I think it's just around the corner. The 'boomers are about five years away from developing prostate problems on a massive scale, and the ladies will have issues with incontinence too, they'll all be pumped-full of prescription drugs that keep them hopping up to piss. As soon as complaints about this start showing up, the producers are going to start making a scene in the movie that's a nice segue to an intermission, just like on TV. Thos of us who can hold our bladders will likely have to sit through commercials asking us to take the opportunity to refill our beverages and popcorn for a modest fee, and to check out next year's Ford lineup.
The funny thing is that cable TV was originally commercial-free, you PAID to not watch advertisements. I remember those days faintly, one of my neighbors had cable and it was quite a hoot at cookouts and block parties.
Somehow cable became so common and people became so passive that cable now has just as much advertising as broadcast, and the quality of the ads and programming is generally lower on cable.
So now we pay the content providers to watch the content, and the advertisers pay them to slip us ads. We even get advertised to when paying the ultimate in high-prices at the theaters. I think that in a decade's time you'll see movies with one or two commercial-filled 'intermissions' under the pretense of letting elderly folks use the potty. Just watch.
Taking an existing drive design and reducing it's capacity doesn't reduce cost at all.
As far as I know, adding capacity to an existing design by adding heads and platters costs VERY little. Hard drive manufacturers make sure they break-even on the lowest-end drive in the lineup, and the rest is real money.
The cost of making a 60GB drive is almost the same as making a 250GB drive with the same technology. The industry expenses are in the R&D for pushing the data density per-platter up.
All the components are (essentially) the same, except the hard drive. The hard drive for the iPod is a 1.8" part, and the mini uses an even smaller 'microdrive' that is the same form as a compactflash card, IIRC.
The 4GB microdrive costs about as much as the 15GB minidrive, hence the cost parity.
Rarely does speed/size equate exactly with price, there's a bottom-limit and a steep upward-curve as you move from low-cost to high-end electronics.
Apple's own 'Disk Utility' has included the 'restore' funtionality from Netrestore since 10.3 came out. It's just a front-end to 'asr'.
I use command-line stuff and Carbon Copy Cloner to make a master, and use target-mode, disk utility, and several long firewire cables to restore labs.
If I had to automate it, I'd make a non-automounted partition with just Darwin on it, copy 'asr' and any needed files to it and boot to that to reimage the 'OS X' partition.
As far as I know, there are no multicast mac imaging solutions available from Apple or OSS outlets.
Get a UPS and ground it and any equipment touching the computer, including outside connections (cable modems)
Boost humidity in the computer area to 70%-90%.
Operate with the case on at all times.
Don't screw drives into the chassis very tightly, just enough to hold them in without vibrating.
Make sure your case is properly (not 'excessively', just 'properly') cooled. Does an air channel pass over the drive?
Replace your power supply and cabling, don't use that 'rounded cable' crap (it's more delicate).
keep the drive/PC running all the time. Use power management features to save juice while leaving the drive on most of the time (mine is set for two hours idle to spin-down).
Buy drives new, get them from a vendor that recently got OVER a quality issue. I have a feeling that the IBM Deskstar models after the 60 and 75 are mighty-good, I run a few GXP-180s 24/7 just fine. Vendors boost quality a LOT when they have to recover their good name.
Well if you recall, Mozilla 1.0, 1.4, and 1.7 were all 'extra stable' codebases, designed for vendor repackaging and forking.
It would be unwise for Sun to run Mozilla 1.5 or 1.6, because in between the 'extra stable' releases a lot of things change and (historically) break.
Once a year or so, the code gets the big projects landed and the tree gets a more thorough debugging than normal, any forks happen (camino, netscape, galeon), and a 'benchmark' release is made.
Well, where I work a 'Tactical Deployment' is when a user takes vacation or a sick day and I reimage their box with a newer OS. There's a lot of folks who cling dearly to their old familiar software.
Well in Linux I'm sure the X Virtual Framebuffer can be hacked into that. XvFB is like a 'fake' screen that has no video hardware, it's just a framebuffer in regular RAM.
The G5 isn't really so hot of a CPU, the current ones put out about as much wattage at 2GHz as my Athlon-XP at 1.8GHz. It puts out about 50 watts, about half as much as the latest P4 offerings. And contrary to popular belief, it's OK to run your CPU hot, there's no need to cool it down to 50C if it's rated and tested at 80C.
The G5 tower's design is designed for the case to be cool to the touch and quiet, hence the big fans, the system puts out less heat than an off-the-shelf P4 box. I'm sure Apple would implement some sort of case-as-heatsink for any SFF systems they cook up (pun intended).
