Because people who need better than consumer level junk cannot buy quality motherboard chipsets for the athlon.
In other words, the AC was complaining about the lack of quality high end motherboard for athlon. Luckilly, with athlon 64, and Opteron, this will no longer be a problem, as the memory controller and hypertransport controller will be built onto the CPU die. This leaves only PCI/AGP bridges, and sound/video/ethernet/1394/IDE/etc. as the primary things 3rd parties will provide. This will eliminate the cause of AMD's piss poor reputation of unstable, low performing, low quality motherboard chipsets produced by shitty companies that have no quality control (VIA, ALi,SiS come to mind).
who cares if it is reversed engineered? No skin off NVIDIA's back. If they keep their licensed IP closed source, and someone reverse engineers it, that isn't NVIDIA's fault.
Unless of course, the license terms disallow opensource derravitives from the licensed work.
But I can't believe the/. crowd believes this load of BS. The drivers are closed source because most of the 3D logic and techniques are contained in SOFTWARE, not hardware, in the form of a driver that runs on NVIDIA's "3D" chip. How else do you think they got sometimes a 20% performance increase with new versions of their drivers and no chip upgrade? If they were to release the source to their video drivers, they would be giving away all their long time developed 3D techniques that makes NVIDIA _the_ 3D video card manufacturer.
Those of you who think a driver is just a layer of software that translates hardware specifics to a standard interface to the OS are fooling yourselves. The Geforce chips are not just pieces of hardware. They are practically CPU's with 3D software that runs on it. It would be equivilent to Adobe releasing the source to all their photoshop filters. It ain't going to happen.
agreed. A real conservative is not what current right wingers or republicans are mostly... I will have to agree with robert, using his form of the term. Same for so called liberals.
But if we are going to talk about the true meaning of conservative, then my points change slightly. There really aren't many real conservatives left [in US, no comment on the rest], even while looking at today's liberals. "Liberals" mostly want change, or at least change of today's policy back in line with traditional values, or change in the direction of compassion to fellow individuals. And the people that call themselves conservative today mostly want radical change in government and society and in the opposite direction from traditional.
It would be easy for most of today's "Liberals" to call themselves conservative in dictionary.com's definition of the word. The only problem is not many of them will call themselves that because of all the lunatics who claim the same name, usually because they think that religious beliefs are the only "traditional" beliefs that matter, and smaller government only means larger military and law enforcement in place of compassionate government programs and reasonable laws [read: more social control and less social help]. See all the new legislation passed by so called "conservatives" for evidence. I wouldn't consider gutting the constitution as any shape or form of traditional or conservative values. The fact that our economy is dependant on war when we are at peace is some more evidence for my point.
"It only takes 20 years for a liberal to become a conservative without changing a single idea." Robert Anton Wilson
I think robert has it backwards.. the quote implies that the world is slowly drifting to the left. In reality, it is drifting right(at least in america). if you could go backwards in time, he might have a point.
But then you have to take apart what liberal and conservative really mean. Some would say it means large vs small government. Others think it means government protection of human rights vs. government protection of coporate rights. Still others think it means strong yet unintrusive military vs. huge intrusive military. IMO, all this is wrong, but it is how the words are currently used. A better set of words could be compassion vs. compassionless, or maybe intelligence vs. ignorance (there is a fine line between the 4. that line is hard to determine for most people).
Drive down to the ISP and fill out a paper form insted. Calling on the telephone is the lamest thing to do to file a complaint. It all goes into the bitbucket. That's how the world works and especially how incompetant businesses work.
Not only this, but since the cable company laid their wires without any government compensation, solely off profit they made from offering the service, I do not see why the cable lines should be taxed to subsudize their COMPETITOR's lines laying to the same house. That is just fucking ridiculous.
Telephone lines are taxed because that exact tax is what brought most people that line. If cable lines are taxed as well, they should be able to draw from the same fund as the fucking monopoly bells that lay their wire for free in the wilderness using taxpayers money.
I'm not defending cable companies. They could be considered just as evil. But the fact that I cannot get digital internet cable is because they don't get money from this fund. But the fact that I CAN get DSL internet for roughly 4 times the price as my local cable provider charges 1 block away from me really fucking pisses me off. Especially when I know my lines these $65/month lines were laid with taxpayers money, and the $30/month (and higher bandwidth, lower pings, better QoS) lines are laid using nothing but profit of the cable co.
Telephone companies already have an unfair advantage over cable companies in the internet arena (at least in my state: arkansas) And taxing cable another 10% to give the money back to the bells is just the most insane idea I have ever heard (SBC owns all the last mile telephone lines in arkansas, and paid practically nothing for them compared to the cable companies, who paid full price for their higher quality lines, and are able to charge LESS for service)
Your forgetting the biggest point of AS: EOL of 5 years, not 1 year.
You see, when RH decides to EOL their consumer product 1 year after its release, you will no longer find security updates on redhat's website. In fact, there will be no security updates. you must manually go in and find the bugs yourself, and patch them yourself, with code you wrote yourself (unles someone else does it for you for free).
