I heard that if you light your cigarette off a candle, a sailor dies.
I think SOME things at Microsoft have changed in the last decade. I found the W2K and XP service pack releases to have a LOT fewer problems than those of NT4. The stakes are MUCH higher now considering that the NT kernel is now the only kernel and that the number of installed machines and potential problems is much higher.
What was so bad about XP SP1? It seemed like improvements on all fronts to me.
Argh. What's the point? A television screen in the USA is 15hz, but the signal is interlaced for a total perception of 30fps.
That's right, your TV can only do 30fps out-of-the-box. If you send it a signal from a unit outputting a 60fps signal, you're just throwing bits to the wind.
This is the sort of thing that drives me nuts. My monitor runs at 85hz, any framerate over 85fps is wasted CPU/GPU cycles. LCD monitors run at 60hz, for the most part.
Sorry pal, according to the documentation, the 9200 has FULL 3D support in XFree 4.4 and XOrg 6.7.0.
I've got a 7500 and a 9200 here working with only open-source drivers, and the 3D is fully accelerated, no closed binary drivers whatsoever.
The 9200 is an RV280, which has in-kernel 3D support.
If you're not getting full OpenGL acceleration the problem is on your end (prolly with your xorg.conf or kernel options). I'm right here if you've got any questions.
Hmm, considering my plans, if anyone sees my heathen self tonight wallowing in booze, please tilt my head to the side so I don't choke on my own vomit. I'll be the only guy in Providence with an NTK t-shirt on. Leave your slashdot nick written on my forehead with a sharpie so I know who to mod-up in the future.
Well, it's hoped that most of the hardware involved is connected to a SAN anyway, so it wouldn't matter much.
But in my experience using Mac-On-Linux, which lets you virtualize Mac OS (and OS X) sessions on Linux, the disk performance is actually INCREASED. What you have is one OS providing the virtualization services (network, disks, memory segment) for the others. When the hosted system does IO it gets handed to the master OS right away, all IO becomes asynchronous when it hits the hosting layer.
What you end up with is an OS that's tuned for great IO providing services with 'native dummy' drivers to the hosted systems. the net result is increased performance.
Overall though, partitioning is really nothing new, my G3 can run partitioned sessions of Linux, Mac OS, Mac OS X, and I'll bet getting *BSD wouldn't be rocket science. I think this announcement by IBM is really just market-driven hype of a feature they've already got. They might just be reimplementing their way of doing it too. I can see the next PPC970 hardware running the 'hosting OS' as embedded Linux on the PC4XX 'control chip' present on all the 9XX motherboards. This would essentially be an IBM/PowerPC-enhanced version of LinuxBIOS. It would be LinuxBIOS that booted your OS inside a virtualized environment. Way cool, and way useful!
My regrets. I spend most of my time in Providence, which has about 15 Dunkin Donuts' within 2km.
I was RAISED on the stuff, a lot of my friends started drinking DD when we were in sixth grade. People from my work (near Boston) joke about the Rhode Island Coffee Fix, meaning that I'll duck-out after lunch and drive past two coffee shops to get to the nearest DD. In the summer I usually slam a 24-ounce iced coffee in about two minutes, it gives me an ice-cream headache.
The trick to Dunkin Donuts is that the brew a light roast, which is not usually found in most 'classy' coffee shops. In my opinion, a light roast is EXACTLY the sort of caffiene pick-up I want after lunch, not a heavy black brew that taints my blood (save that stuff to get me out of bed!).
BTW, if you ever get down to Rhode Island, if you order coffee at Dunkin Donuts 'regular' you get about four times the cream and sugar you expect. Make sure to tell them to go 'easy on the cream and sugar'.
Er, MAC filtering is the LEAST safe way to lock-out wireless. MAC addresses are EASILY picked-out of the air, and all you have to do is push the address you want to your wifi card to 'steal' one.
MAC filtering is not encryption, even if you MAC filter, I can come by with any number of 'tools' and leech all your traffic without having to do any work. Perhaps the only thing MAC filtering does is keep the non-technical neighbor upstairs off your signal.
This article refers to another way to crack networks that are actually encrypted, which was generally enough of a hassle that someone would want to specifically target YOU before going through the trouble. As with all encryption though, cracking what's out there gets easier every day, time to move up to something else!