I don't think it would be tough to build a really small G5, you can always underclock/undervolt the bus and CPU to save on heat and energy. Apple also has a history of placing the AC-adapter outside the system, which saves a lot of space.
What they oughtta-do is pay Hans Reiser some 'hush money' to use reiser4 core code, and implement their own repacking and indexing daemons. They could also write filesystem plugins for NTFS-style ACLs, compression, and encryption. I can't see why that wouldn't work.
I had some issues back in the 2.2 days, but ever since 2.4 came out I've been somewhat judicious about hardware purchases and I've not had a problem since.
See, I don't bitch about hardware compatability because my box is top-to-bottom Linux clean. It's really not hard to get a system built that's 100% Linux-friendly, you just have to go for the big brands, stay away from bleeding-edge stuff, and only buy stuff with chipsets you can identify.
Right now I'm in the market for a FireWire card for my box, but I can't figure out which chipset will yeild the maximum performance, so I'm holding off until I do more research.
I wouldn't buy a car without knowing what kind of brakes it had, and I don't buy computers if I don't know what kind of chips they contain.
Personally, I think they use those hot-button issues as a way to divvy everything right down the line at 50%.
I've stopped thinking that we manipulate the politicians, and started to realize that it's the other way around. If not then wouldn't there be more McCains and Chafees?
I don't think this is what the person asking wants though. This might limit BANDWIDTH, but it doesn't do anything to the real latency.
What the submitter wants is to simulate a line that already exists, and already has it's own bandwidth and latency.
Several people have answered this without understanding what the problem is. Some people say "just hose your connection with other traffic!" but that will cause packet-loss where a dedicated serial line (like a WAN) generally won't have any if it's not overloaded. Others say "Limit bandwidth!" but that will only do just that, it will limit the amount of data you can fling, but it will not properly give you the end-user experience of a true high-latency connection.
SMB is a great example, I can dial into my work and browse the LAN quite painlessly, it's a low-band low-latency connection. When I do dial in I can authenticate and browse directories very well, but downloading a file takes forever. When I VPN-in over a cable connection to the same network from a far-away ISP (high-band, high-latency) I find it painful to browse and authenticate, but pulling files is super-quick.
SMB works fine on low-band connections, but it will choke-and-die if you drop packets, so you can't get a good feel for how your dialup users will feel if you simulate a connection with the wrong properties. Many database programs and web-based applications are the same way.
Agreed. I personally don't have the niceness option set, but the grandparent of the thread seems to have a lower-end machine that might benefit from some niceness with his emerges.
true, but from experience, it's not so bad to fully charge them before storing them.
If I had only a few machines to deal with I'd be going for 50%, but I've got over 70 batteries here, and it would take a LONG time to get them all calibrated (up to 100%) and then discharge them to 50%. I'm trying to minimize battery loss while also minimizing the time I spend fiddling with this aspect of my job.
Alright. I understand. I think the 'slot' issue with kde-3.3 going into a different slot will prevent that from happening this time, but here's an alternate.
During the time you're using the machine, compile but don't install the kdelibs package. There's an option for that in emerge, it will build a tarball of the binary for you. Later, before you go to bed, do the command for a full emerge of kde and it will install the libs from the package.
LOL, I went to high school with 'Kersh. I remember how he showed me the first UN*X I've ever seen, mkLinux on his PowerBook 3400. The man is single-handedly responsible for both my affection for Apple and for getting me into Linux. Not to mention that he showed my friends and I 'South Park' long before it was ever on TV (it was '97 or '98 when he showed us jesus-vs-santa).
Now he's the guy behind kismet, which I use to monitor WiFi at work.
Thanks 'Kersh! I wish you much success with career and hobby, and hope you find a real-life anime chick to settle down with. Send me some tentacle-shots when you do.:-)
LiIon oxidizes when it runs low, permanently damaging it.
It's best to leave them plugged-in as much as possible, and if the machine ever foces itself to sleep, you need to get it to a power adapter ASAP and LEAVE IT PLUGGED IN until it's fully charged. That calibrates the battery!
Then there's something REALLY wrong with your hardware or your config.
Linux is the PREMIER multitasking OS. You should be able to emerge stuff for the system, play an MP3, robo-re-tag your entire collection of MP3s, and compile do other stuff all at the same time without a big slowdown.
