I would never rely on a piece of hardware (or *anything*) bought in Aldi. I mean, have you ever even looked at what they're selling?
Yes, and the components that I've seen listed on the spec sheets of the machines sold by both Aldi and Lidl are pretty good - Seagate discs, Samsung memory, ASUS motherboards, Sony optical drives, etc. You'd be hard pressed to put together a machine of similar spec yourself, and that's excluding the software (which, given this is/. may actually have negative value) and warranty it all for 3 or 5 years as both Aldi and Lidl do. About the only concern with these machines is the quality of the phone support, but again, given this is/.....
Interestingly, a German friend told me a few years back that Medion/Aldi had secured 11.6% of the PC market in Germany.
Not sure how to do this with linux, but with freebsd all i did was change a single line in the ppp.conf** file
Pretty similar: look into the bsdcomp, deflate, predictor1 options in pppd and the CONFIG_PPP_BSDCOMP and CONFIG_PPP_DEFLATE kernel options. Of course, the hard thing - at least in the UK - is finding an ISP who supports either of those options. I've never experienced that, so I've always had to rely upon my modems' built-in compression. With serial port tuning (i.e. 230400baud), I was able to achieve 20Kbyte/s on a V.90 modem downloading files consisting entirely of 0x00 or 0xff. In reality 5-7Kbyte/s was much more realistic for ASCII text.
For us that live in coldish countries, and I'd place Scotland in this group, as long as you have regulated heating, heat from PSUs is just as good as any other heat.
Sort of: electricity is currently (ha!) about four times more expensive per kWh than gas in the UK. Presumably this is down to a) conversion losses at the power station b) transmission losses c) value - electricity can be used for more purposes in a typical home than gas. If you're heating your home with electricity, you're effectively doing chemical->thermal->[kinetic->]electricity->thermal , rather than just chemical->thermal, and the overall energy efficiency will be accordingly lower.
Seconded. Good information security should ideally be transparent, and with a bit of work on the part of the people implementing it, often can be. Sometimes, it's even possible for the good security to facilitate working practices that wouldn't have previously been considered possible.
Seconded. I have a PC here which, after a mains glitch, sometimes fails to start, even after disconnecting the PSU from the mains for a whole day, or even disconnecting the PSU from the motherboard.
The last time this happened, I disconnected the PSU from the mains and the motherboard and reluctantly started stripping it for parts. Whilst removing the AGP graphics card, the motherboard-mounted LED to warn that the motherboard still had power started flickering on and off (eek!)
I hurriedly started yanking cables from the back, and I'm pretty sure the LED went off when I yanked one of the cables from my externally-powered USB hubs. I reconnected everything back up, and it started first time. My hubs are no longer externally-powered, and I've had no problems since, touch wood.
Hence it's best practice to configure your RAID (be it hardware or software) to use slightly less than the full capacity of a drive, in case you can't get an exact match if/when you need to get a replacement.
If the "magic key" length was something completely random like 6385492, then I would be more suspicious.
But the best backdoors are the ones that are plausibly deniable, so that your potential targets don't get er... spooked and switch from using your software (presumably with other backdoors) to something more secure.
That depends. In my experience, hardware budgets are lumpy - sometimes you can get shiny, expensive new hardware easy as pie, sometimes you need to make do with the hand-me-down desktop machine. Individual software packages are generally pretty easy to get bought, but 'enterprise' software can be very difficult, unless there's a good discount available for educational users (e.g. Microsoft does well here, since it's in their interest to get the next generation of graduates hooked on their stuff). Maintenance/support is generally extremely hard to get bought, which further limits the choice of 'enterprise' software you can obtain.
On the other hand, if you want to play with new Free/Open Source software, you'll have much more chances to do so in education, than in the private sector, simply because at times you'll be expected to achieve miracles with nothing more than some string, bubble gum, sticky tape, your ingenuity, a hand-me-down PC and a fast Internet link. If you like working with FOSS, sometimes having too much money sloshing around is a disadvantage.:-)
Ah, I think I see: you seem to have interpreted my post as meaning that I believed that VT meant VMware, or something like it, would be redundant. I don't, and never have.