So if you want free security updates, get debian. If you just don't want security updates after 1 year, get redhat consumer. If you want security updates for 5 years and want to pay for them, and pay for QA, then get redhat AS.
All the points you mention in your post are secondary to this main reason for the very existance of RH AS.
"If you need vendor support from Oracle and/or Red Hat, I'd go with AS. Otherwise, the standard releases should work fine for you"
The standard release should work fine for you... for a year. Then when EOL comes out, BOOM. No security updates. No bug fixes. No patches. No nothing. YOU must write your own redhat specific security patches (if there are any to be written) If someone finds a redhat 8.0 vulnerability and posts it to bugtraq, YOU have to figure out how to secur that specific area of RH 8.0 OS. You can't just download the latest bugfixed RPM's off redhat's website. Because there ARE NONE.
And this is the very point of AS. To get the ppl that need longer than 1 year lifetimes out of an Operating System to pay the cost of maintenance development.
You see, if you are a software vendor, you can only maintain a version of your software for so long at such a price. If you sell 20 million copies of RHL 8.0, that only covers the cost to maintain RHL 8.0 for a predetermined period of time. Unless you charge more, then you can extend the lifetime of 8.0. This is exactly the whole point of RHL AS.
(it amazes me that nobody has yet to read what AS really means off redhat's website, and are still commenting that it is about technical support and finger pointing. Which it isn't)
the enterprise version has NOTHING to do with "shit hitting the fan" and EVERYTHING to do with EOL cycles.
with the consumer RHL, you get an EOL of 1 year, and after 1 year, its GUARANTEED to hit the fan, because there are no more security fixes put out by redhat, meaning you have to search the internet for security updates to individual packages you may or may not be running...
With the enterprise version, you have an EOL of 5 years. This means you will be notified about security bugs and given patches when they are available. For five years.
So it all depends on when you want shit to hit the fan. If you want shit to hit the fan in 1 year, then use consumer products. If you want shit to hit the fan in 5 years, use enterprise version..
But rest assured, the "support" you are talking about isn't redhat fixing your crashed server, or having someone to blame if it fucks up. It is in the form of 5 years worth of security patches and updates to the product which are not available under RH consumer products.
I can guarantee you that if your server gets hacked and its running an unsupported version of RHL in which they stoped releasing security updates for, and your boss finds out, you WILL be fired. Period.
I think its funny people are still submitting comments with specific version recommendations when the real question of the article is asking Enterprise vs Standard.
If you need a 5 year product life, and not a 1 year product life, you go with enterprise. It's as simple as that. Talking about specific versions of some other unrelated 1 yaer EOL OS to install on mission critical servers that will be there for longer than a year is just the silliest thing I have ever heard in my life.
No, its not a pipe dream. It means redhat will stop making patches for their customers when that product is EOL. In other words, If you run redhat 7.3 after the end of this year, you will have to keep track of securit updates yourself. you will have to write security patches yourself. If there is a redhat specific security problem, you will have to fix it yourself. Redhat will not release a patch for a redhat specific bug after EOL. That is the point. They charge less for products that they don't have to support for more than a year. They charge more for products that need security patches and updates for more than a year. End of story. period.
Nobody is stopping you from patching an old version of redhat yourself, except yourself, or your boss. If you need tested patches on production systems, your best place to get them is the vendor of the software. In this case, that is Redhat, not some 3rd party that never tested their software on redhat v6.x or whatever else unsupported software that nobody uses anymore. Why would they? And why would any sane administrator go hacking away at a production system just to "save" a couple bucks by going with a consumer level OS?
I'll tell you what, you go change your servers over to redhat 9, and after redhat EOL's it, and you can't upgrade it to RH 10.0 (or whatever their new version will be by then) because of proprietary software written for redhat 9, and you start finding security bugs specific to redhat 9 after the EOL, and you go and hack a patch up for it, and roll it out to your servers, and it crashes or gets hacked, unles your boss is a total fucking idiot, you will be fired.
I can just hear him say it now... "WHY THE FUCK ARE WE RUNNING AN UNSUPPORTED VERSION OF REDHAT WHEN WE COULD BE USING A FIVE YEAR EOL VERSION INSTED??????"
Just a thought, Maybe you didn't mean enterprise level. maybe you just ment home PC level, in which case, you would be right, redhat 9 will do you just fine.
"the Red Hat layer of obfuscation over the text files in/etc,"
"the more sense it makes to go with LFS, Gentoo or perhaps Debian."
I don't know about gentoo or LFS (which btw aren't tested extensively for your configuration) but debian also obfuscates the text files in/etc. They do this to allow for the possibility of a configuration tool to help you configure them. Editing the files by hand are still useful, but if it can be done by a program, then that capability should be there.. and in RedHat, and Debian both, it is. The old way of managing/etc files is simple, and works (and has worked for decades). The new layers just try to allow more flexability while maintaining backwards compatibility with the old way. If you don't like the new way, then I suggest either start liking it, or start using something else.