Yes, someone could. It would then take about ten minutes for one of the people out there who DOES check the signatures of binaries they download (maybe 1% of the population?) to discover and report it.
There is room for improvement, but the 'many eyes' thing helps with securing releases as well.
Agreed, though I'm more worried about defacement of my bike than theft.
Having been an urban biker for about ten years, I can tell you that having a bike defaced/kicked/smashed/broken/scratched every day is overall much more disheartening than just having one stolen every few years.
It's one thing to know that someone REALLY wanted something you have; It's another to realize how many ignorant morons with stock in the oil companies will kick someone's bike.
The Apple updates _DO_ include updates to the open-source stuff, like Apache and Samba. But you don't get updates to the latest/greatest versions from the source, you get patched versions from when Apple snapshotted the code.
For instance, 10.3 Server ran Samba 3.0.2, the last update, 10.3.6 updated samba to 3.0.5, even though samba is now at 3.0.10. Apple tries (like almost all vendors) to keep everything the same, while patching for bugs, it's a good idea. I'd be pissed if they tracked the latest versions, this isn't Gentoo, it's supposed to be a stable system to the core.
You see, version bumps in Apple's OS X only happen when they're nescessary. If Apple can't backport a security fix to a package until it version-bumps it, that's when you'll see the update.
Alright, so far I've been frustrated by packages that don't build, and packages that require dependencies that could easily be satisfied by 'native' OS X stuff, but instead require the building of strings of dependencies that seem unnescessary.
for some things, portage on OS X is great, CLI tools seem to work (and look) fantastic. I still get a lot of collissions when trying to install stuff though. I hope it gets better, and that I get better at understanding the direction it's moving so I can test and submit bugs for it.
BTW, I figured that portage for OS X would do stuff like let me build firefox for aqua, but it seems to build for X11, which seems lame, I think it should aim to build targets as native as possible. Granted, you could never do this with a lot of packages, but those with aqua ports should be built with them. Of course, then there's the folks running straight Darwin with X11 on top, and they'd want the X11 version, sounds like time to keyword out stuff like that.
Has anyone managed to get an aqua GTK+ wrapper out? that would make 'native' aqua ports a lot easier, methinks.
Apple makes REALLY solid workstation/server-class desktops. My G3 circa 1999 has 64-bit PCI slots, shipped with Ultra2Wide SCSI hardware, and a 66MHz PCI slot (double-speed). You can't find reasonably-priced PC hardware like that, it's just not available.
Now we all know that OS X is really cool, but it packs some serious overhead for the graphics and sound systems, and while it's perfectly acceptable for a desktop OS (or a server if you're running it on newer metal), I'd much rather have the lower latency, higher performance, and reduced resource use of Linux on my more utilitarian boxes.
Also, if you're running a server on Apple software, you're limited to UFS, HFS+, and VFAT file systems, which is alright if you don't do anything wild, but I often need to pack 120,000 data files into a single folder and then manipulate them, that's something I find much more appealing when it's happening on a reiserfs partition, which only works under Linux.
I run Linux on a second partition on my AlBook as well, and I can tell you that the hardware support is quite good, Apple uses mostly commodity parts, so there's usually existing drivers in the kernel for Apple's hardware. I particularly like the feature where you can switch the 'apple sleep light' on the PowerBook and iBook into a hard drive activity indicator, very useful.
Another thing, I just realy LIKE Linux, I enjoy booting to an XDM session that kicks-off WindowMaker. It CAN be done on OS X, but it's a lot of work if all you want is to use *NIX apps. I like portage, and the OS X implementation is still mostly useless.
meta name="GENERATOR" content="Microsoft FrontPage 5.0"
Strike one.
visible interpolation of images on the main page (it looks like some fool blew things up WAY beyond native resolution).
Strike two.
Link on front page to the thing about 'regula' (upper right-hand corner) is a dead link.
Three strikes.
Methinks this admin will get what's coming to him in his own good time. Some kiddie will stumble on this site someday and have a ball. People who can't make sure they have working links on their front pages, and let users install whatever they want in labs generally don't know much about security.
Speech recognition shipped on my 25MHz Mac Quadra 660av back in '93, on 'System 7.1' (wasn't called 'Mac OS' back then). It was cool and useful, worked well, and Windows isn't at that level yet eleven years later.
