What would be your guesstimate as to the total number of tabs you have open? I've regularly had ~550 open spread across three->five windows. The only symptoms that I've noticed are slow scrolling, tab switching, and text input response times... Firefox doesn't vanish under my workloads.
Do you have the official Flash plugin installed? If so, try uninstalling it. Firefox used to vanish or hard-lock frequently until I removed the Flash 9 plugin. I can't speak about the quality of the Flash 10 plugin, as I don't have it installed.
The acceptable latency is really independent of the desired frame rate; the required bandwidth depends on the frame rate. Latency affects responsiveness, sure, but the same latency will provide the same responsiveness no matter what framerate you have.
What happens to an unbuffered video stream when the next frame can't be delivered on time? (Is this "unbuffered video stream" notion the right way for me to be thinking of this "server-side rendered" MMO?) What happens to the client who has, say a 500ms ping to his server? Is his character just *really* slow to react? (I'm imagining something like: * 250ms client command to server (includes ACK of previous frame transmission) * 10ms server renders frame * 250ms server sends frame to client
Or is it a completely insane idea to wait to render the next frame until we get an ACK from the client that the previous one has been received?) [The more I think about it, the more insane this scheme that I've cooked up seems to me.]
You might drop frames if you had something else going on on the client computer, so that in the span between the display of one frame and the display of the next frame, more than one new frame was received. In that case, you would drop any undisplayed frame between the last frame displayed and the most recent frame received.
^^^ This kinda dovetails in with this next quote:
Latency won't make him able to update his position at 1/10 the rate of everyone else. Latency will just make his current view of the world farther behind everyone else's--if it takes a packet 70msec to get to one client and 35msec to get to another...
Aye. 35ms vs. 70ms is not a big deal. This happens all the time. What happens when you have two players that provide positioning data to two separate entities... one with a ping of 70ms, the other with a ping of 700ms. (Let's assume that both clients are not CPU bound... they can achieve a 60+FPS render for anything that the server can throw at them.) You can't have the 700ms player's view of the world be *eternally* and *increasingly* behind the 70ms player's... otherwise they're no longer controlling the same simulation.
(Sorry. I feel like I'm not being clear here. I'll get some sleep and come back with a better explanation of my thoughts. [Assuming that you don't suss out what I'm trying to say in the interim.])
You may have something there. Please let me voice my objections and thoughts on the matter. Your critique is welcome and encouraged.:)
* Latency: Consider that a ping of no more than ~34 msec would be required to stream responsive 30fps video. We're not talking about pre-rendered video that can be buffered to guard against future network issues, I imagine that this all would be generated on the fly.
IDK about you, but it's a good day when I get 50 msec to the servers that I play on. An average day sees 70 msec.
* Thoughts on Latency's affect on live video: I'm having trouble figuring out what would happen to an interactive, unbuffered video stream (that would probably be transported over UDP) when the latency on a client's connection rises to an unacceptable level. Would you have to drop frames? (You've seen the HPB [0] in your FPS game who skips (or teleports) around the map 'cause his ping is soo bad that he can only update his position at 1/10th the rate of everyone else, right?) Would the user see the video stream skip in the same way that folks with acceptable latency see the HPB skipping around?
* Random trick and conclusion: I bet that some good things could be done by rendering things like the HUD and mouse client-side, but you're gonna have to have a faster and more reliable connection than most USian residential customers have.
Um... a) I don't see how your reply addresses the PP's point. Did you read what he had to say?
...were it possible to have different gun laws for the urban dweller versus the rural dweller...
b) There are cities -such as the one that I'm currently living in- which make it unlawful for a civilian to discharge a weapon within city limits. Concealed carry permits may be obtained. However, if you discharge your weapon -even in self-defence- outside of a designated area (such as an indoor gun range) you will be charged and legally obligated to haul your ass in to court. Folks that live in rural areas typically don't operate under such restrictions.