Make sure you have hdparm configged (here's mine):/etc/conf.d/hdparm:
disc1_args="A1 -a64 -c1 -d1 -m16 -u1 -W1"
# rc-update add hdparm default
Also make sure your CFLAGS were sane. Here's mine:
CFLAGS="-pipe -O3 -march=athlon-xp -mfpmath=sse"
and that you have enough RAM. Compiling KDE will eat about 80MB/thread at times, so it's nice to have 256MB.
Well if you mistreat your battery it wil do that. It's not just Apple either. I killed one battery by leaving the PowerBook in the car in the summer, and an iBook battery by forgetting it overnight in the winter (we had some -5 degree nights last winter). I also killed a cell batery on one of those occasions, and a Compaq LiIon battery on another.
A great way to kill your batteries is to let them sit when they're discharged. LiIon needs to be charged as much as possible. We lost about forty batteries at my work last summer because nobody plugged the laptop carts in after the end of school.
You're essentially saying that if I step on your neck for a while, then punch you and let you go, you're better off that I punched you, because it was overall better than having your neck stepped on.
Sorry. The people who were against the sanctions (at least the ones I know) were always a tiny minority, whereas about half of folks were against the war. The people you saw who wanted the sanctions kept up were NOT the same who were asking for them to be lifted before. Both the sanctions and the war were the wrong answer for the problem. If sanctions worked, we'd have never had to go to war, and if war 'worked' Iraq wouldn't be a war-zone right now.
Saddam wasn't really the mass-murdering butcher we make him out to be. The guy was not morally right, by any standard, but the several incidents of mass-murder we blame him for were perpetrated under CIA support. Saddam even told our ambassador that he intended to invade Kuwait in '91 and our representative said "We have no intention of stepping into a middle-east turf-war."
And regardless of how many lives we 'saved' under your logic, it's still never right to invade a soverign territory unless you are attacked. As far as I know, Saddam didn't have troops standing-by to storm the beaches of Massachusetts.
Slashdot Mirror
I think it's just around the corner. The 'boomers are about five years away from developing prostate problems on a massive scale, and the ladies will have issues with incontinence too, they'll all be pumped-full of prescription drugs that keep them hopping up to piss. As soon as complaints about this start showing up, the producers are going to start making a scene in the movie that's a nice segue to an intermission, just like on TV. Thos of us who can hold our bladders will likely have to sit through commercials asking us to take the opportunity to refill our beverages and popcorn for a modest fee, and to check out next year's Ford lineup.
The funny thing is that cable TV was originally commercial-free, you PAID to not watch advertisements. I remember those days faintly, one of my neighbors had cable and it was quite a hoot at cookouts and block parties.
Somehow cable became so common and people became so passive that cable now has just as much advertising as broadcast, and the quality of the ads and programming is generally lower on cable.
So now we pay the content providers to watch the content, and the advertisers pay them to slip us ads. We even get advertised to when paying the ultimate in high-prices at the theaters. I think that in a decade's time you'll see movies with one or two commercial-filled 'intermissions' under the pretense of letting elderly folks use the potty. Just watch.
Taking an existing drive design and reducing it's capacity doesn't reduce cost at all.
As far as I know, adding capacity to an existing design by adding heads and platters costs VERY little. Hard drive manufacturers make sure they break-even on the lowest-end drive in the lineup, and the rest is real money.
The cost of making a 60GB drive is almost the same as making a 250GB drive with the same technology. The industry expenses are in the R&D for pushing the data density per-platter up.
Here's why:
All the components are (essentially) the same, except the hard drive. The hard drive for the iPod is a 1.8" part, and the mini uses an even smaller 'microdrive' that is the same form as a compactflash card, IIRC.
The 4GB microdrive costs about as much as the 15GB minidrive, hence the cost parity.
Rarely does speed/size equate exactly with price, there's a bottom-limit and a steep upward-curve as you move from low-cost to high-end electronics.
Apple's own 'Disk Utility' has included the 'restore' funtionality from Netrestore since 10.3 came out. It's just a front-end to 'asr'.
I use command-line stuff and Carbon Copy Cloner to make a master, and use target-mode, disk utility, and several long firewire cables to restore labs.
If I had to automate it, I'd make a non-automounted partition with just Darwin on it, copy 'asr' and any needed files to it and boot to that to reimage the 'OS X' partition.
As far as I know, there are no multicast mac imaging solutions available from Apple or OSS outlets.
Be nicer to your hardware.
Get a UPS and ground it and any equipment touching the computer, including outside connections (cable modems)
Boost humidity in the computer area to 70%-90%.
Operate with the case on at all times.
Don't screw drives into the chassis very tightly, just enough to hold them in without vibrating.
