The way i see it is that SSDs are useful if performance is your top priority. For a server, the capacity of the device may not matter if the device is too slow. That's why 15kRPM hard drives exist. They are smaller and more expensive, but are useful as system drives. I keep my other stuff on 7.2kRPM drives because they are cheaper and have higher capacity and I do not really care about the speed for those files. I also archive rarely used files to tape, even though it has much slower random access than hard drives, tapes are a bit cheaper and (hopefully) more reliable in long term storage.
I actually like the UI of my N93. It could be improved, but in my opinion it's good as it is. If Nokia decides to change the UI, I hope it does not listen to the people who think that the UI of MS Excel 2007 is good.
I mainly use nokia phones.I wouldn't want a phone with Windows on it (or rather, I wouldn't mind a phone with normal Windows on it, not the WM7). It'sa shamethat Symbian is almost dead. Anyway, maybe Nokia will manage to get better by the time my current phone breaks down.
I don't think Mozilla can afford to distribute a copy of CoreAVC to everybody who doesn't already own a copy of CoreAVC.
They do not need to distribute it, I already have it, just use it. A lot of other people have ffdshow or similar codecs, just use them.
Mozilla does not distribute Flash (or Silverlight, or Java) with the browser too, but it can use the plugin if it is available on the system (it even shows where I can downlaod it if I do not have it). Why should the codec be any different?
A lot of such downloaded videos will use MPEG-4 Part 2 (DivX/Xvid) or Windows Media Video, not necessarily AVC.
Most of the videos I see on the pirate bay are either divx/xvid (SD) or h264 (HD). If you want HD you most likely have AVC. Some newer phones and cameras film in h264 too.
Flash Player doesn't crash Firefox anymore; it crashes the plug-in container.
It might crash the system. Also, the codecs can also be invoked by a separate process so they do not crash FF.
So there's no good way at all unfortunately to utilize CoreAVC inside the browser.
So, just download the file and watch it locally, that will be safer? Probably not. And I would have to download the file to watch it locally if it was too high resolution to be decoded by the less efficient codec.
So, instead they only provide the less popular codec, the copy of which I still would have to have in my system (so I could watch the downloaded videos) and now if a vulnerability is found inside the codec, I would have to update Firefox and the system one.
Also, for h264 I use CoreAVC, because it uses less resources and (mainly) manages to use all 4 cores of my CPUs (my main PC has two dual-core CPUs) about evenly, so I can watch 1080p video without problems, even at 60fps. Now let's say that for whatever reason Mozilla now can include h264 codec in the browser, they would probably choose ffdshow, which is less efficient than CoreAVC. Same thing about WebM. If the decoder is single threaded (or even if it is multithreaded but uses the cores unevenly, like some games), I will probably not be able to watch 1080p video. And even if CoreCodec makes a multithreaded and efficient decoder, I would not be able to use it on web videos, I would have to download the video and then play it.
And if Mozilla cares about choice, they can always make the internal codecs as a second choice - if the correct system codecis not found, try internal ones, if no suitable internal codec found, display which codec is needed to view the video.
For example, the end user might be missing a codec, which is likely if the end user is on Windows XP Home Edition, Windows XP Professional, Windows Vista Home Basic, Windows Vista Business, Windows 7 Starter, or any freely redistributable GNU/Linux distribution.
I really doubt that it is likely. Someone who watches videos on Youtube most likely also watches downloaded videos (maybe even downloads from Youtube), so he will have the codecs needed. I am sure that there are some people who just watch Youtube and similar videos but do not download any videos, however I really doubt that there is a lot of them.
Second, Mozilla doesn't want users to blame Firefox if a defect in a system codec causes a crash or intrusion.
So, instead it wants the users to blame Firefox for not showing some videos? Or if flash causes a crash?
Except browsers like Chrome and Firefox aren't actually using the system decoders.
Well, it's not like they cannot use the system codecs. Which means that this is just a stupid choice, the same way that I would insist on buying a PC power supply that is designed for 110V then using it with a transformer that converts the 230V in the outlet to 110V and complaining that the power supply was quite expensive because I needed to buy it from the US and then needed to buy the transformer.
Even if these guys were right, and with modern science you are able to find that one gene whose loss of function causes cancer, and they are able to fix it, all it means is you will not die of cancer, but will die of other geriatric diseases.
Well, maybe we can take them one at a time. Cure cancer then move on to whatever other disease now kills a lot of people and so on. You might die from getting hit by a bus, but that does not mean that eliminating diseases is pointless.
allow both the body and the brain to die together painlessly.
Bullet to the brain? Poison? Though nobody would ever know if a death if painful or not, since the dead person cannot tell you.
Wouldn't that be true of any fatal disease? Let's say that there was only one fatal disease left, so people who do not die due to injury or poison would eventually get that disease and die, so it would be "disease x would kill all humans if the other y% did not die from some other cause first".
technically speaking, two routers with two public IP's, each with half the office behind them each is "load balancing". if you don't want a cheep hack form of load balancing (where connections are handled properly based on line load) talk to your ISP. they're happy to help.