Maybe one day there will be standard formats to config files that are machine editable insted of a buncha shit randomly thrown into a file that configures a specific package in a random spagetti like way depending on which distribution you use.
The old way of configuring the same package differently depending on your distribution needs to end. And this extra proprietary "layer" on top are steps in the right direction. All we need to do now is make these layers un-proprietary;)
The point is that there are no security updates put out by redhat.
So if its a redhat specific security problem, and your software is EOL, your SOL unless you RTFM and RTFS and use VI to write a security patch to compile in GCC.
"Um, this has *got* to be a troll. First off, any company that doesn't chase and fix bugs should (and will) go out of business."
This is where you don't understand the differences between their "Enterprise" and "standard" editions...
First of all, microsoft doesn't chase and fix bugs, and they are not out of business. In fact, they are the most profiting company in this half of the world (probably the whole world).
Secondly, That is what RedHat means about a "5 year product life" It means that the "Enterprise" edition of their software will be supported for five years, as opposed to RedHat Linux 9.0's support which will last maybe 1.5 years if your lucky.
The point is that if you are a hobbiest, you will want the latest and greatest version of linux. And you will be forced to upgrade to the latest and greatest version if you want support (Read: patches and updates to the software). If you want support (Patches and updates) for more than a typical hobbiest needs, then you need to go with the "Enterprise" version, which will be officially supported by redhat with updates and security patches for at minimum of 5 years.
If you don't need the telephone support and just need updates and patches, I suggest bypassing redhat's services altogether and going with Debian Linux, which has simlar long term support networks in place by default.
The bottom line is that if you go with the "enterprise" version, RedHat will train and maintain a technical support staff that is capeable of troubleshooting your version of Enterprise Linux for five years. They will also release security updates to your version for five years. If you don't go enterprise, no matter what kind of support services you need, your version of Redhat will be defunct (read: no more security updates on unsupported versions) in probably less than a year. This is not good for enterprise servers that don't need to be upgraded to the latest and buggiest software every 9 months.
Yea... So then if the GPL isn't enforceable, then you have no legal right to download or copy my code and are in violation. You now owe me $25,000 dollars in copyright infringement damages.
Get the picture yet? If the GPL isn't "Enforceable" then whoever is defending themselves in court has no rights to the code in the first place. hence, why nobody is going to take this story to a judge. Because either way, they lose.
I'd be willing to bet you don't know what their 10K looks like...
Just an assumption tho... Either way, wether it reported as revenue through the Netscape division or not, the company knows the division is a leverage tool. That is why they bought it in the first place... Just because their lawsuit is over doesn't mean that Netscape has no more leverage. If anything, Netscape has given them more leverage over the years than they could have possibly imagined.
Seven Hundred Fifty Million Dollars plus 7 year license for IE, plus licensing for Windows Media Player and formats and codecs and protocols isn't enough revenue for the Netscape division to be self sufficient?
Unles you can show me some numbers that prove the Netscape division costs more than that, I suggest you come up with a different argument.
Sorry about that, I didn't fully grasp that quote at first...
But now I am reminded of that game "Tank Plus" for atari 2600.. There was an invisible mode where the only time you were visable was when you were firing your cannon, or moving (making noise)...
They could use this method to determine if the server should send coordinates of the player in the dark room to the client or not... If he is making footsteps, send the coordinates (because they could be calcuated anyway by the stero audio signal by a program). If he is firing his weapon at something, send the coordinates.. If he isn't, then the player shouldn't see him, and thus, coordinates should not be sent. If you are running into the enemy because you don't see him, then send the coordinates, because IRL, you would know there is a body in front of you. To the client who is cheating, the enemy would just disappear (or never appear if he isn't making noise/shooting). To a normal player, it would be just like normal...
There are a lot of rules that would need to be made that the server should follow in order to keep the client informed, but not too informed. These techniques could be developed as time goes on, and with each release of a new engine, they are improved upon. It's easy to calculate wether or not to show the coordinates of a player to the client or not. It's not very easy to render the entire scene for them, compress it into an MPEG video stream, and send it over a low bandwidth/high latency internet link such as any current or near future consumer broadband link.
I'm not trying to argue here. After exploring the topic more, I realise that encrpytion isn't the only answer... But it is one answer. Just like viruses/anti-viruses, it's an arms race, and the only way to guarantee to not get a virus is to be smart as to who you trust while using your computer/having sex. I believe the same is true when you play multiplayer internet video games(FPS or 3D or not). But I still don't buy the server side rendering model. I don't believe that it can solve any problem that can't be solved with current technology and a few simple principles as outlined above.
