the main reason for decreasing the frequency is so that the client has less data to process
<4> or, less data to fit through their tiny connection, really
<4> *shrug*
<4> anyway, just got an email saying they want to fly me out heh
<3> Congrats :D
<3> Where are they
<4> seattle :D
<4> redmond, specifically
<7> we don't like your kind in redmond
<3> MS or just some remora
<3> If it's MS, make them do something with their Shadowrun license
<4> nah, not MS
<4> http://www.zipperint.com/
<2> http://breakingnews.iol.ie/news/story.asp?j=143258260&p=y43z5884x
<5> You're aware sony is going to die pretty soon right? ;)
<2> Sony is going to die?
<8> Yeah right...
<8> Microsoft too
<4> lol
<9> i think sony'll outlive me tbh
<5> Sony's losing so much money they don't know what to put in banks
<5> And the PS3 won't be out this century
<5> => death
<9> hmmmm, where's the info from?
<5> Which one?
<9> that they're losing money
<5> In all their reports?
<9> hmm, ok then
<9> i never bother checking those things
<5> They don't fire a bunch of employees just for fun ;)
<9> haha
<9> well some do it, midway games for example
<9> "omg let's fire a bunch of employees at one place, then hire a whole bunch of new ones in another place"
<5> Sony's dropping 10k people (from last year to 2008)
<9> holy ****, ok, that's not a good sign
<5> But then when it'll be close to armageddon
<5> I'll buy sony for $1
<5> Release the PS3
<5> And become god
<5> And don't try to steal my plan, cause I patented it
<9> alright then, sure!
<9> :D
<4> lol
<9> oh and i'll call it ps3.5 just incase
<5> Might as well call it PS360
<9> hahaha
<4> heh
<9> i think i'll then get a big red court order from microsoft
<2> http://www.duiblog.com/2006/02/25#a357
<9> "omg another console with 360 in it's name!"
<5> Late 2007, I'll get Sony
Return to
#c++
or
Go to some related
logs:
#php
#linux
What thirst quencher was promoted with the line 'Probably the best lager in the
ucarp stateful failover
netforce misskitten
#mirc
hidebehind dewez
#AllNiteCafe
#linux
#MissKitten
|
|