The problem I had was that one of the connections was unreliable, but I still wanted to use it, since 512kbps half of the time is about equal to 256kbps all the time. At first I tried two virtual machines, each connected to a different router and running uTorrent, but it was more harm than good, since if the WiFi connection went down, all downloads stopped even though my DSL connection was OK. So, I googled and found a way to make one uT instance use both conections, now if the WiFi connection went down the speed would just drop to the speed of the DSL connection and that's it. When I restarted the WiFi connection, the speed went back up.
- Making example.com:21 and example.com:80 be separate servers. (in the cisco world) create an ACL patching each port, and set them on a route-map with the next-hop being the server(s) in question.
Good to know, any way of doing that on Linux? I have been using DNAT (iptables) (and SNAT on the replies) any time I needed that.
Conversely, making example.com:80 and example1.com:80 be one server while example.com:21 and example1.com:21 are another server. EASY to do with a load balancer. (though many will say it's basically the same thing.)
For a load balancer to keep the source IP intact (so all connections have the correct source and do not appear to originate from the load balancer), it would have to work as NAT. Or at least it would be possible to make NAT using it.
- Transparent proxies. already covered in the comments above. your router can easily move flows of TCP port 80 traffic to another destination.
And when the reply packet from the proxy arrives, the router will have to change the source IP so that the client accepts it (it would not work if I sent a packet to google.com and received a reply from proxy.ip), so it is (or can be used as) NAT.
Dial-up Internet access in the United States was typically billed at a flat monthly rate. A few ISPs imposed a limit of 100 to 150 hours a month to free up modem pools.
At the time, in Lithuania, it was by the minute. I later actually use dial-up and paid by the minute. As I recall, there was a plan with constant monthly fee, but the fee was very high so it was cheaper to pay by the minute and only use the connection a few minutes per day.
A lot of other Slashdot users explain to me that they play online because A. their friends live hundreds of miles or hundreds of km away and they can't afford to be a frequent flyer, or B. they play with strangers because their friends don't play the same video games.
C. My friend lives about 10km away, but he has something else to do, so, while we can play the game for an hour or so, going there would take too long and would be less convenient than just playing the game online. Also, FPS games (and most of the other games) are better with keyboard+mouse (in my opinion) and I do not have a console.
Load balancing two internet connections without any cooperation from the ISPs. Making example.com:21 and example.com:80 be separate servers. Conversely, making example.com:80 and example1.com:80 be one server while example.com:21 and example1.com:21 are another server. Transparent proxies.
The "game asks/installer still starts" behavior is a bug in the game. They've failed to (correctly) implement IQueryCancelAutoPlay. A proper game would suppress autoplay after asking for a DVD until seeing the game DVD.
May be, but with disabled Autorun I can just launch the setup.exe file on the CD if I want to.
I do not like things auto starting when I insert a CD. Or connect to the internet (in case where the connection is not always-on). My CD player requires me to press play after I insert the CD, which is good. Windows wants to automatically start playing a music CD, which is not good. Nobody ever thought that I might insert the CD now to play it later?
As for Autorun not working on floppies, well, that's not a Microsoft design decision. Autorun triggers on the insertion notification, and most floppy drives don't have those. That's why you had to hit any key after changing a floppy in a multi-disk installer.
Hmm... So I gave Microsoft too much credit. OTOH, if Autorun was possible on floppies, the Autorun viruses would have started to spread much earlier and the Autorun would have been already disabled by the time USB flash drives became popular.
Why not as a home user? Do you seriously think your $20 NAT box, with the last firmware update 2003 is more secure than your OS?
My Linux or windows based router (which i could use without NAT if needed) ir probably secure enough, and not having the firewall enabled on the internal network adds to the performance, since now my computers can use the segmentation offload capabilities in my Intel network cards.
If your system has the port closed you have nothing to worry about. If the port is open it should be open in the firewall anyway.
So, no such thing as internal services, for example, shared drives, network printers etc?
You can just as well do it properly with a proxy or load balancer, most likely with better performance as well.
The proxy either needs to be specific for the protocol used or (in case of simple port forwarding, "nc -l -p 1234 | nc 1.2.3.4 4321" it will make all connections appear to originate from the device that does the port forwarding).
Okay, you got me. You can do IPv4 transparent load balancing on flow level between different ISPs using NAT. That is hard to do without it. Not that such load balancing tends to work very well (servers tend to bind your identity to the IP address, breaking stuff like authentication).
