Yes it does play some role, not as big as it should, but a very small role.
If you really think this article is so sensationalized, go sit through two weeks of paid training on your next vacation. Looking at it from the inside, you won't think the article overly sensational anymore.
And yes I realise that two weeks of vacation is all most people get a year, and working at one of these places is hardly a vacation, but my point is from the inside it is obvious quality is of little to no concern to management.
Because you have to run the scam for at least a few months to make profits. Call logs are the metrics the whole contract hinges on. There are customer satisfation surveys, done by the client at random, if every call just automaticly hung up even after 20 minutes, the client would know whats up very quickly.
compare it to the cheating that goes on with the seti client. They can only realisticly detect cheats that do the work in a ridiculusly shot period of time. The rest, that claims reasonable time for the work done. Beyond that your left with random sampling and statistics.
If you did no work, they know what's up instantly, if you do half assed work they have to get about 3-6 months worth of poor satisfation surveys before you have a trustworthy statistical analysis. Even then they'll usually give the client a chance to improve, so as not to incur the expense of switching to another provider.
Bottom line, statistical analysis of quality lags far behind that of quantity. The client doesn't know they're getting fleeced for at least a few months, as long as you'r doing something.
Well, your wrong. I've worked in a few of these places, and for the outsourced operations this is always how it is. In-house is a completely different story, your employer wants you to fix it. Outsourcing, you're there to take calls.
The problem is, as mentioed earlier, their busisness model. No one seems to have developed a resonable and equitable way to pay outsourcers, because the per call method simply does not work.
Usually the drive to run businesses this way comes from execs with profit sharing, they get their money and leave before the client is completely pissed, and their resumes look like waves of profitability!
I worked at a company w/ an oustource division, that hired an exec that had just finished pulling this stunt on another company. When i got their, our segment was unprofitable (as support should be), in 3 months we had our first $1 million month, followed by a 2$mil and then $4mil. Folowed by a pissed client, and exec jumping ship, and the downward spiral to loosing the client and the profits.
The exec? He moved on to his next job and showed off the exponential profit growth that this company had under his fine direction!
The way windows gets around this (sometimes) is by placing the old library(dll) in the same directory as the exe calling, while the system retains it's newer version. It's a reduementary form of sandboxing, at least on the file level.
I think what is needed is similar version snadboxing for packages. When installing a new pkg.y that call for and upgrade to lib.x.5, but lib.x.5 causes a cacading dependency problem for the rest of the system, it should give you the option to install lib.x.5 and pkg.y together, without upgrading lib.x.4 that the rest of the system needs. It may need to be recompiled to use the lib.x.5 in a non-standard location, but it would help greatly.
Now, all of this may not work too well if you plan to use pkg.y to talk to pgk.z if they use different version of lib.x, but it's a start. You could allways move pkg.z into the 'sandbox' as well, I guess.
Though your point is valid, the Unreal was one of the best software renderers I've seend.
At work, I ran the Unreal Tournament on a 450 XEON w/ 256MB RAM in software mode on the crappy onboard card (S3?), and it was actually quite playable. Sure a better card would have done wonders, and I know an equivilant PII 450 would not have kept up in software rendering, but I was suprised as hell it ran so well!
As to the original question, it's really a toss up how they determine the minimum requirments, and it varies from game to game. If you don't have a fast enough cpu, it will surely install, but may not run the best. Some games will benifit more from a better GPU, some from memory (SW: Galaxies loves 1GB+), some from CPU (Quake3).
I wouldn't take SYS REQ's at face value, as long as you don't have gerneration gaps in software, a few mhz rarely makes a difference that can't be overlooked.
This would only work if they realigned the scores every year to make 100 the median, making comparing the scores from different years entirely useless.
They don't readjust that often, but they have readjusted twice since the system's inception, both 10 point moves to inflate the ever decreasing scores. So 80 (borderline retarded) on the original test now nets you 100 (average). Pretty sad huh?
As I recall, wasn't their last supposed ddos rather questionable as well? I recall being able to access ftp.sco.com (one ip away) and investor.sco.com was up with their complete website.
That time there was at least some data to suggest that an attack had taken place, but this time I really have to wonder. After all, in the lab the worm ONLY resolved www.sco.com, no attack was made regardless of the date. By pulling the dns records they elimnate all data regarding how big the attack is, or if it even took place.
It all stinks if you ask me, and I really have begun to doubt there is an attack at all.
is make the the alterations to the TOS to be clear on what the limits are, send the tos out to all customers, and give them a grace period of 60 days to terminate their contracts without penalty. That would be the right thing to do, not that they'll do it.
If they want to offer tiered level of services, they should introduce a second data class on their line, requiring different hardware and operating at a symmetrical speed, like 1.5Mb up / 1.5Mb down, or even 3u/3d, and unlimited xfers. They should be wary of overpricing such an offering, but who am I kidding, they overprice everything.
The only reason I'm with comcast is because they were offering $20/mo for 6 months, and they only reason I stay is the new 3Mb down. If they send me a letter and disconnect me, I'll just get DSL and wait for the class action suit that will inevitably follow all this.