As far as modifying display settings to light things up... Quake3 players are religious when you try to take away their rendering variables... I don't like the fact that vertex lighting (which is slower than lightmap lighting on modern cards) is an available option. But anybody who plays on hardcore servers has to enable vertex lighting if they want to be able to compete. I think that sux, and Doom3 is going to let the server admins lock rendering variables for the clients to not allow them to "brighten" their screen.. (all assuming they are using a legit client). But mostly when people brighten their displays with render variables, people don't consider them cheaters because everyone else already does it. They just want to even the playing field for people that don't want to look at total ugly video game to be able to compete with the "best" players, and provide a better gaming experience...
Of course, if these tweakers want to still tweak, nobody will stop them from creating a doom3 mod that will allow them to tweak their video settings...
You don't seem to understand... Proxy bots could still be written for server side rendered gaming... They can intercept the video stream, detect exactly where the enemy is on the screen, and automatically aim for you. This is EXACTLY how proxy aim bots currently work, except insted of using a graphic to determing the position of the enemy, they use coordinates the server sends it...
If there were a dumb terminal displaying the video game, you could still write a program that captures the terminal's video, detects the head of the enemy, and alters your cursor to hover over the head... This method will not stop aim bots...
Also, there is no reason that servers cannot limit the amount of information they send to the client. If an enemy is behind a wall, the server shouldn't send their coordinates to the client. It currently is NOT possible to server side render 3D FPS video games, and it isn't going to happen any time in the near future. Even if the bandwidth were available (such as at a LAN) there is no such thing as hardware that can render several clients in real time. It is beyond real hardware that is made in even today's supercomputers.
Like I said, Quake3 and other FPS's could fix their protocols to only tell you about players that should be visible to the client, but currently they do not because of prediction reasons (which is good, because without prediction, you would really notice that 100ms lag to the extent that the game would NOT be playable). I read somewhere that the doom3 multiplayer engine will support this, and that would stop most radar and see through walls video card driver cheats. There is no need for any server to EVER do the rendering for you. It is a misconception that has been talked about, and never acted upon in the history of internet gaming (and likely won't happen for a long LONG time, if ever.) So the solutions that are readilly available, and would stop most of today's cheating, exists in packet encrpytion (to eliminate proxy bots), servers only sending relevant data to the client, trusted clients, and last but most importantly, trusted players.
Here is a scenerio that would allow you to trust the client: the server could send a request for an encrypted checksum that only an algorithm behind encrypted code can determine. If (or more likely when) someone breaks the encryption, or reverse engineers the algorithm, update the clients in a security patch. Rinse, repeate.
From my conclusions above, I don't believe it is possible to stop aim bots 100%, even with server side rendering. But it is possible to stop see through walls cheats, and radar cheats, even if you can't trust the client, using today's standard network protocols and client side rendering.
Finally, as I said before, the only solution to stop all bots 100% of the time is to trust the people behind each keyboard that is in the video game.
The server does a checksum of the client before allowing it to connect... I don't know easy it would be to fake the checksum, but apparently it is pretty secure code. But I don't really know the details...
"Better would be to not let the client know anything that it shouldn't know: transmit all keystrokes/mouse movements to the server, and then have the server send the screen to display back"
This is essentially how current games work. But instead of sending 30 bitmaps per second back in a video stream, they send 30 descriptions of the state of the game back per second, and allow the video card to render it.
The problem isn't with trusted clients. They already weed out any modded clients using the pure server mode, which requires the client to send checksums of itself to the server on connect. The problem is mainly that a proxy is ALTERING the keystroke patterns that is sent to the server. The proxy bots will keep track of where players are using the information the server sends the client (Note: depending on the game engine, the server might transmit every player's info to the client, or just players that are near his visible field of view, Massive Multi Player [MMP] games are usually pretty good about not telling you about people you shouldn't know about for obvious bandwidth reasons, but games like Q3, etc tell you about everyone to keep client side prediction code working good, and this is one reason why Q3 based games don't scale well past ~32 players/server). When you press the fire button, the packet that contains that command gets altered to include a movement command that will aim the cursor before the fire command.
A proxy bot could still intercept screen captures and aim the cursor for you, even if your server side render method were possible within the next 10 years (which it isn't). In fact, there are other bots that are standalone programs that do screen captures, finds the head of the enemy, and aims your mouse cursor to it when you want to fire. All this without using a proxy, or altered gamecode (note: this method is not solved via encrypted packets, but it is much more difficult to implement than proxy bots, and very uncommon from my experience).
We need a combination of encrypted data packets, a good way to trust clients (like checksums, AOL does it with their IM client even), and a good community of players that you know and trust to not cheat.
I play FPS's a lot. I run into bots rarely. Usually because I know most of the players I play with.
Slashdot Mirror
no...
Because people who need better than consumer level junk cannot buy quality motherboard chipsets for the athlon.
In other words, the AC was complaining about the lack of quality high end motherboard for athlon. Luckilly, with athlon 64, and Opteron, this will no longer be a problem, as the memory controller and hypertransport controller will be built onto the CPU die. This leaves only PCI/AGP bridges, and sound/video/ethernet/1394/IDE/etc. as the primary things 3rd parties will provide. This will eliminate the cause of AMD's piss poor reputation of unstable, low performing, low quality motherboard chipsets produced by shitty companies that have no quality control (VIA, ALi,SiS come to mind).
who cares if it is reversed engineered? No skin off NVIDIA's back. If they keep their licensed IP closed source, and someone reverse engineers it, that isn't NVIDIA's fault.