Make sure your case is properly (not 'excessively', just 'properly') cooled. Does an air channel pass over the drive?
Replace your power supply and cabling, don't use that 'rounded cable' crap (it's more delicate).
keep the drive/PC running all the time. Use power management features to save juice while leaving the drive on most of the time (mine is set for two hours idle to spin-down).
Buy drives new, get them from a vendor that recently got OVER a quality issue. I have a feeling that the IBM Deskstar models after the 60 and 75 are mighty-good, I run a few GXP-180s 24/7 just fine. Vendors boost quality a LOT when they have to recover their good name.
Well if you recall, Mozilla 1.0, 1.4, and 1.7 were all 'extra stable' codebases, designed for vendor repackaging and forking.
It would be unwise for Sun to run Mozilla 1.5 or 1.6, because in between the 'extra stable' releases a lot of things change and (historically) break.
Once a year or so, the code gets the big projects landed and the tree gets a more thorough debugging than normal, any forks happen (camino, netscape, galeon), and a 'benchmark' release is made.
Well, where I work a 'Tactical Deployment' is when a user takes vacation or a sick day and I reimage their box with a newer OS. There's a lot of folks who cling dearly to their old familiar software.
They don't even know what hit them.
...The phallus with the clit stimulator is pretty popular.
Thank you! I'm flattered. I finally got some good PR.
Well in Linux I'm sure the X Virtual Framebuffer can be hacked into that. XvFB is like a 'fake' screen that has no video hardware, it's just a framebuffer in regular RAM.
Well the pictures linked are bunk but...
The G5 isn't really so hot of a CPU, the current ones put out about as much wattage at 2GHz as my Athlon-XP at 1.8GHz. It puts out about 50 watts, about half as much as the latest P4 offerings. And contrary to popular belief, it's OK to run your CPU hot, there's no need to cool it down to 50C if it's rated and tested at 80C.
The G5 tower's design is designed for the case to be cool to the touch and quiet, hence the big fans, the system puts out less heat than an off-the-shelf P4 box. I'm sure Apple would implement some sort of case-as-heatsink for any SFF systems they cook up (pun intended).
I don't think it would be tough to build a really small G5, you can always underclock/undervolt the bus and CPU to save on heat and energy. Apple also has a history of placing the AC-adapter outside the system, which saves a lot of space.
What they oughtta-do is pay Hans Reiser some 'hush money' to use reiser4 core code, and implement their own repacking and indexing daemons. They could also write filesystem plugins for NTFS-style ACLs, compression, and encryption. I can't see why that wouldn't work.
I had some issues back in the 2.2 days, but ever since 2.4 came out I've been somewhat judicious about hardware purchases and I've not had a problem since.
See, I don't bitch about hardware compatability because my box is top-to-bottom Linux clean. It's really not hard to get a system built that's 100% Linux-friendly, you just have to go for the big brands, stay away from bleeding-edge stuff, and only buy stuff with chipsets you can identify.
Right now I'm in the market for a FireWire card for my box, but I can't figure out which chipset will yeild the maximum performance, so I'm holding off until I do more research.
I wouldn't buy a car without knowing what kind of brakes it had, and I don't buy computers if I don't know what kind of chips they contain.
Personally, I think they use those hot-button issues as a way to divvy everything right down the line at 50%.
I've stopped thinking that we manipulate the politicians, and started to realize that it's the other way around. If not then wouldn't there be more McCains and Chafees?
I don't think this is what the person asking wants though. This might limit BANDWIDTH, but it doesn't do anything to the real latency.
What the submitter wants is to simulate a line that already exists, and already has it's own bandwidth and latency.
Several people have answered this without understanding what the problem is. Some people say "just hose your connection with other traffic!" but that will cause packet-loss where a dedicated serial line (like a WAN) generally won't have any if it's not overloaded. Others say "Limit bandwidth!" but that will only do just that, it will limit the amount of data you can fling, but it will not properly give you the end-user experience of a true high-latency connection.
SMB is a great example, I can dial into my work and browse the LAN quite painlessly, it's a low-band low-latency connection. When I do dial in I can authenticate and browse directories very well, but downloading a file takes forever. When I VPN-in over a cable connection to the same network from a far-away ISP (high-band, high-latency) I find it painful to browse and authenticate, but pulling files is super-quick.
SMB works fine on low-band connections, but it will choke-and-die if you drop packets, so you can't get a good feel for how your dialup users will feel if you simulate a connection with the wrong properties. Many database programs and web-based applications are the same way.