Also worth nothing that my Dad bought an 8MHz Mac Plus (for $2,500) back in '87 and it could do decent text-to-speech. My friends and I learned a lot about phonetics, A.I., typing, and computers by playing with that little applet.
Word! I'm actually holding off on replacing my PowerPC file server until I can get a Pentium-M based desktop board to use instead. It's a shame these aren't out in full-force yet; I bet a speedy but low-power/low-heat silent pM would outsell a lot of what Dell is foisting these days.
I would get a VIA EPIA, but they really lack the solid feel and performance I get from VIA, Intel, and Apple boards. Also, the VERY lame compiler support for VIA's CPUs is most discouraging.
The oter option is a last-gen pIII on a decent micro-atx board, which I might have to resort to if the aforementioned pM mobos don't come around soon.
The pM is most pleasing, it really does a fantastic job of keeping cool and delivering the goods, at least on the laptops I've worked on.
Er... As far as I know the naming of the 'generation' is independent of the instruction set (which remains largely the same) and more about the method by which the CPU processes them. The p4 SHOULD be called a 786, all signs clearly point to this, it just isn't for some reason. I think a lot of it has to do with it being much less pretty to say:
"[four|five|six]-eighty-six" sounds good, "seven-eighty-six" is sorta ugly.
* the pentium M looks a lot like the p4 on paper, but is actually closer to the pIII in architecture.
** I don't think anyone actually referred to the p4 as the 786, even though it represented a dramatic-enough redesign to warrant it.
It should also be noted that the difference that defined the 686 from the 586 was that 686 chips are actually more like RISC chips EMULATING x86. If you were in IT when the PentiumPro and PII came out, you'd recall that the 686-based chips ran 16-bit apps much slower than their Pentium counterparts. Also note that the pII and pIII were essentially the same beast, with added instructions and minor reworks to take advantage of SIMD and miniaturization.
Well my numbers are a guess, but my setup is atypical.
I've got a beefy Athlon Barton workstation, a PowerPC file server with two 10K RPM SCSI drives, a slew of netwoking equipment, and a 19" CRT that stays on 24/7 in the winter. And my boxes don't sit idle most of the time, I've got them automagically rebuilding Gentoo stages on both x86 and PPC from a daily portage snapshot, to test if anything added to portage interferes with a clean build.
My setup burns some serious wattage, I'm sure. But paying for ANY wattage that gets invested into the great outdoors seems foolish to me these days considering current energy prices.
Well I understand your case of not having the space, I don't have that problem and sort of like the idea that the 800 watts of my whole rig (two boxes, one monitor, plus gear) are helping to warm the INSIDE of my house, keeping the heater from having to kick-on so often.
I've got a decent-sized apartment, but there is DEFINITELY a benefit from having the equivalent of half a space heater running all the time. Also helps justify leaving them on 24/7, the excess energy use is reducing my other energy costs.
Speding money on electricity to warm a CPU in the garage seems foolish to me. If you want safe backups buy a DVD burner or an old tape drive, or a removable hard drive to tote to another location weekly.
You do realize that your chance of contracting HIV (as a male) from a random one-night-hetero-stand in North America is about 1 in 30,000 right? You can have about 1,500 unprotected partners before your chances of infection approach 5% with those numbers.
You also realize that those odds get dramatically better if you're sleeping with someone who didn't IV-inject, is white, and didn't have a blood transfusion over 15 years ago, right?
HIV is not a social disease like HPV, it's not something the average American has to worry about. HIV is a disease that primarily preys upon cultures that engage in massive wanton unprotected sex, drug use, war crimes (mass rape), and other vices. Hell, I bet we could -LEGISLATE- and -LITIGATE- HIV to very low numbers in the USA, where we haven't been overrun yet and there are no major risk factors.
I'd worry more about all the other little STDs or the odd chance she got pregnant.
Hell, I worry more about cleaning the sheets after a one night stand than I do about anything else. I try to keep my dick in my pants when the nasty girls are about, it dramatically reduces your chances of any sort of infection if you look, smell, and taste before you take the plunge.
They do pour a disproportionate amount of R&D into the iPod. But remember that if you take a great product that posts a profit and throw millions of R&D at it it might really eat into that profitability. There's VERY little Apple can get from R&D on an iPod, the whole thing is made up of commodity parts. The only thing Apple does is provide an OS for the embedded ARM chip and put all the pieces together inside a nice pretty case (yes, they design the board in there too).