While it did not work very well (I actually used it), it did allow me to get better upload and download speeds in BitTorrent. My ISP allowed WiFi access and I had DSL. Since ADSL is not known for fast upload (in my case it was about 800kbps), I decided to use the WiFi (since the access point was within range). However, trying to use different virtual machines (each for one WAN connection) wasn't all that great, since the WiFi connection sometimes failed (and i had to reset it manually, because I had to log in to use that connection). So, I used pfsense to load balance both connections, and guess what? My upload speed increased to approximately the sum of upload speeds of both connections. So did my download speed. And if the wifi connection failed, the download would still finish on DSL only.
The alternative is using a transport protocol actually built for Mobility, like SCTP. Or IPv6 with mobility extensions. And then you suddenly don't need NAT and things will actually work...
Does it work with BitTorrent? Can it be used without any cooperation from the ISP(s)?
No, you don't need NAT for it. NAT is an ugly hack to solve a very specific problem created by ISPs trying to prevent customers from having too many systems connected. A hack that many unskilled network administrators use as their golden hammer. NAT is something holding us back, locking us into a situation were we can not deploy new protocols and systems that the Internet badly needs.
So, how do I get transparent proxy without NAT? I need to somehow redirect the packet to the computer where the proxy software is running.
Also, as with h264 vs WebM or open vs closed source (but still $0), I do not care about the ideology as long as stuff works. Same thing here. NAT may be an "ugly hack", but it works (at least for me) and it offers other options beside its original purpose. But that's the point - NAT is useful for me, so I want the option of NAT on IPv6. You can choose not to use it, it won't affect me. Or the fact that I choose to use it won't affect you.
I understand that the "ideal" of the internet is every computer having at least two IP addresses (one that changes whenever the ISP changes, one local that does not depend on the ISP), maybe more, local DNS and DHCP servers that do not crash, ever, so nobody would need to enter the IP address anywhere, except maybe when setting the IP of a computer or configuring the firewall. And no one would want to make two computers appear as one, for any reason, ever.
The entire point of the Internet and IP was that two addresses could communicate.
Sure, I can forward ports, just tell me which ones. Even without NAT, I would still have to know which ports to open in the firewall (you are not suggesting that I should be without a firewall, right?).
It's bad design to have to do weird hole-punching type hacks to get around it.
It's bad design to not trust the header of the packet and embed the copy of the IP address inside the data section of the packet. There is very little reason to do so and should not be done unless really needed.
OTOH, with NAT, I can make example.com:80 and example.com:21 actually be different servers without resorting to www.example.com and ftp.example.com. Or, I could make two external IPs point to a single server. I could also make example.com:80 and example1.com:80 be one server, while example.com:21 and example1.com:21 be another one. I can also load balance two internet connections with no cooperation from the ISPs. Or, I can have a transparent proxy.
See, all that fun stuff needs NAT. Not just making a bunch of computers use only one external IP.
Autorun made some sense when it worked only on CD-ROM disks, though sometimes it still was annoying (start a game, the game asks for the CD, insert the CD and the installer starts - this on slow PCs with little memory and slow CD drives). It did not work on floppies, so maybe someone saw that it would be bad. When USB flash drives replaced floppies in every day use it was only a matter of time before virus writers took advantage of Autorun.
But what's stopping users from navigating to that CD or flash drive and executing the code themselves? Aren't they the ones connecting the devices or putting the disks in their computer in the first place?
Nothing, but a lot of the infections were due to unintentional activations of the virus. Insert flash drive, go to My Computer and double click on the icon, you have a virus. OTOH, some people actually know that they should not double click on executable files if they do not know what the file does.
Not always. a year or so ago Windows XP did not completely implement that policy. That is, it would not auto run it, but would still read the autorun.inf file and let it change/replace the context menu and the default behavior when you double click on the drive icon. So a virus could still be unintentionally started by double clicking the drive icon. A registry hack was needed to prevent windows from reading autorun.inf. Later, Microsoft released an update that made Windows work as intended with the autorun.in file, but you had to turn that behavior on in the registry (the setting was named HonorAutoRunSetting).
Indie developers have little or no access to consoles, leaving them without a viable platform for which to develop games in shared-screen genres.
It's not like the PC somehow prevents them from making a shared screen game. I know because I played Need For Speed 2 shared screen all the time (back then I did not have an internet connection, not to mention that even if I had it, it would most likely have been dial-up and charged by the minute). Either my friend would come to me or I would come to him and we would play the game.
You'd still benefit indirectly, as source code is useful to developers of mods that you might get the itch to play.
Could be, but then again, the game could be made so that mods are possible. Half Life (both the original and the Source engine) is one example. Also, I have noticed that in some cases open source is a bit worse than closed source when it comes to support - as you can sometimes get the "well, it's open source - go study programming for 5 years and just modify it yourself" answer when reporting a bug or requesting a new feature.
This way, the games can draw wide open spaces filled with thousands of objects. Yet despite the low detail, all the objects are easy to make out.
Great, though I still like having more detail. Compare it to this - the story/acting/etc of a movie does not depend on the recording medium, that is, I can watch the movie on a black&white TV with geometry distortions and would still be able to enjoy it (after all, a lot of good movies/TV shows were made before there was color TV or film), if the movie actually has an interesting story and so on, on the other hand watching it in HD (or at least color) is better.