You cannot appoint a military organization whose effectiveness depends on ignorance of its capabilities and vulnerabilities to protect civilian infosec. The only way any newly discovered vulns will ever be disclosed to the public by this sort of watchdog is if it is felt that "The Enemy" already knows about them and has a workaround, and that the disclosure would not compromise the position of any spies/well placed janitors.
After all, we're *all* generally using the same basic computing infrastructure these days.
This isn't news if the users of McD's wifi have to click through a page that discloses the surveillance program run by Clean Pipes before transmitting a packet to the Internet.
IMO, McD would be insane to set up the system in any other way.
On the other hand, one MIGHT have expected MS to make life a little easier on those who developed them.
You are aware that MSFT used to have "plugfests" with the SAMBA team? These were little shakedown meetings where the SAMBA folks would get together with some MSFT folks and test their code against several versions and configurations of Windows, and get to ask detailed questions of the MSFT folks. IIRC, the "plugfests" stopped in the late-mid to late 1990's. There may have been another plugfest or two recently, what with the EU breathing down MSFT's neck.
Or, why are we so up-tight about doctors and civil engineers when they should have the same laissez-faire setup as software engineers?
Hokay, I'll bite. Last I checked, not just anyone had access to the surgeon's tools, equipment, and operating theater. Only trained, authorized professionals or -at the surgeon's pleasure- authorized professionals-in-training have access. The same goes for the civil engineer and his contractors. The CE specs the design and materials. Trained, professional builders or -under their supervision- professional builders-in-training use his specced materials and design to create the CE's structure.
a) Handing an untrained or malicious user an account on computer, is just as risky as same as replacing a surgeon's scalpel with a band saw, or replacing the CE's concrete with marshmallow. Moreover, that untrained user can often accomplish this without being aware that he's screwed up. (Getting bitten by the malware-of-the-week would do it.)
b) The CE certifies a design. The contractor build said design. The software analogue to the CE's certified design is a system with *very* specific revisions of each and every piece of software. Changes to any piece of software invalidate the certification and/or warranty, having much the same effect as changing the CE's steel rebar with candy cane would. Most folks call such a system an "appliance".
Moving from one to two cores for me has meant that browsers can't suck down 100% of my CPU and prevent me from even closing them for minutes at a time. This had better not let Firefox use up 100% of my machine again...
So, what was it that you were *trying* to say? FWIW, I understood your statement in the way that Eunuchswear did.
What happens when you have a safe that is believed to contain -among other things- documents that may be (fuck me, IDK the legalese, but bear with me) important to the case at hand? What sort of warrant does the judge approve then? 'Cause that is *exactly* the same type of situation that you have when a warrant is issued to seize and search a hard drive.
If you were smart, you'd write up a little script that scraped my user page for new posts every ten minutes or so and posted some of your copypasta to each one. If you were *really* smart, you'd do all this through a good proxy, so the admins here wouldn't catch on.
That plasmoid that I mentioned is *really* beta software and is still in heavy development. It's possible that Kbuntu packaged a crappy version of it. (The version that I have [from April 26] works just fine for wired and wireless connections.) I'll check out wicd, thanks for the mention.
Also, IIRC, TrollTech doesn't have all *that* much control over KDE as a whole, and doesn't have any control over Amarok.
Slashdot Mirror
What would be your guesstimate as to the total number of tabs you have open? I've regularly had ~550 open spread across three->five windows. The only symptoms that I've noticed are slow scrolling, tab switching, and text input response times... Firefox doesn't vanish under my workloads.
Do you have the official Flash plugin installed? If so, try uninstalling it. Firefox used to vanish or hard-lock frequently until I removed the Flash 9 plugin. I can't speak about the quality of the Flash 10 plugin, as I don't have it installed.
Odds are good that that mouse was made by Logitech, under contract from MSFT. :)
Was the PDP-8 powered down at the time that it was discovered?
The acceptable latency is really independent of the desired frame rate; the required bandwidth depends on the frame rate. Latency affects responsiveness, sure, but the same latency will provide the same responsiveness no matter what framerate you have.