Why destroy a profit-center by spending excessively on it? Especially when you're not gonna get much out of it.
Don't forget the cool system-wide features provided by the NeXT-derived cocoa framework.
Did you know that you can run ANY text control in the system through a spell-checker that shares a common user dictionary? That means that ALL of your apps learn how to spell ALL of your custom words at once.
ColorSync? QuickTime? These are awesome toolkits that make writing apps easy and using them even easier.
What about the entire graphics system running 'display postscript' on top of an OpenGL subsystem? Do you realize the implications of this for Apple when 200dpi LCD displays become available? It will be like moving from regular broadcast to HDTV, you won't even know you're looking at an array of pixels. Microsoft is several years behind Apple on this front.
Our policy for September was 'wait and see' because we hadn't tested it. Now it seems OK but the network managers are dragging their feet on giving us the go-ahead (it's not really their say anyway, it's desktop's). I've been pushing it since October, and now the desktop team is pushing it to any machine that gets reimaged or comes through the shop.
As for educating to use the firewall, there's more to it than that. we've got two desktop techs (myself included) for 1,300 direct-users and about 300 home machines. Most of the userbase doesn't know how to install AIM, that wireless only works if you're near a base unit, or that 'right-clicking' doesn't mean you're doing it 'wrong' now. That's another issue altogether.
We simply don't have the resources either to teach computer security to the users (short staffed!) or to get the time from the admins to teach the faculty and students ("these kids are going to be upper-class, why should we teach them techie stuff?" -overall sentiment).
The firewall is useless in our environment anyway, we're VLAN'd out the wazoo, so worms really don't get very far in here, and when we have left it on it causes more support calls than it's worth. It also prevents some of the scanning software we use from working, I like to be able to portscan borked machines from my desk.
Slashdot Mirror
I heard that if you light your cigarette off a candle, a sailor dies.
I think SOME things at Microsoft have changed in the last decade. I found the W2K and XP service pack releases to have a LOT fewer problems than those of NT4. The stakes are MUCH higher now considering that the NT kernel is now the only kernel and that the number of installed machines and potential problems is much higher.
What was so bad about XP SP1? It seemed like improvements on all fronts to me.
Argh. What's the point? A television screen in the USA is 15hz, but the signal is interlaced for a total perception of 30fps.
That's right, your TV can only do 30fps out-of-the-box. If you send it a signal from a unit outputting a 60fps signal, you're just throwing bits to the wind.
This is the sort of thing that drives me nuts. My monitor runs at 85hz, any framerate over 85fps is wasted CPU/GPU cycles. LCD monitors run at 60hz, for the most part.
Sorry pal, according to the documentation, the 9200 has FULL 3D support in XFree 4.4 and XOrg 6.7.0.
I've got a 7500 and a 9200 here working with only open-source drivers, and the 3D is fully accelerated, no closed binary drivers whatsoever.
The 9200 is an RV280, which has in-kernel 3D support.
If you're not getting full OpenGL acceleration the problem is on your end (prolly with your xorg.conf or kernel options). I'm right here if you've got any questions.
Hmm, considering my plans, if anyone sees my heathen self tonight wallowing in booze, please tilt my head to the side so I don't choke on my own vomit. I'll be the only guy in Providence with an NTK t-shirt on. Leave your slashdot nick written on my forehead with a sharpie so I know who to mod-up in the future.
Thanks in advance!
Well, it's hoped that most of the hardware involved is connected to a SAN anyway, so it wouldn't matter much.
But in my experience using Mac-On-Linux, which lets you virtualize Mac OS (and OS X) sessions on Linux, the disk performance is actually INCREASED. What you have is one OS providing the virtualization services (network, disks, memory segment) for the others. When the hosted system does IO it gets handed to the master OS right away, all IO becomes asynchronous when it hits the hosting layer.
What you end up with is an OS that's tuned for great IO providing services with 'native dummy' drivers to the hosted systems. the net result is increased performance.