And inability to concisely express system requirements is one of the major faults of PC gaming.
Yes, that is a problem, but it is understandable - there are so many different possible configurations that it is impossible to test on all of them. For example, my PC is quite non-standard, while it has two CPUs with two cores each (so, similar to quad core), the CPUs are only 2GHz, so single threaded games run worse and multi threaded games run better than, say, on a CPU that has two cores 3GHz each (let's assume that GHz equals speed for the sake of this comparison).
Say you have little Abigail and little Chester who live together and want to play together, but only one of the PCs in the household is a gaming PC, and the head of household has bought only one copy of the game for them to share....
Well, in that case, console would most likely be better, seeing as there are more games that support local multiplayer (this is not a limitation of the PC platform, it's just that there are very few games that support it). OTOH, there is a different take on this - my friend does not have to come to my house (or I to his house) for us to play Borderlands (a game that is quite long) over the internet. Even if we live 100km away from each other, we can still play together. Consoles usually allow both, but as I said, I do not really need the local multiplayer option.
But all PLAYSTATION 3 consoles and all Xbox 360 consoles after the first few batches have HDMI out, and many newer LCD monitors tend to have HDMI in or DVI in with HDMI audio support.
While my monitor would does not support HDMI, I could manage to connect these consoles using HDFury or a similar device. The problem is video mode. 1080p is 1920x1080@60Hz, while my monitor would support that (though I would have to adjust the image height), a lot of LCD monitors only support lower resolutions. A lot of CRT monitors would not support that video mode either. On a PC, I can set the resolution all the way from 640x480 to 2048x1536, so I can find one that my monitor supports.
So allow me to rephrase your standard as I understand it: if you don't have at least Xbox/Wii class video hardware and comparable asset authoring capacity, keep it 2D.
Yes, good looking 2D is better than blocky 3D. A lot of adventure games have painted backgrounds and they look really good and also can be played on older computers.
Consider a choice between a video game that is free and Free [google.com] vs. one that is proprietary and $50 per player. Or a video game that runs on hardware you own vs. a video game that needs a new CPU, a new GPU, and more RAM. At what point do graphics override other considerations?
Well, I said similar games, and the price would be in the comparison. I do not care about GNU Free though, as I am not a good programmer, the source code would be useless to me. If a video game needs a new CPU, GPU and more RAM, then I will not play it now, but when I upgrade my computer (which I would still do once in a while), I'll play it. If the game is much more expensive, then I'll probably choose the cheaper game. It also depends on the rest of the game. System Shock 2 is great as it is (even though the human models look awful), Borderlands would not be as good without the wide open spaces, ability to see an enemy from far away and so on. I also am using a HD texture pack for Minecraft, it made the game better for me (and I would like even higher resolution textures than 32x32, but probably the game cannot handle it). Bad game with good graphics would still be a bad game, but good graphics make a good game better. I enjoyed the original Half Life, but I am really waiting for Black Mesa (Original Half Life with HL2 graphics).
Of course there is. Just like with consoles - can you play a PS3 game on a PS2? just that with PCs it is not that clearly defined. Some new games work on older hardware some don't.
But how many people can use your PC at once? Console games are more likely to allow four players on one large monitor.
Seeing that I play single player games or multiplayer over the internet, this is not a problem for me. On the other hand, console controllers (in my opinion) are worse than keyboard+mouse (especially for FPS games, or even point-and-click adventure). Also, it may be difficult to connect a console to a computer monitor for single player use (my CRT monitor would most likely work with this as long as the console has a VGA (or RGB) output, but most LCDs and even CRTs wouldn't).
Let me guess: you can't tolerate La Mulana, Eversion, I Wanna Be the Guy, or other games with retraux art style [tvtropes.org] that calls back to the 8-bit days.
I went found a screenshot of La Mulana. The graphics are OK, since I can still see stuff. OK, I'll try to explain: SMB style games with those graphics are OK - the graphics do not have to be 3D, but they have to be high enough resolution. On the other hand, games like NetHack are out. Also, some 3D games have very low resolution models and textures, to the point that it is difficult to recognize what I am looking at.
Still, for me to play a game with low quality graphics it has to be very good in other areas. I played The Longest Journey because I really liked the story (and tolerated the 640x480 fixed resolution), though I would have been really happy if the game had better graphics. I have also played the original GTA, but to me, GTA3 and newer are much better, in no small part due to being 3D.
Given a choice between to very similar games, I'd pick the one with better graphics.
Slashdot Mirror
$100 means to you much more when you earn $200/month than when you earn $10k/month.