I wonder if the problem is more with having enough nodes in a given area, and the cost of adding nodes, or if the real problem is with the amount of upstream bandwidth that they have to buy.
If the problem is not enough nodes, it's their problem. They are responsible for maintaining the network and providing adequate room for growth, and for current users.
If the problem lies in purchasing upstream bandwidth, it's still their problem, but they just exposed themselves to potentially %100 more swing in their needs by doubling their speed limits.
I think they failed to plan for this 3Mb change and are ill prepared to deal with it. If they have to scale back to deal with it then I guess that's what they have to do.
If so, then maybe look at replacing the stock computer alltogether. Haltec and Autronic make some very nice replacement computers for overclocking your car (so to speak) that have built in com ports and datalogging capabilities. Your really only need to go this route if your trying to get a lot out of your engine, but certain applications it's the only thing that will do (500+hp 1.3 liter rotaries! fun fun!!)
First of all, I have to say I don't agree with what MS is doing, however it seems perfectly within their rights to do, at least for FAT32. The FAT16 patents should be expired by now, and FAT16 is good for up to 2GB, so this should only really affect new devices >2GB. When said devices come out they probably will come in formatted and unformatted versions, so consumers can have a choice.
I haven't purchased floppies in a long time, but as I recall they came in mac, pc, or unformatted at different prices. Did they pay a liscence then, or was it just the production cost of actually formating the disk that made the price different?
In the end 25 cents passed on to me, the end user, on a device greater than 2GB seems like no big deal, especialy given the current price of such a thing. As long as they don't use the precedent to then go after other FAT driver implementations in other OS's, I can live with it.
I think MS once recieved a very large batch of new 3com cards, all with the same MAC address. Needless to say it was a big mistake, and took some time to troubleshoot. Imagine upgrading/deploying ~300 NICS, all get link but they wont talk. No matter how you troubleshoot, your just not going to figure it out until you start sniffing.
can anybody really think of anything that they could possibly do, that would both make windows faster AND make linux slower?
From my understanding it emulates a generic s3 video card, not your actual card, so the driver argument is null. I don't know how it handles this for other hardware, but I would guess the same.
Any malice towards linux, real or imagined, would not take precedence over product funtionality for their target users. If you look at what they need to maintain compatibility with (NT4 NT351 MSDOS) I can't see what they could possibly do to make windows faster, that wouldn't do the same in linux. Conversly, I can't think of anything that would slow down linux and not adversly affect their OSs.
Slashdot Mirror
Yes it does play some role, not as big as it should, but a very small role.
If you really think this article is so sensationalized, go sit through two weeks of paid training on your next vacation. Looking at it from the inside, you won't think the article overly sensational anymore.
And yes I realise that two weeks of vacation is all most people get a year, and working at one of these places is hardly a vacation, but my point is from the inside it is obvious quality is of little to no concern to management.
Because you have to run the scam for at least a few months to make profits. Call logs are the metrics the whole contract hinges on. There are customer satisfation surveys, done by the client at random, if every call just automaticly hung up even after 20 minutes, the client would know whats up very quickly.
compare it to the cheating that goes on with the seti client. They can only realisticly detect cheats that do the work in a ridiculusly shot period of time. The rest, that claims reasonable time for the work done. Beyond that your left with random sampling and statistics.
If you did no work, they know what's up instantly, if you do half assed work they have to get about 3-6 months worth of poor satisfation surveys before you have a trustworthy statistical analysis. Even then they'll usually give the client a chance to improve, so as not to incur the expense of switching to another provider.
Bottom line, statistical analysis of quality lags far behind that of quantity. The client doesn't know they're getting fleeced for at least a few months, as long as you'r doing something.
Well, your wrong. I've worked in a few of these places, and for the outsourced operations this is always how it is. In-house is a completely different story, your employer wants you to fix it. Outsourcing, you're there to take calls.
The problem is, as mentioed earlier, their busisness model. No one seems to have developed a resonable and equitable way to pay outsourcers, because the per call method simply does not work.
Usually the drive to run businesses this way comes from execs with profit sharing, they get their money and leave before the client is completely pissed, and their resumes look like waves of profitability!
I worked at a company w/ an oustource division, that hired an exec that had just finished pulling this stunt on another company. When i got their, our segment was unprofitable (as support should be), in 3 months we had our first $1 million month, followed by a 2$mil and then $4mil. Folowed by a pissed client, and exec jumping ship, and the downward spiral to loosing the client and the profits.
The exec? He moved on to his next job and showed off the exponential profit growth that this company had under his fine direction!
The way windows gets around this (sometimes) is by placing the old library(dll) in the same directory as the exe calling, while the system retains it's newer version. It's a reduementary form of sandboxing, at least on the file level.
I think what is needed is similar version snadboxing for packages. When installing a new pkg.y that call for and upgrade to lib.x.5, but lib.x.5 causes a cacading dependency problem for the rest of the system, it should give you the option to install lib.x.5 and pkg.y together, without upgrading lib.x.4 that the rest of the system needs. It may need to be recompiled to use the lib.x.5 in a non-standard location, but it would help greatly.