What happens to an unbuffered video stream when the next frame can't be delivered on time? (Is this "unbuffered video stream" notion the right way for me to be thinking of this "server-side rendered" MMO?) What happens to the client who has, say a 500ms ping to his server? Is his character just *really* slow to react? (I'm imagining something like:
* 250ms client command to server (includes ACK of previous frame transmission)
* 10ms server renders frame
* 250ms server sends frame to client
Or is it a completely insane idea to wait to render the next frame until we get an ACK from the client that the previous one has been received?) [The more I think about it, the more insane this scheme that I've cooked up seems to me.]
You might drop frames if you had something else going on on the client computer, so that in the span between the display of one frame and the display of the next frame, more than one new frame was received. In that case, you would drop any undisplayed frame between the last frame displayed and the most recent frame received.
^^^ This kinda dovetails in with this next quote:
Latency won't make him able to update his position at 1/10 the rate of everyone else. Latency will just make his current view of the world farther behind everyone else's--if it takes a packet 70msec to get to one client and 35msec to get to another...
Aye. 35ms vs. 70ms is not a big deal. This happens all the time.
What happens when you have two players that provide positioning data to two separate entities... one with a ping of 70ms, the other with a ping of 700ms. (Let's assume that both clients are not CPU bound... they can achieve a 60+FPS render for anything that the server can throw at them.)
You can't have the 700ms player's view of the world be *eternally* and *increasingly* behind the 70ms player's... otherwise they're no longer controlling the same simulation.
(Sorry. I feel like I'm not being clear here. I'll get some sleep and come back with a better explanation of my thoughts. [Assuming that you don't suss out what I'm trying to say in the interim.])
Hio. I have some thoughts on this matter. If you'd care to spend a few minutes reading them, I'd really like to hear your feedback.
http://slashdot.org/comments.pl?sid=1233885&cid=27971371
You may have something there. Please let me voice my objections and thoughts on the matter. Your critique is welcome and encouraged. :)
* Latency:
Consider that a ping of no more than ~34 msec would be required to stream responsive 30fps video. We're not talking about pre-rendered video that can be buffered to guard against future network issues, I imagine that this all would be generated on the fly.
IDK about you, but it's a good day when I get 50 msec to the servers that I play on. An average day sees 70 msec.
* Thoughts on Latency's affect on live video:
I'm having trouble figuring out what would happen to an interactive, unbuffered video stream (that would probably be transported over UDP) when the latency on a client's connection rises to an unacceptable level. Would you have to drop frames? (You've seen the HPB [0] in your FPS game who skips (or teleports) around the map 'cause his ping is soo bad that he can only update his position at 1/10th the rate of everyone else, right?) Would the user see the video stream skip in the same way that folks with acceptable latency see the HPB skipping around?
* Random trick and conclusion:
I bet that some good things could be done by rendering things like the HUD and mouse client-side, but you're gonna have to have a faster and more reliable connection than most USian residential customers have.
[0] High Ping Bastard
Um...
a) I don't see how your reply addresses the PP's point. Did you read what he had to say?
...were it possible to have different gun laws for the urban dweller versus the rural dweller...
b) There are cities -such as the one that I'm currently living in- which make it unlawful for a civilian to discharge a weapon within city limits. Concealed carry permits may be obtained. However, if you discharge your weapon -even in self-defence- outside of a designated area (such as an indoor gun range) you will be charged and legally obligated to haul your ass in to court.
Folks that live in rural areas typically don't operate under such restrictions.
What's update-manager's resident size?
I don't use Ubuntu, so I can't check for myself. :(
You found the contact information that I added to my TOR exit node. Good sleuthing! :rolleyes:
https://server.privacyfoundation.de/torstatus/router_detail.php?FP=68cf91bd0e58e0128dd40ce7bdec824fda49a925
^^^^ THIS.
You cannot appoint a military organization whose effectiveness depends on ignorance of its capabilities and vulnerabilities to protect civilian infosec. The only way any newly discovered vulns will ever be disclosed to the public by this sort of watchdog is if it is felt that "The Enemy" already knows about them and has a workaround, and that the disclosure would not compromise the position of any spies/well placed janitors.