The way i see it is that SSDs are useful if performance is your top priority. For a server, the capacity of the device may not matter if the device is too slow. That's why 15kRPM hard drives exist. They are smaller and more expensive, but are useful as system drives. I keep my other stuff on 7.2kRPM drives because they are cheaper and have higher capacity and I do not really care about the speed for those files. I also archive rarely used files to tape, even though it has much slower random access than hard drives, tapes are a bit cheaper and (hopefully) more reliable in long term storage.
I actually like the UI of my N93. It could be improved, but in my opinion it's good as it is. If Nokia decides to change the UI, I hope it does not listen to the people who think that the UI of MS Excel 2007 is good.
I mainly use nokia phones.I wouldn't want a phone with Windows on it (or rather, I wouldn't mind a phone with normal Windows on it, not the WM7). It'sa shamethat Symbian is almost dead. Anyway, maybe Nokia will manage to get better by the time my current phone breaks down.
I don't think Mozilla can afford to distribute a copy of CoreAVC to everybody who doesn't already own a copy of CoreAVC.
They do not need to distribute it, I already have it, just use it. A lot of other people have ffdshow or similar codecs, just use them.
Mozilla does not distribute Flash (or Silverlight, or Java) with the browser too, but it can use the plugin if it is available on the system (it even shows where I can downlaod it if I do not have it). Why should the codec be any different?
A lot of such downloaded videos will use MPEG-4 Part 2 (DivX/Xvid) or Windows Media Video, not necessarily AVC.
Most of the videos I see on the pirate bay are either divx/xvid (SD) or h264 (HD). If you want HD you most likely have AVC. Some newer phones and cameras film in h264 too.
Flash Player doesn't crash Firefox anymore; it crashes the plug-in container.
It might crash the system. Also, the codecs can also be invoked by a separate process so they do not crash FF.
So there's no good way at all unfortunately to utilize CoreAVC inside the browser.
So, just download the file and watch it locally, that will be safer? Probably not. And I would have to download the file to watch it locally if it was too high resolution to be decoded by the less efficient codec.
So, instead they only provide the less popular codec, the copy of which I still would have to have in my system (so I could watch the downloaded videos) and now if a vulnerability is found inside the codec, I would have to update Firefox and the system one.
Also, for h264 I use CoreAVC, because it uses less resources and (mainly) manages to use all 4 cores of my CPUs (my main PC has two dual-core CPUs) about evenly, so I can watch 1080p video without problems, even at 60fps. Now let's say that for whatever reason Mozilla now can include h264 codec in the browser, they would probably choose ffdshow, which is less efficient than CoreAVC. Same thing about WebM. If the decoder is single threaded (or even if it is multithreaded but uses the cores unevenly, like some games), I will probably not be able to watch 1080p video. And even if CoreCodec makes a multithreaded and efficient decoder, I would not be able to use it on web videos, I would have to download the video and then play it.
And if Mozilla cares about choice, they can always make the internal codecs as a second choice - if the correct system codecis not found, try internal ones, if no suitable internal codec found, display which codec is needed to view the video.
For example, the end user might be missing a codec, which is likely if the end user is on Windows XP Home Edition, Windows XP Professional, Windows Vista Home Basic, Windows Vista Business, Windows 7 Starter, or any freely redistributable GNU/Linux distribution.
I really doubt that it is likely. Someone who watches videos on Youtube most likely also watches downloaded videos (maybe even downloads from Youtube), so he will have the codecs needed. I am sure that there are some people who just watch Youtube and similar videos but do not download any videos, however I really doubt that there is a lot of them.
Second, Mozilla doesn't want users to blame Firefox if a defect in a system codec causes a crash or intrusion.
So, instead it wants the users to blame Firefox for not showing some videos? Or if flash causes a crash?
Except browsers like Chrome and Firefox aren't actually using the system decoders.
Well, it's not like they cannot use the system codecs. Which means that this is just a stupid choice, the same way that I would insist on buying a PC power supply that is designed for 110V then using it with a transformer that converts the 230V in the outlet to 110V and complaining that the power supply was quite expensive because I needed to buy it from the US and then needed to buy the transformer.
Even if these guys were right, and with modern science you are able to find that one gene whose loss of function causes cancer, and they are able to fix it, all it means is you will not die of cancer, but will die of other geriatric diseases.
Well, maybe we can take them one at a time. Cure cancer then move on to whatever other disease now kills a lot of people and so on. You might die from getting hit by a bus, but that does not mean that eliminating diseases is pointless.
allow both the body and the brain to die together painlessly.
Bullet to the brain? Poison? Though nobody would ever know if a death if painful or not, since the dead person cannot tell you.
Wouldn't that be true of any fatal disease? Let's say that there was only one fatal disease left, so people who do not die due to injury or poison would eventually get that disease and die, so it would be "disease x would kill all humans if the other y% did not die from some other cause first".
technically speaking, two routers with two public IP's, each with half the office behind them each is "load balancing". if you don't want a cheep hack form of load balancing (where connections are handled properly based on line load) talk to your ISP. they're happy to help.