Overall though, partitioning is really nothing new, my G3 can run partitioned sessions of Linux, Mac OS, Mac OS X, and I'll bet getting *BSD wouldn't be rocket science. I think this announcement by IBM is really just market-driven hype of a feature they've already got. They might just be reimplementing their way of doing it too. I can see the next PPC970 hardware running the 'hosting OS' as embedded Linux on the PC4XX 'control chip' present on all the 9XX motherboards. This would essentially be an IBM/PowerPC-enhanced version of LinuxBIOS. It would be LinuxBIOS that booted your OS inside a virtualized environment. Way cool, and way useful!
And on the seventh day, he rested; after he disconnected.
My regrets. I spend most of my time in Providence, which has about 15 Dunkin Donuts' within 2km.
I was RAISED on the stuff, a lot of my friends started drinking DD when we were in sixth grade. People from my work (near Boston) joke about the Rhode Island Coffee Fix, meaning that I'll duck-out after lunch and drive past two coffee shops to get to the nearest DD. In the summer I usually slam a 24-ounce iced coffee in about two minutes, it gives me an ice-cream headache.
The trick to Dunkin Donuts is that the brew a light roast, which is not usually found in most 'classy' coffee shops. In my opinion, a light roast is EXACTLY the sort of caffiene pick-up I want after lunch, not a heavy black brew that taints my blood (save that stuff to get me out of bed!).
BTW, if you ever get down to Rhode Island, if you order coffee at Dunkin Donuts 'regular' you get about four times the cream and sugar you expect. Make sure to tell them to go 'easy on the cream and sugar'.
Er, MAC filtering is the LEAST safe way to lock-out wireless. MAC addresses are EASILY picked-out of the air, and all you have to do is push the address you want to your wifi card to 'steal' one.
MAC filtering is not encryption, even if you MAC filter, I can come by with any number of 'tools' and leech all your traffic without having to do any work. Perhaps the only thing MAC filtering does is keep the non-technical neighbor upstairs off your signal.
This article refers to another way to crack networks that are actually encrypted, which was generally enough of a hassle that someone would want to specifically target YOU before going through the trouble. As with all encryption though, cracking what's out there gets easier every day, time to move up to something else!
Yes, someone could. It would then take about ten minutes for one of the people out there who DOES check the signatures of binaries they download (maybe 1% of the population?) to discover and report it.
There is room for improvement, but the 'many eyes' thing helps with securing releases as well.
Agreed, though I'm more worried about defacement of my bike than theft.
Having been an urban biker for about ten years, I can tell you that having a bike defaced/kicked/smashed/broken/scratched every day is overall much more disheartening than just having one stolen every few years.
It's one thing to know that someone REALLY wanted something you have; It's another to realize how many ignorant morons with stock in the oil companies will kick someone's bike.
The Apple updates _DO_ include updates to the open-source stuff, like Apache and Samba. But you don't get updates to the latest/greatest versions from the source, you get patched versions from when Apple snapshotted the code.
For instance, 10.3 Server ran Samba 3.0.2, the last update, 10.3.6 updated samba to 3.0.5, even though samba is now at 3.0.10. Apple tries (like almost all vendors) to keep everything the same, while patching for bugs, it's a good idea. I'd be pissed if they tracked the latest versions, this isn't Gentoo, it's supposed to be a stable system to the core.
You see, version bumps in Apple's OS X only happen when they're nescessary. If Apple can't backport a security fix to a package until it version-bumps it, that's when you'll see the update.
Alright, so far I've been frustrated by packages that don't build, and packages that require dependencies that could easily be satisfied by 'native' OS X stuff, but instead require the building of strings of dependencies that seem unnescessary.
for some things, portage on OS X is great, CLI tools seem to work (and look) fantastic. I still get a lot of collissions when trying to install stuff though. I hope it gets better, and that I get better at understanding the direction it's moving so I can test and submit bugs for it.
BTW, I figured that portage for OS X would do stuff like let me build firefox for aqua, but it seems to build for X11, which seems lame, I think it should aim to build targets as native as possible. Granted, you could never do this with a lot of packages, but those with aqua ports should be built with them. Of course, then there's the folks running straight Darwin with X11 on top, and they'd want the X11 version, sounds like time to keyword out stuff like that.
Has anyone managed to get an aqua GTK+ wrapper out? that would make 'native' aqua ports a lot easier, methinks.