After all, we're *all* generally using the same basic computing infrastructure these days.
This isn't news if the users of McD's wifi have to click through a page that discloses the surveillance program run by Clean Pipes before transmitting a packet to the Internet.
IMO, McD would be insane to set up the system in any other way.
OT comment about your sig:
Whaddya think of KDE 4.2.x? :D
Also, I've never used QT Creator. *downloads it*
Bash has a tool called ulimit that allows you to limit -among other things- how much CPU time and memory a process run from that shell can consume.
This sounds like maybe it might be a solution to your poorly behaved application.
That's where ionice comes in to play.
http://linux.die.net/man/1/ionice
Available since 2.6.13.
On the other hand, one MIGHT have expected MS to make life a little easier on those who developed them.
You are aware that MSFT used to have "plugfests" with the SAMBA team? These were little shakedown meetings where the SAMBA folks would get together with some MSFT folks and test their code against several versions and configurations of Windows, and get to ask detailed questions of the MSFT folks.
IIRC, the "plugfests" stopped in the late-mid to late 1990's. There may have been another plugfest or two recently, what with the EU breathing down MSFT's neck.
Or, why are we so up-tight about doctors and civil engineers when they should have the same laissez-faire setup as software engineers?
Hokay, I'll bite.
Last I checked, not just anyone had access to the surgeon's tools, equipment, and operating theater. Only trained, authorized professionals or -at the surgeon's pleasure- authorized professionals-in-training have access.
The same goes for the civil engineer and his contractors. The CE specs the design and materials. Trained, professional builders or -under their supervision- professional builders-in-training use his specced materials and design to create the CE's structure.
a) Handing an untrained or malicious user an account on computer, is just as risky as same as replacing a surgeon's scalpel with a band saw, or replacing the CE's concrete with marshmallow. Moreover, that untrained user can often accomplish this without being aware that he's screwed up. (Getting bitten by the malware-of-the-week would do it.)
b) The CE certifies a design. The contractor build said design. The software analogue to the CE's certified design is a system with *very* specific revisions of each and every piece of software. Changes to any piece of software invalidate the certification and/or warranty, having much the same effect as changing the CE's steel rebar with candy cane would. Most folks call such a system an "appliance".
Moving from one to two cores for me has meant that browsers can't suck down 100% of my CPU and prevent me from even closing them for minutes at a time. This had better not let Firefox use up 100% of my machine again...
So, what was it that you were *trying* to say? FWIW, I understood your statement in the way that Eunuchswear did.
Heh. WTF are you doing that fucks your system so hard? Are you SSH'd in, or running a xterm (of some sort)?
What happens when you have a safe that is believed to contain -among other things- documents that may be (fuck me, IDK the legalese, but bear with me) important to the case at hand? What sort of warrant does the judge approve then? 'Cause that is *exactly* the same type of situation that you have when a warrant is issued to seize and search a hard drive.
If you were smart, you'd write up a little script that scraped my user page for new posts every ten minutes or so and posted some of your copypasta to each one.
If you were *really* smart, you'd do all this through a good proxy, so the admins here wouldn't catch on.
Let's see how smart you are. :)
Also, I have outstanding questions:
http://slashdot.org/comments.pl?sid=1219095&cid=27809231
Until they are answered to my standards, I have nothing more for you.
And, moreover, what would they *do* with your IP address if they had it?
You're a security expert, and they're a bunch of wannabes, remember?
No, no, no.
How do they know *your* IP address?
Where's the copypasta that you promised me? I have a new reply to someone who's not you. It's more than an hour old.
How could the Ars folks differentiate between your IP and mine?
Moreover, why wouldn't you browse the forums through Tor? That works just fine.
That plasmoid that I mentioned is *really* beta software and is still in heavy development. It's possible that Kbuntu packaged a crappy version of it. (The version that I have [from April 26] works just fine for wired and wireless connections.)
I'll check out wicd, thanks for the mention.
Also, IIRC, TrollTech doesn't have all *that* much control over KDE as a whole, and doesn't have any control over Amarok.