The problem I had was that one of the connections was unreliable, but I still wanted to use it, since 512kbps half of the time is about equal to 256kbps all the time. At first I tried two virtual machines, each connected to a different router and running uTorrent, but it was more harm than good, since if the WiFi connection went down, all downloads stopped even though my DSL connection was OK. So, I googled and found a way to make one uT instance use both conections, now if the WiFi connection went down the speed would just drop to the speed of the DSL connection and that's it. When I restarted the WiFi connection, the speed went back up.
- Making example.com:21 and example.com:80 be separate servers.
(in the cisco world) create an ACL patching each port, and set them on a route-map with the next-hop being the server(s) in question.
Good to know, any way of doing that on Linux? I have been using DNAT (iptables) (and SNAT on the replies) any time I needed that.
Conversely, making example.com:80 and example1.com:80 be one server while example.com:21 and example1.com:21 are another server.
EASY to do with a load balancer. (though many will say it's basically the same thing.)
For a load balancer to keep the source IP intact (so all connections have the correct source and do not appear to originate from the load balancer), it would have to work as NAT. Or at least it would be possible to make NAT using it.
- Transparent proxies.
already covered in the comments above. your router can easily move flows of TCP port 80 traffic to another destination.
And when the reply packet from the proxy arrives, the router will have to change the source IP so that the client accepts it (it would not work if I sent a packet to google.com and received a reply from proxy.ip), so it is (or can be used as) NAT.
And how do I make the transparent proxy accept the packet if the destination IP in the header does not belong to it?
with IPv4 and iptables, i can do
-j DNAT --to-destination proxy.ip:8080
I have a LS120 drive too (and since I have disabled autorun I wouldn't know if it worked on LS120).
But no, it wasn't an OS design decision to not have autorun on floppies.
So I gave Microsoft too much credit.
Dial-up Internet access in the United States was typically billed at a flat monthly rate. A few ISPs imposed a limit of 100 to 150 hours a month to free up modem pools.
At the time, in Lithuania, it was by the minute. I later actually use dial-up and paid by the minute. As I recall, there was a plan with constant monthly fee, but the fee was very high so it was cheaper to pay by the minute and only use the connection a few minutes per day.
A lot of other Slashdot users explain to me that they play online because A. their friends live hundreds of miles or hundreds of km away and they can't afford to be a frequent flyer, or B. they play with strangers because their friends don't play the same video games.
C. My friend lives about 10km away, but he has something else to do, so, while we can play the game for an hour or so, going there would take too long and would be less convenient than just playing the game online. Also, FPS games (and most of the other games) are better with keyboard+mouse (in my opinion) and I do not have a console.
I can name a few of them:
Load balancing two internet connections without any cooperation from the ISPs.
Making example.com:21 and example.com:80 be separate servers.
Conversely, making example.com:80 and example1.com:80 be one server while example.com:21 and example1.com:21 are another server.
Transparent proxies.
The "game asks/installer still starts" behavior is a bug in the game. They've failed to (correctly) implement IQueryCancelAutoPlay. A proper game would suppress autoplay after asking for a DVD until seeing the game DVD.
May be, but with disabled Autorun I can just launch the setup.exe file on the CD if I want to.
I do not like things auto starting when I insert a CD. Or connect to the internet (in case where the connection is not always-on). My CD player requires me to press play after I insert the CD, which is good. Windows wants to automatically start playing a music CD, which is not good. Nobody ever thought that I might insert the CD now to play it later?
As for Autorun not working on floppies, well, that's not a Microsoft design decision. Autorun triggers on the insertion notification, and most floppy drives don't have those. That's why you had to hit any key after changing a floppy in a multi-disk installer.
Hmm... So I gave Microsoft too much credit. OTOH, if Autorun was possible on floppies, the Autorun viruses would have started to spread much earlier and the Autorun would have been already disabled by the time USB flash drives became popular.
Why not as a home user? Do you seriously think your $20 NAT box, with the last firmware update 2003 is more secure than your OS?
My Linux or windows based router (which i could use without NAT if needed) ir probably secure enough, and not having the firewall enabled on the internal network adds to the performance, since now my computers can use the segmentation offload capabilities in my Intel network cards.
If your system has the port closed you have nothing to worry about. If the port is open it should be open in the firewall anyway.
So, no such thing as internal services, for example, shared drives, network printers etc?
You can just as well do it properly with a proxy or load balancer, most likely with better performance as well.
The proxy either needs to be specific for the protocol used or (in case of simple port forwarding, "nc -l -p 1234 | nc 1.2.3.4 4321" it will make all connections appear to originate from the device that does the port forwarding).
Okay, you got me. You can do IPv4 transparent load balancing on flow level between different ISPs using NAT. That is hard to do without it. Not that such load balancing tends to work very well (servers tend to bind your identity to the IP address, breaking stuff like authentication).