I use Linux on a Mac, for several reasons:
Apple makes REALLY solid workstation/server-class desktops. My G3 circa 1999 has 64-bit PCI slots, shipped with Ultra2Wide SCSI hardware, and a 66MHz PCI slot (double-speed). You can't find reasonably-priced PC hardware like that, it's just not available.
Now we all know that OS X is really cool, but it packs some serious overhead for the graphics and sound systems, and while it's perfectly acceptable for a desktop OS (or a server if you're running it on newer metal), I'd much rather have the lower latency, higher performance, and reduced resource use of Linux on my more utilitarian boxes.
Also, if you're running a server on Apple software, you're limited to UFS, HFS+, and VFAT file systems, which is alright if you don't do anything wild, but I often need to pack 120,000 data files into a single folder and then manipulate them, that's something I find much more appealing when it's happening on a reiserfs partition, which only works under Linux.
I run Linux on a second partition on my AlBook as well, and I can tell you that the hardware support is quite good, Apple uses mostly commodity parts, so there's usually existing drivers in the kernel for Apple's hardware. I particularly like the feature where you can switch the 'apple sleep light' on the PowerBook and iBook into a hard drive activity indicator, very useful.
Another thing, I just realy LIKE Linux, I enjoy booting to an XDM session that kicks-off WindowMaker. It CAN be done on OS X, but it's a lot of work if all you want is to use *NIX apps. I like portage, and the OS X implementation is still mostly useless.
Alright, I just went there too...
meta name="GENERATOR" content="Microsoft FrontPage 5.0"
Strike one.
visible interpolation of images on the main page (it looks like some fool blew things up WAY beyond native resolution).
Strike two.
Link on front page to the thing about 'regula' (upper right-hand corner) is a dead link.
Three strikes.
Methinks this admin will get what's coming to him in his own good time. Some kiddie will stumble on this site someday and have a ball. People who can't make sure they have working links on their front pages, and let users install whatever they want in labs generally don't know much about security.
could be overcome by the diversity of the company's clients
I swear I hear one more thing about diversity overcoming problems, I'll wring someone's neck.
Speech recognition shipped on my 25MHz Mac Quadra 660av back in '93, on 'System 7.1' (wasn't called 'Mac OS' back then). It was cool and useful, worked well, and Windows isn't at that level yet eleven years later.
Also worth nothing that my Dad bought an 8MHz Mac Plus (for $2,500) back in '87 and it could do decent text-to-speech. My friends and I learned a lot about phonetics, A.I., typing, and computers by playing with that little applet.
Word! I'm actually holding off on replacing my PowerPC file server until I can get a Pentium-M based desktop board to use instead. It's a shame these aren't out in full-force yet; I bet a speedy but low-power/low-heat silent pM would outsell a lot of what Dell is foisting these days.
I would get a VIA EPIA, but they really lack the solid feel and performance I get from VIA, Intel, and Apple boards. Also, the VERY lame compiler support for VIA's CPUs is most discouraging.
The oter option is a last-gen pIII on a decent micro-atx board, which I might have to resort to if the aforementioned pM mobos don't come around soon.
The pM is most pleasing, it really does a fantastic job of keeping cool and delivering the goods, at least on the laptops I've worked on.
Er... As far as I know the naming of the 'generation' is independent of the instruction set (which remains largely the same) and more about the method by which the CPU processes them. The p4 SHOULD be called a 786, all signs clearly point to this, it just isn't for some reason. I think a lot of it has to do with it being much less pretty to say:
"[four|five|six]-eighty-six" sounds good, "seven-eighty-six" is sorta ugly.
the x in x86 is a generational thing.
There have been different brands of chip based on the same generation technology, for instance:
586 = pentium
686 = ppro, pII, pIII, pM*
786** = p4
* the pentium M looks a lot like the p4 on paper, but is actually closer to the pIII in architecture.
** I don't think anyone actually referred to the p4 as the 786, even though it represented a dramatic-enough redesign to warrant it.
It should also be noted that the difference that defined the 686 from the 586 was that 686 chips are actually more like RISC chips EMULATING x86. If you were in IT when the PentiumPro and PII came out, you'd recall that the 686-based chips ran 16-bit apps much slower than their Pentium counterparts. Also note that the pII and pIII were essentially the same beast, with added instructions and minor reworks to take advantage of SIMD and miniaturization.
Well my numbers are a guess, but my setup is atypical.