Agreed. I personally don't have the niceness option set, but the grandparent of the thread seems to have a lower-end machine that might benefit from some niceness with his emerges.
true, but from experience, it's not so bad to fully charge them before storing them.
If I had only a few machines to deal with I'd be going for 50%, but I've got over 70 batteries here, and it would take a LONG time to get them all calibrated (up to 100%) and then discharge them to 50%. I'm trying to minimize battery loss while also minimizing the time I spend fiddling with this aspect of my job.
Alright. I understand. I think the 'slot' issue with kde-3.3 going into a different slot will prevent that from happening this time, but here's an alternate.
During the time you're using the machine, compile but don't install the kdelibs package. There's an option for that in emerge, it will build a tarball of the binary for you. Later, before you go to bed, do the command for a full emerge of kde and it will install the libs from the package.
LOL, I went to high school with 'Kersh. I remember how he showed me the first UN*X I've ever seen, mkLinux on his PowerBook 3400. The man is single-handedly responsible for both my affection for Apple and for getting me into Linux. Not to mention that he showed my friends and I 'South Park' long before it was ever on TV (it was '97 or '98 when he showed us jesus-vs-santa).
:-)
Now he's the guy behind kismet, which I use to monitor WiFi at work.
Thanks 'Kersh! I wish you much success with career and hobby, and hope you find a real-life anime chick to settle down with. Send me some tentacle-shots when you do.
That's right!
LiIon oxidizes when it runs low, permanently damaging it.
It's best to leave them plugged-in as much as possible, and if the machine ever foces itself to sleep, you need to get it to a power adapter ASAP and LEAVE IT PLUGGED IN until it's fully charged. That calibrates the battery!
I didn't mean to come off as condescending. My apologies.
I use Occam's Razor, in which the reason a person's stuff is broken is probably the simplest reason: they abused it.
In your case, it seems like you got a somewhat bunk battery, or there's something wrong with your box.
In any case, you'll have to go out and buy a battery now. Are there folks making third-party PowerBook battery units?
Then there's something REALLY wrong with your hardware or your config.
/etc/conf.d/hdparm:
Linux is the PREMIER multitasking OS. You should be able to emerge stuff for the system, play an MP3, robo-re-tag your entire collection of MP3s, and compile do other stuff all at the same time without a big slowdown.
Make sure you have hdparm configged (here's mine):
disc1_args="A1 -a64 -c1 -d1 -m16 -u1 -W1"
# rc-update add hdparm default
Also make sure your CFLAGS were sane. Here's mine:
CFLAGS="-pipe -O3 -march=athlon-xp -mfpmath=sse"
and that you have enough RAM. Compiling KDE will eat about 80MB/thread at times, so it's nice to have 256MB.
So what?
/etc/make.conf
Set PORTAGE_NICENESS="15" in
# ACCEPT_KEYWORDS="~x86" emerge kde
go about your normal business, it takes about ten hours to compile on my 1.4GHz Athlon. You can stiill use your system while it compiles, you know.
You'll still have KDE-3.3 months before most people can get it in thoer shrinkwrapped distros.
Well if you mistreat your battery it wil do that. It's not just Apple either. I killed one battery by leaving the PowerBook in the car in the summer, and an iBook battery by forgetting it overnight in the winter (we had some -5 degree nights last winter). I also killed a cell batery on one of those occasions, and a Compaq LiIon battery on another.
A great way to kill your batteries is to let them sit when they're discharged. LiIon needs to be charged as much as possible. We lost about forty batteries at my work last summer because nobody plugged the laptop carts in after the end of school.
You're essentially saying that if I step on your neck for a while, then punch you and let you go, you're better off that I punched you, because it was overall better than having your neck stepped on.
Sorry. The people who were against the sanctions (at least the ones I know) were always a tiny minority, whereas about half of folks were against the war. The people you saw who wanted the sanctions kept up were NOT the same who were asking for them to be lifted before. Both the sanctions and the war were the wrong answer for the problem. If sanctions worked, we'd have never had to go to war, and if war 'worked' Iraq wouldn't be a war-zone right now.
Saddam wasn't really the mass-murdering butcher we make him out to be. The guy was not morally right, by any standard, but the several incidents of mass-murder we blame him for were perpetrated under CIA support. Saddam even told our ambassador that he intended to invade Kuwait in '91 and our representative said "We have no intention of stepping into a middle-east turf-war."
And regardless of how many lives we 'saved' under your logic, it's still never right to invade a soverign territory unless you are attacked. As far as I know, Saddam didn't have troops standing-by to storm the beaches of Massachusetts.