While it did not work very well (I actually used it), it did allow me to get better upload and download speeds in BitTorrent. My ISP allowed WiFi access and I had DSL. Since ADSL is not known for fast upload (in my case it was about 800kbps), I decided to use the WiFi (since the access point was within range). However, trying to use different virtual machines (each for one WAN connection) wasn't all that great, since the WiFi connection sometimes failed (and i had to reset it manually, because I had to log in to use that connection). So, I used pfsense to load balance both connections, and guess what? My upload speed increased to approximately the sum of upload speeds of both connections. So did my download speed. And if the wifi connection failed, the download would still finish on DSL only.
The alternative is using a transport protocol actually built for Mobility, like SCTP. Or IPv6 with mobility extensions. And then you suddenly don't need NAT and things will actually work...
Does it work with BitTorrent? Can it be used without any cooperation from the ISP(s)?
No, you don't need NAT for it. NAT is an ugly hack to solve a very specific problem created by ISPs trying to prevent customers from having too many systems connected. A hack that many unskilled network administrators use as their golden hammer. NAT is something holding us back, locking us into a situation were we can not deploy new protocols and systems that the Internet badly needs.
So, how do I get transparent proxy without NAT? I need to somehow redirect the packet to the computer where the proxy software is running.
Also, as with h264 vs WebM or open vs closed source (but still $0), I do not care about the ideology as long as stuff works. Same thing here. NAT may be an "ugly hack", but it works (at least for me) and it offers other options beside its original purpose. But that's the point - NAT is useful for me, so I want the option of NAT on IPv6. You can choose not to use it, it won't affect me. Or the fact that I choose to use it won't affect you.
I understand that the "ideal" of the internet is every computer having at least two IP addresses (one that changes whenever the ISP changes, one local that does not depend on the ISP), maybe more, local DNS and DHCP servers that do not crash, ever, so nobody would need to enter the IP address anywhere, except maybe when setting the IP of a computer or configuring the firewall. And no one would want to make two computers appear as one, for any reason, ever.
The entire point of the Internet and IP was that two addresses could communicate.
Sure, I can forward ports, just tell me which ones.
Even without NAT, I would still have to know which ports to open in the firewall (you are not suggesting that I should be without a firewall, right?).
It's bad design to have to do weird hole-punching type hacks to get around it.
It's bad design to not trust the header of the packet and embed the copy of the IP address inside the data section of the packet. There is very little reason to do so and should not be done unless really needed.
OTOH, with NAT, I can make example.com:80 and example.com:21 actually be different servers without resorting to www.example.com and ftp.example.com.
Or, I could make two external IPs point to a single server.
I could also make example.com:80 and example1.com:80 be one server, while example.com:21 and example1.com:21 be another one.
I can also load balance two internet connections with no cooperation from the ISPs.
Or, I can have a transparent proxy.
See, all that fun stuff needs NAT. Not just making a bunch of computers use only one external IP.
Autorun made some sense when it worked only on CD-ROM disks, though sometimes it still was annoying (start a game, the game asks for the CD, insert the CD and the installer starts - this on slow PCs with little memory and slow CD drives). It did not work on floppies, so maybe someone saw that it would be bad. When USB flash drives replaced floppies in every day use it was only a matter of time before virus writers took advantage of Autorun.
But what's stopping users from navigating to that CD or flash drive and executing the code themselves? Aren't they the ones connecting the devices or putting the disks in their computer in the first place?
Nothing, but a lot of the infections were due to unintentional activations of the virus. Insert flash drive, go to My Computer and double click on the icon, you have a virus. OTOH, some people actually know that they should not double click on executable files if they do not know what the file does.
Not always. a year or so ago Windows XP did not completely implement that policy. That is, it would not auto run it, but would still read the autorun.inf file and let it change/replace the context menu and the default behavior when you double click on the drive icon. So a virus could still be unintentionally started by double clicking the drive icon. A registry hack was needed to prevent windows from reading autorun.inf. Later, Microsoft released an update that made Windows work as intended with the autorun.in file, but you had to turn that behavior on in the registry (the setting was named HonorAutoRunSetting).
There is no security benefit of NAT, but NAT offers some conveniences too, at the price of breaking a few badly designed protocols.
Indie developers have little or no access to consoles, leaving them without a viable platform for which to develop games in shared-screen genres.
It's not like the PC somehow prevents them from making a shared screen game. I know because I played Need For Speed 2 shared screen all the time (back then I did not have an internet connection, not to mention that even if I had it, it would most likely have been dial-up and charged by the minute). Either my friend would come to me or I would come to him and we would play the game.
You'd still benefit indirectly, as source code is useful to developers of mods that you might get the itch to play.
Could be, but then again, the game could be made so that mods are possible. Half Life (both the original and the Source engine) is one example. Also, I have noticed that in some cases open source is a bit worse than closed source when it comes to support - as you can sometimes get the "well, it's open source - go study programming for 5 years and just modify it yourself" answer when reporting a bug or requesting a new feature.