/. crowd believes this load of BS. The drivers are closed source because most of the 3D logic and techniques are contained in SOFTWARE, not hardware, in the form of a driver that runs on NVIDIA's "3D" chip. How else do you think they got sometimes a 20% performance increase with new versions of their drivers and no chip upgrade? If they were to release the source to their video drivers, they would be giving away all their long time developed 3D techniques that makes NVIDIA _the_ 3D video card manufacturer.
Unless of course, the license terms disallow opensource derravitives from the licensed work.
But I can't believe the
Those of you who think a driver is just a layer of software that translates hardware specifics to a standard interface to the OS are fooling yourselves. The Geforce chips are not just pieces of hardware. They are practically CPU's with 3D software that runs on it. It would be equivilent to Adobe releasing the source to all their photoshop filters. It ain't going to happen.
agreed. A real conservative is not what current right wingers or republicans are mostly... I will have to agree with robert, using his form of the term. Same for so called liberals.
;)
But if we are going to talk about the true meaning of conservative, then my points change slightly. There really aren't many real conservatives left [in US, no comment on the rest], even while looking at today's liberals. "Liberals" mostly want change, or at least change of today's policy back in line with traditional values, or change in the direction of compassion to fellow individuals. And the people that call themselves conservative today mostly want radical change in government and society and in the opposite direction from traditional.
It would be easy for most of today's "Liberals" to call themselves conservative in dictionary.com's definition of the word. The only problem is not many of them will call themselves that because of all the lunatics who claim the same name, usually because they think that religious beliefs are the only "traditional" beliefs that matter, and smaller government only means larger military and law enforcement in place of compassionate government programs and reasonable laws [read: more social control and less social help]. See all the new legislation passed by so called "conservatives" for evidence. I wouldn't consider gutting the constitution as any shape or form of traditional or conservative values. The fact that our economy is dependant on war when we are at peace is some more evidence for my point.
all this IMHO of course
"It only takes 20 years for a liberal to become a conservative without changing a single idea." Robert Anton Wilson
I think robert has it backwards.. the quote implies that the world is slowly drifting to the left. In reality, it is drifting right(at least in america). if you could go backwards in time, he might have a point.
But then you have to take apart what liberal and conservative really mean. Some would say it means large vs small government. Others think it means government protection of human rights vs. government protection of coporate rights. Still others think it means strong yet unintrusive military vs. huge intrusive military. IMO, all this is wrong, but it is how the words are currently used. A better set of words could be compassion vs. compassionless, or maybe intelligence vs. ignorance (there is a fine line between the 4. that line is hard to determine for most people).
"C) His ISP's DNS is fine, this is the result of spyware on his -- yes, the submitter's -- computer."
are you a wacko?
read the article, it clearly states his ISP is the one with the hacked DNS server.
If he got a trojan, why would he be calling the fucking FBI over it? That would be the lamest thing I have heard in the history of reasoning.
"uhh, FBI, someone infected me with code red, can you investigate?????" give me a break.
Drive down to the ISP and fill out a paper form insted. Calling on the telephone is the lamest thing to do to file a complaint. It all goes into the bitbucket. That's how the world works and especially how incompetant businesses work.
Yes, but in context, my quote makes much more sence...
"'Are they going to stop chasing bugs in the consumer division because of the obvious conflict of interest with their revenue stream selling support?'
Um, this has *got* to be a troll. First off, any company that doesn't chase and fix bugs should (and will) go out of business."
My post was in the same context of what is described above in the origional post.
But I do agree with you on those points outside of this particular context.
Not only this, but since the cable company laid their wires without any government compensation, solely off profit they made from offering the service, I do not see why the cable lines should be taxed to subsudize their COMPETITOR's lines laying to the same house. That is just fucking ridiculous.
Telephone lines are taxed because that exact tax is what brought most people that line. If cable lines are taxed as well, they should be able to draw from the same fund as the fucking monopoly bells that lay their wire for free in the wilderness using taxpayers money.
I'm not defending cable companies. They could be considered just as evil. But the fact that I cannot get digital internet cable is because they don't get money from this fund. But the fact that I CAN get DSL internet for roughly 4 times the price as my local cable provider charges 1 block away from me really fucking pisses me off. Especially when I know my lines these $65/month lines were laid with taxpayers money, and the $30/month (and higher bandwidth, lower pings, better QoS) lines are laid using nothing but profit of the cable co.
Telephone companies already have an unfair advantage over cable companies in the internet arena (at least in my state: arkansas) And taxing cable another 10% to give the money back to the bells is just the most insane idea I have ever heard (SBC owns all the last mile telephone lines in arkansas, and paid practically nothing for them compared to the cable companies, who paid full price for their higher quality lines, and are able to charge LESS for service)
ummm.. actually, it does.