Now, all of this may not work too well if you plan to use pkg.y to talk to pgk.z if they use different version of lib.x, but it's a start. You could allways move pkg.z into the 'sandbox' as well, I guess.
Though your point is valid, the Unreal was one of the best software renderers I've seend.
At work, I ran the Unreal Tournament on a 450 XEON w/ 256MB RAM in software mode on the crappy onboard card (S3?), and it was actually quite playable. Sure a better card would have done wonders, and I know an equivilant PII 450 would not have kept up in software rendering, but I was suprised as hell it ran so well!
As to the original question, it's really a toss up how they determine the minimum requirments, and it varies from game to game. If you don't have a fast enough cpu, it will surely install, but may not run the best. Some games will benifit more from a better GPU, some from memory (SW: Galaxies loves 1GB+), some from CPU (Quake3).
I wouldn't take SYS REQ's at face value, as long as you don't have gerneration gaps in software, a few mhz rarely makes a difference that can't be overlooked.
This would only work if they realigned the scores every year to make 100 the median, making comparing the scores from different years entirely useless.
They don't readjust that often, but they have readjusted twice since the system's inception, both 10 point moves to inflate the ever decreasing scores. So 80 (borderline retarded) on the original test now nets you 100 (average). Pretty sad huh?
As I recall, wasn't their last supposed ddos rather questionable as well? I recall being able to access ftp.sco.com (one ip away) and investor.sco.com was up with their complete website.
That time there was at least some data to suggest that an attack had taken place, but this time I really have to wonder. After all, in the lab the worm ONLY resolved www.sco.com, no attack was made regardless of the date. By pulling the dns records they elimnate all data regarding how big the attack is, or if it even took place.
It all stinks if you ask me, and I really have begun to doubt there is an attack at all.
is make the the alterations to the TOS to be clear on what the limits are, send the tos out to all customers, and give them a grace period of 60 days to terminate their contracts without penalty. That would be the right thing to do, not that they'll do it.
If they want to offer tiered level of services, they should introduce a second data class on their line, requiring different hardware and operating at a symmetrical speed, like 1.5Mb up / 1.5Mb down, or even 3u/3d, and unlimited xfers. They should be wary of overpricing such an offering, but who am I kidding, they overprice everything.
The only reason I'm with comcast is because they were offering $20/mo for 6 months, and they only reason I stay is the new 3Mb down. If they send me a letter and disconnect me, I'll just get DSL and wait for the class action suit that will inevitably follow all this.
I wonder if the problem is more with having enough nodes in a given area, and the cost of adding nodes, or if the real problem is with the amount of upstream bandwidth that they have to buy.
If the problem is not enough nodes, it's their problem. They are responsible for maintaining the network and providing adequate room for growth, and for current users.
If the problem lies in purchasing upstream bandwidth, it's still their problem, but they just exposed themselves to potentially %100 more swing in their needs by doubling their speed limits.
I think they failed to plan for this 3Mb change and are ill prepared to deal with it. If they have to scale back to deal with it then I guess that's what they have to do.
If so, then maybe look at replacing the stock computer alltogether. Haltec and Autronic make some very nice replacement computers for overclocking your car (so to speak) that have built in com ports and datalogging capabilities. Your really only need to go this route if your trying to get a lot out of your engine, but certain applications it's the only thing that will do (500+hp 1.3 liter rotaries! fun fun!!)
First of all, I have to say I don't agree with what MS is doing, however it seems perfectly within their rights to do, at least for FAT32. The FAT16 patents should be expired by now, and FAT16 is good for up to 2GB, so this should only really affect new devices >2GB. When said devices come out they probably will come in formatted and unformatted versions, so consumers can have a choice. I haven't purchased floppies in a long time, but as I recall they came in mac, pc, or unformatted at different prices. Did they pay a liscence then, or was it just the production cost of actually formating the disk that made the price different? In the end 25 cents passed on to me, the end user, on a device greater than 2GB seems like no big deal, especialy given the current price of such a thing. As long as they don't use the precedent to then go after other FAT driver implementations in other OS's, I can live with it.
I think MS once recieved a very large batch of new 3com cards, all with the same MAC address. Needless to say it was a big mistake, and took some time to troubleshoot. Imagine upgrading/deploying ~300 NICS, all get link but they wont talk. No matter how you troubleshoot, your just not going to figure it out until you start sniffing.
can anybody really think of anything that they could possibly do, that would both make windows faster AND make linux slower?
From my understanding it emulates a generic s3 video card, not your actual card, so the driver argument is null. I don't know how it handles this for other hardware, but I would guess the same.
Any malice towards linux, real or imagined, would not take precedence over product funtionality for their target users. If you look at what they need to maintain compatibility with (NT4 NT351 MSDOS) I can't see what they could possibly do to make windows faster, that wouldn't do the same in linux. Conversly, I can't think of anything that would slow down linux and not adversly affect their OSs.