This way, the games can draw wide open spaces filled with thousands of objects. Yet despite the low detail, all the objects are easy to make out.
Great, though I still like having more detail. Compare it to this - the story/acting/etc of a movie does not depend on the recording medium, that is, I can watch the movie on a black&white TV with geometry distortions and would still be able to enjoy it (after all, a lot of good movies/TV shows were made before there was color TV or film), if the movie actually has an interesting story and so on, on the other hand watching it in HD (or at least color) is better.
And inability to concisely express system requirements is one of the major faults of PC gaming.
Yes, that is a problem, but it is understandable - there are so many different possible configurations that it is impossible to test on all of them. For example, my PC is quite non-standard, while it has two CPUs with two cores each (so, similar to quad core), the CPUs are only 2GHz, so single threaded games run worse and multi threaded games run better than, say, on a CPU that has two cores 3GHz each (let's assume that GHz equals speed for the sake of this comparison).
Say you have little Abigail and little Chester who live together and want to play together, but only one of the PCs in the household is a gaming PC, and the head of household has bought only one copy of the game for them to share....
Well, in that case, console would most likely be better, seeing as there are more games that support local multiplayer (this is not a limitation of the PC platform, it's just that there are very few games that support it).
OTOH, there is a different take on this - my friend does not have to come to my house (or I to his house) for us to play Borderlands (a game that is quite long) over the internet. Even if we live 100km away from each other, we can still play together. Consoles usually allow both, but as I said, I do not really need the local multiplayer option.
But all PLAYSTATION 3 consoles and all Xbox 360 consoles after the first few batches have HDMI out, and many newer LCD monitors tend to have HDMI in or DVI in with HDMI audio support.
While my monitor would does not support HDMI, I could manage to connect these consoles using HDFury or a similar device. The problem is video mode. 1080p is 1920x1080@60Hz, while my monitor would support that (though I would have to adjust the image height), a lot of LCD monitors only support lower resolutions. A lot of CRT monitors would not support that video mode either. On a PC, I can set the resolution all the way from 640x480 to 2048x1536, so I can find one that my monitor supports.
So allow me to rephrase your standard as I understand it: if you don't have at least Xbox/Wii class video hardware and comparable asset authoring capacity, keep it 2D.
Yes, good looking 2D is better than blocky 3D. A lot of adventure games have painted backgrounds and they look really good and also can be played on older computers.
Consider a choice between a video game that is free and Free [google.com] vs. one that is proprietary and $50 per player. Or a video game that runs on hardware you own vs. a video game that needs a new CPU, a new GPU, and more RAM. At what point do graphics override other considerations?
Well, I said similar games, and the price would be in the comparison. I do not care about GNU Free though, as I am not a good programmer, the source code would be useless to me.
If a video game needs a new CPU, GPU and more RAM, then I will not play it now, but when I upgrade my computer (which I would still do once in a while), I'll play it. If the game is much more expensive, then I'll probably choose the cheaper game.
It also depends on the rest of the game. System Shock 2 is great as it is (even though the human models look awful), Borderlands would not be as good without the wide open spaces, ability to see an enemy from far away and so on. I also am using a HD texture pack for Minecraft, it made the game better for me (and I would like even higher resolution textures than 32x32, but probably the game cannot handle it). Bad game with good graphics would still be a bad game, but good graphics make a good game better. I enjoyed the original Half Life, but I am really waiting for Black Mesa (Original Half Life with HL2 graphics).
But there's still a limit.
Of course there is. Just like with consoles - can you play a PS3 game on a PS2? just that with PCs it is not that clearly defined. Some new games work on older hardware some don't.
But how many people can use your PC at once? Console games are more likely to allow four players on one large monitor.
Seeing that I play single player games or multiplayer over the internet, this is not a problem for me.
On the other hand, console controllers (in my opinion) are worse than keyboard+mouse (especially for FPS games, or even point-and-click adventure). Also, it may be difficult to connect a console to a computer monitor for single player use (my CRT monitor would most likely work with this as long as the console has a VGA (or RGB) output, but most LCDs and even CRTs wouldn't).
Let me guess: you can't tolerate La Mulana, Eversion, I Wanna Be the Guy, or other games with retraux art style [tvtropes.org] that calls back to the 8-bit days.
I went found a screenshot of La Mulana. The graphics are OK, since I can still see stuff. OK, I'll try to explain: SMB style games with those graphics are OK - the graphics do not have to be 3D, but they have to be high enough resolution. On the other hand, games like NetHack are out. Also, some 3D games have very low resolution models and textures, to the point that it is difficult to recognize what I am looking at.
Still, for me to play a game with low quality graphics it has to be very good in other areas. I played The Longest Journey because I really liked the story (and tolerated the 640x480 fixed resolution), though I would have been really happy if the game had better graphics. I have also played the original GTA, but to me, GTA3 and newer are much better, in no small part due to being 3D.
Given a choice between to very similar games, I'd pick the one with better graphics.