Your forgetting the biggest point of AS: EOL of 5 years, not 1 year.
You see, when RH decides to EOL their consumer product 1 year after its release, you will no longer find security updates on redhat's website. In fact, there will be no security updates. you must manually go in and find the bugs yourself, and patch them yourself, with code you wrote yourself (unles someone else does it for you for free).
So if you want free security updates, get debian. If you just don't want security updates after 1 year, get redhat consumer. If you want security updates for 5 years and want to pay for them, and pay for QA, then get redhat AS.
All the points you mention in your post are secondary to this main reason for the very existance of RH AS.
"If you need vendor support from Oracle and/or Red Hat, I'd go with AS. Otherwise, the standard releases should work fine for you"
The standard release should work fine for you... for a year. Then when EOL comes out, BOOM. No security updates. No bug fixes. No patches. No nothing. YOU must write your own redhat specific security patches (if there are any to be written) If someone finds a redhat 8.0 vulnerability and posts it to bugtraq, YOU have to figure out how to secur that specific area of RH 8.0 OS. You can't just download the latest bugfixed RPM's off redhat's website. Because there ARE NONE.
And this is the very point of AS. To get the ppl that need longer than 1 year lifetimes out of an Operating System to pay the cost of maintenance development.
You see, if you are a software vendor, you can only maintain a version of your software for so long at such a price. If you sell 20 million copies of RHL 8.0, that only covers the cost to maintain RHL 8.0 for a predetermined period of time. Unless you charge more, then you can extend the lifetime of 8.0. This is exactly the whole point of RHL AS.
(it amazes me that nobody has yet to read what AS really means off redhat's website, and are still commenting that it is about technical support and finger pointing. Which it isn't)
the enterprise version has NOTHING to do with "shit hitting the fan" and EVERYTHING to do with EOL cycles.
with the consumer RHL, you get an EOL of 1 year, and after 1 year, its GUARANTEED to hit the fan, because there are no more security fixes put out by redhat, meaning you have to search the internet for security updates to individual packages you may or may not be running...
With the enterprise version, you have an EOL of 5 years. This means you will be notified about security bugs and given patches when they are available. For five years.
So it all depends on when you want shit to hit the fan. If you want shit to hit the fan in 1 year, then use consumer products. If you want shit to hit the fan in 5 years, use enterprise version..
But rest assured, the "support" you are talking about isn't redhat fixing your crashed server, or having someone to blame if it fucks up. It is in the form of 5 years worth of security patches and updates to the product which are not available under RH consumer products.
I can guarantee you that if your server gets hacked and its running an unsupported version of RHL in which they stoped releasing security updates for, and your boss finds out, you WILL be fired. Period.
I think its funny people are still submitting comments with specific version recommendations when the real question of the article is asking Enterprise vs Standard.
If you need a 5 year product life, and not a 1 year product life, you go with enterprise. It's as simple as that. Talking about specific versions of some other unrelated 1 yaer EOL OS to install on mission critical servers that will be there for longer than a year is just the silliest thing I have ever heard in my life.
No, its not a pipe dream. It means redhat will stop making patches for their customers when that product is EOL. In other words, If you run redhat 7.3 after the end of this year, you will have to keep track of securit updates yourself. you will have to write security patches yourself. If there is a redhat specific security problem, you will have to fix it yourself. Redhat will not release a patch for a redhat specific bug after EOL. That is the point. They charge less for products that they don't have to support for more than a year. They charge more for products that need security patches and updates for more than a year. End of story. period.
Nobody is stopping you from patching an old version of redhat yourself, except yourself, or your boss. If you need tested patches on production systems, your best place to get them is the vendor of the software. In this case, that is Redhat, not some 3rd party that never tested their software on redhat v6.x or whatever else unsupported software that nobody uses anymore. Why would they? And why would any sane administrator go hacking away at a production system just to "save" a couple bucks by going with a consumer level OS?
I'll tell you what, you go change your servers over to redhat 9, and after redhat EOL's it, and you can't upgrade it to RH 10.0 (or whatever their new version will be by then) because of proprietary software written for redhat 9, and you start finding security bugs specific to redhat 9 after the EOL, and you go and hack a patch up for it, and roll it out to your servers, and it crashes or gets hacked, unles your boss is a total fucking idiot, you will be fired.
I can just hear him say it now... "WHY THE FUCK ARE WE RUNNING AN UNSUPPORTED VERSION OF REDHAT WHEN WE COULD BE USING A FIVE YEAR EOL VERSION INSTED??????"
Just a thought, Maybe you didn't mean enterprise level. maybe you just ment home PC level, in which case, you would be right, redhat 9 will do you just fine.
"the Red Hat layer of obfuscation over the text files in /etc,"
/etc. They do this to allow for the possibility of a configuration tool to help you configure them. Editing the files by hand are still useful, but if it can be done by a program, then that capability should be there.. and in RedHat, and Debian both, it is. The old way of managing /etc files is simple, and works (and has worked for decades). The new layers just try to allow more flexability while maintaining backwards compatibility with the old way. If you don't like the new way, then I suggest either start liking it, or start using something else.