I've got a beefy Athlon Barton workstation, a PowerPC file server with two 10K RPM SCSI drives, a slew of netwoking equipment, and a 19" CRT that stays on 24/7 in the winter. And my boxes don't sit idle most of the time, I've got them automagically rebuilding Gentoo stages on both x86 and PPC from a daily portage snapshot, to test if anything added to portage interferes with a clean build.
My setup burns some serious wattage, I'm sure. But paying for ANY wattage that gets invested into the great outdoors seems foolish to me these days considering current energy prices.
Well I understand your case of not having the space, I don't have that problem and sort of like the idea that the 800 watts of my whole rig (two boxes, one monitor, plus gear) are helping to warm the INSIDE of my house, keeping the heater from having to kick-on so often.
I've got a decent-sized apartment, but there is DEFINITELY a benefit from having the equivalent of half a space heater running all the time. Also helps justify leaving them on 24/7, the excess energy use is reducing my other energy costs.
Speding money on electricity to warm a CPU in the garage seems foolish to me. If you want safe backups buy a DVD burner or an old tape drive, or a removable hard drive to tote to another location weekly.
You do realize that your chance of contracting HIV (as a male) from a random one-night-hetero-stand in North America is about 1 in 30,000 right? You can have about 1,500 unprotected partners before your chances of infection approach 5% with those numbers.
You also realize that those odds get dramatically better if you're sleeping with someone who didn't IV-inject, is white, and didn't have a blood transfusion over 15 years ago, right?
HIV is not a social disease like HPV, it's not something the average American has to worry about. HIV is a disease that primarily preys upon cultures that engage in massive wanton unprotected sex, drug use, war crimes (mass rape), and other vices. Hell, I bet we could -LEGISLATE- and -LITIGATE- HIV to very low numbers in the USA, where we haven't been overrun yet and there are no major risk factors.
I'd worry more about all the other little STDs or the odd chance she got pregnant.
Hell, I worry more about cleaning the sheets after a one night stand than I do about anything else. I try to keep my dick in my pants when the nasty girls are about, it dramatically reduces your chances of any sort of infection if you look, smell, and taste before you take the plunge.
They do pour a disproportionate amount of R&D into the iPod. But remember that if you take a great product that posts a profit and throw millions of R&D at it it might really eat into that profitability. There's VERY little Apple can get from R&D on an iPod, the whole thing is made up of commodity parts. The only thing Apple does is provide an OS for the embedded ARM chip and put all the pieces together inside a nice pretty case (yes, they design the board in there too).
Why destroy a profit-center by spending excessively on it? Especially when you're not gonna get much out of it.
Don't forget the cool system-wide features provided by the NeXT-derived cocoa framework.
Did you know that you can run ANY text control in the system through a spell-checker that shares a common user dictionary? That means that ALL of your apps learn how to spell ALL of your custom words at once.
ColorSync? QuickTime? These are awesome toolkits that make writing apps easy and using them even easier.
What about the entire graphics system running 'display postscript' on top of an OpenGL subsystem? Do you realize the implications of this for Apple when 200dpi LCD displays become available? It will be like moving from regular broadcast to HDTV, you won't even know you're looking at an array of pixels. Microsoft is several years behind Apple on this front.
I'm sorry, I didn't make myself clear.
Our policy for September was 'wait and see' because we hadn't tested it. Now it seems OK but the network managers are dragging their feet on giving us the go-ahead (it's not really their say anyway, it's desktop's). I've been pushing it since October, and now the desktop team is pushing it to any machine that gets reimaged or comes through the shop.
As for educating to use the firewall, there's more to it than that. we've got two desktop techs (myself included) for 1,300 direct-users and about 300 home machines. Most of the userbase doesn't know how to install AIM, that wireless only works if you're near a base unit, or that 'right-clicking' doesn't mean you're doing it 'wrong' now. That's another issue altogether.
We simply don't have the resources either to teach computer security to the users (short staffed!) or to get the time from the admins to teach the faculty and students ("these kids are going to be upper-class, why should we teach them techie stuff?" -overall sentiment).
The firewall is useless in our environment anyway, we're VLAN'd out the wazoo, so worms really don't get very far in here, and when we have left it on it causes more support calls than it's worth. It also prevents some of the scanning software we use from working, I like to be able to portscan borked machines from my desk.