;)
"the more sense it makes to go with LFS, Gentoo or perhaps Debian."
I don't know about gentoo or LFS (which btw aren't tested extensively for your configuration) but debian also obfuscates the text files in
Maybe one day there will be standard formats to config files that are machine editable insted of a buncha shit randomly thrown into a file that configures a specific package in a random spagetti like way depending on which distribution you use.
The old way of configuring the same package differently depending on your distribution needs to end. And this extra proprietary "layer" on top are steps in the right direction. All we need to do now is make these layers un-proprietary
You can't do this with debian either. The kernel is updated in major releases, so you must reboot to change the kernel.
I guess you were sleeping when you wrote that comment tho, so I won't hold it against you...
The point is that there are no security updates put out by redhat.
So if its a redhat specific security problem, and your software is EOL, your SOL unless you RTFM and RTFS and use VI to write a security patch to compile in GCC.
"Um, this has *got* to be a troll. First off, any company that doesn't chase and fix bugs should (and will) go out of business."
This is where you don't understand the differences between their "Enterprise" and "standard" editions...
First of all, microsoft doesn't chase and fix bugs, and they are not out of business. In fact, they are the most profiting company in this half of the world (probably the whole world).
Secondly, That is what RedHat means about a "5 year product life" It means that the "Enterprise" edition of their software will be supported for five years, as opposed to RedHat Linux 9.0's support which will last maybe 1.5 years if your lucky.
The point is that if you are a hobbiest, you will want the latest and greatest version of linux. And you will be forced to upgrade to the latest and greatest version if you want support (Read: patches and updates to the software). If you want support (Patches and updates) for more than a typical hobbiest needs, then you need to go with the "Enterprise" version, which will be officially supported by redhat with updates and security patches for at minimum of 5 years.
If you don't need the telephone support and just need updates and patches, I suggest bypassing redhat's services altogether and going with Debian Linux, which has simlar long term support networks in place by default.
The bottom line is that if you go with the "enterprise" version, RedHat will train and maintain a technical support staff that is capeable of troubleshooting your version of Enterprise Linux for five years. They will also release security updates to your version for five years. If you don't go enterprise, no matter what kind of support services you need, your version of Redhat will be defunct (read: no more security updates on unsupported versions) in probably less than a year. This is not good for enterprise servers that don't need to be upgraded to the latest and buggiest software every 9 months.
Yea... So then if the GPL isn't enforceable, then you have no legal right to download or copy my code and are in violation. You now owe me $25,000 dollars in copyright infringement damages.
Get the picture yet? If the GPL isn't "Enforceable" then whoever is defending themselves in court has no rights to the code in the first place. hence, why nobody is going to take this story to a judge. Because either way, they lose.
I'd be willing to bet you don't know what their 10K looks like...
Just an assumption tho... Either way, wether it reported as revenue through the Netscape division or not, the company knows the division is a leverage tool. That is why they bought it in the first place... Just because their lawsuit is over doesn't mean that Netscape has no more leverage. If anything, Netscape has given them more leverage over the years than they could have possibly imagined.
Seven Hundred Fifty Million Dollars plus 7 year license for IE, plus licensing for Windows Media Player and formats and codecs and protocols isn't enough revenue for the Netscape division to be self sufficient?
Unles you can show me some numbers that prove the Netscape division costs more than that, I suggest you come up with a different argument.
"To quote myself..."
Sorry about that, I didn't fully grasp that quote at first...
But now I am reminded of that game "Tank Plus" for atari 2600.. There was an invisible mode where the only time you were visable was when you were firing your cannon, or moving (making noise)...
They could use this method to determine if the server should send coordinates of the player in the dark room to the client or not... If he is making footsteps, send the coordinates (because they could be calcuated anyway by the stero audio signal by a program). If he is firing his weapon at something, send the coordinates.. If he isn't, then the player shouldn't see him, and thus, coordinates should not be sent. If you are running into the enemy because you don't see him, then send the coordinates, because IRL, you would know there is a body in front of you. To the client who is cheating, the enemy would just disappear (or never appear if he isn't making noise/shooting). To a normal player, it would be just like normal...
There are a lot of rules that would need to be made that the server should follow in order to keep the client informed, but not too informed. These techniques could be developed as time goes on, and with each release of a new engine, they are improved upon. It's easy to calculate wether or not to show the coordinates of a player to the client or not. It's not very easy to render the entire scene for them, compress it into an MPEG video stream, and send it over a low bandwidth/high latency internet link such as any current or near future consumer broadband link.
I'm not trying to argue here. After exploring the topic more, I realise that encrpytion isn't the only answer... But it is one answer. Just like viruses/anti-viruses, it's an arms race, and the only way to guarantee to not get a virus is to be smart as to who you trust while using your computer/having sex. I believe the same is true when you play multiplayer internet video games(FPS or 3D or not). But I still don't buy the server side rendering model. I don't believe that it can solve any problem that can't be solved with current technology and a few simple principles as outlined above.
As far as modifying display settings to light things up... Quake3 players are religious when you try to take away their rendering variables... I don't like the fact that vertex lighting (which is slower than lightmap lighting on modern cards) is an available option. But anybody who plays on hardcore servers has to enable vertex lighting if they want to be able to compete. I think that sux, and Doom3 is going to let the server admins lock rendering variables for the clients to not allow them to "brighten" their screen.. (all assuming they are using a legit client). But mostly when people brighten their displays with render variables, people don't consider them cheaters because everyone else already does it. They just want to even the playing field for people that don't want to look at total ugly video game to be able to compete with the "best" players, and provide a better gaming experience...
Of course, if these tweakers want to still tweak, nobody will stop them from creating a doom3 mod that will allow them to tweak their video settings...
You don't seem to understand... Proxy bots could still be written for server side rendered gaming... They can intercept the video stream, detect exactly where the enemy is on the screen, and automatically aim for you. This is EXACTLY how proxy aim bots currently work, except insted of using a graphic to determing the position of the enemy, they use coordinates the server sends it...
If there were a dumb terminal displaying the video game, you could still write a program that captures the terminal's video, detects the head of the enemy, and alters your cursor to hover over the head... This method will not stop aim bots...
Also, there is no reason that servers cannot limit the amount of information they send to the client. If an enemy is behind a wall, the server shouldn't send their coordinates to the client. It currently is NOT possible to server side render 3D FPS video games, and it isn't going to happen any time in the near future. Even if the bandwidth were available (such as at a LAN) there is no such thing as hardware that can render several clients in real time. It is beyond real hardware that is made in even today's supercomputers.
Like I said, Quake3 and other FPS's could fix their protocols to only tell you about players that should be visible to the client, but currently they do not because of prediction reasons (which is good, because without prediction, you would really notice that 100ms lag to the extent that the game would NOT be playable). I read somewhere that the doom3 multiplayer engine will support this, and that would stop most radar and see through walls video card driver cheats. There is no need for any server to EVER do the rendering for you. It is a misconception that has been talked about, and never acted upon in the history of internet gaming (and likely won't happen for a long LONG time, if ever.) So the solutions that are readilly available, and would stop most of today's cheating, exists in packet encrpytion (to eliminate proxy bots), servers only sending relevant data to the client, trusted clients, and last but most importantly, trusted players.
Here is a scenerio that would allow you to trust the client: the server could send a request for an encrypted checksum that only an algorithm behind encrypted code can determine. If (or more likely when) someone breaks the encryption, or reverse engineers the algorithm, update the clients in a security patch. Rinse, repeate.
From my conclusions above, I don't believe it is possible to stop aim bots 100%, even with server side rendering. But it is possible to stop see through walls cheats, and radar cheats, even if you can't trust the client, using today's standard network protocols and client side rendering.
Finally, as I said before, the only solution to stop all bots 100% of the time is to trust the people behind each keyboard that is in the video game.
The server does a checksum of the client before allowing it to connect... I don't know easy it would be to fake the checksum, but apparently it is pretty secure code. But I don't really know the details...
"Better would be to not let the client know anything that it shouldn't know: transmit all keystrokes/mouse movements to the server, and then have the server send the screen to display back"
This is essentially how current games work. But instead of sending 30 bitmaps per second back in a video stream, they send 30 descriptions of the state of the game back per second, and allow the video card to render it.
The problem isn't with trusted clients. They already weed out any modded clients using the pure server mode, which requires the client to send checksums of itself to the server on connect. The problem is mainly that a proxy is ALTERING the keystroke patterns that is sent to the server. The proxy bots will keep track of where players are using the information the server sends the client (Note: depending on the game engine, the server might transmit every player's info to the client, or just players that are near his visible field of view, Massive Multi Player [MMP] games are usually pretty good about not telling you about people you shouldn't know about for obvious bandwidth reasons, but games like Q3, etc tell you about everyone to keep client side prediction code working good, and this is one reason why Q3 based games don't scale well past ~32 players/server). When you press the fire button, the packet that contains that command gets altered to include a movement command that will aim the cursor before the fire command.
A proxy bot could still intercept screen captures and aim the cursor for you, even if your server side render method were possible within the next 10 years (which it isn't). In fact, there are other bots that are standalone programs that do screen captures, finds the head of the enemy, and aims your mouse cursor to it when you want to fire. All this without using a proxy, or altered gamecode (note: this method is not solved via encrypted packets, but it is much more difficult to implement than proxy bots, and very uncommon from my experience).
We need a combination of encrypted data packets, a good way to trust clients (like checksums, AOL does it with their IM client even), and a good community of players that you know and trust to not cheat.
I play FPS's a lot. I run into bots rarely. Usually because I know most of the players I play with.