Obviously, something will need to be present in order to emulate the rest of the system. What I do hope VT (and AMD's Pacifica) will improve is the CPU virtualization subsystems within VMware and friends, that currently need to provide equivalent mechanisms by various software techniques. Your point froim your blog that just because VT/Pacifica provides hardware mechanisms, doesn't mean they'll be any faster than existing software techniques is entirely valid, of course. I trust VMware developers to benchmark all the options and use VT, software, or some hybrid blend as appropriate. Also, I don't overestimate the overall effect VT/Pacifica will have on complete system virtualization; I'm well aware that VMware's CPU virtualization is already very efficient.
Don't get me wrong; I like VMware, and have done since I was one of the early beta testers. I've plugged it to goodness-knows-how-many organizations since then (without commission!), but I look forward to it delivering even better performance by taking advantage of any useful architectural enhancements as they are provided by new CPUs. I hear that VMware has a product in the pipeline for the x86 Macs, which would make them an extremely interersting proposition for anyone who has to work with and support multiple OSs.
I highly doubt that any work required to make Windows XP work with EFI will drastically, or even noticably affect the speed of the machine. [using BobPaul's corrected quote]
...especially seeing as the Core Duo supports the new Vanderpool Virtualization Technology (VT) extensions, making the x86 architecture now completely virtualizable, meaning that the tricks used by VMware and friends are no longer necessary.
Yes, and, as someone else pointed out elsewhere in the comments for this story, if you need to upgrade the webpad or the communications, it can be done seperately, and quite possibly more cheaply. Also, if one knows that a PDA will be superfluous, taking a small BT-enabled phone is much less encumbering than taking a usable-but-clunky all-in-one device.
I'm warming to the idea of keeping these two devices seperate.:-)
I've been looking into this part of the market recently, as I'd like web and SSH access from anywhere, using WiFi where possible, but falling back to GPRS/3G/GSM if necessary. At first I thought the Palm Treo 650 was likely to be the be-and-end-all, but it appears to be fairly seriously flawed in its current state. The Nokia 770 is interesting, but I think I'd be better served by getting a Palm T|X (which has a reasonable HVGA screen and built-in Bluetooth and 802.11b, and of course is PalmOS-based, so plenty of independent/free software) for about 200GBP and some kind of Bluetooth-capable phone for 70-80GBP.
Ideally, I'd like both the PDA/webpad to be in the same physical case as the phone, since I'm more likely to keep the PDA with me at all times, but that doesn't really seem to be a practical option right now... Suggestions welcome, though, if anyone's got any!
Even better, make your/boot or / partition the only active (aka bootable) partition, and put GRUB or LILO on the bootsector of that partition, leaving Windows' MBR code in the MBR. The boot sequence will then be BIOS->Windows MBR->Bootsector of Active Partition (GRUB/LILO)->kernel or bootsector of C:->/sbin/init or NTLOADR.
When calling for a boycott of a company because of their business practices, it seems somewhat wrong to only boycott that one company and say it's ok to buy from other people who have exactly the same business practices.
Not really; if a boycott can get nVidia to change their policies and behaviour, it would set an example to many other companies, because nVidia is viewed, rightly or wrongly, as something of a leader. Certainly, they were one of the first hardware manufacturers to go down the binary-only-driver route for Linux. It can be interpreted that ATI's recent actions in also going binary-only (rather than their previous policy of providing useful documentation to the XFree/Xorg project) are simply following nVidia's lead.
Maybe in America as in Canada we need to put aside our petty differences and vote en bloc to push the neocons out of power before the definition of facism RMS casts at America today becomes applicable in the U.S. and in Canada.
Be very careful who you vote for if you do so; the UK tried that in 1997 in order to get rid of the Conservative government of the previous 18 years.
Seems like humans are wired to try Cargo Cult-like approaches first:
Eventually the cargo cults petered out. But, from time to time, the term "cargo cult" is invoked as an English language idiom, to mean any group of people who imitate the superficial exterior of a process or system without having any understanding of the underlying substance.
I had the same idea back around 1989 when I was about 15. Luckily for my criminal record, I never did get my program to print EAN-13 barcodes working completely right...
WD also sells IDE and SATA RE and RE2 enterprise drives with MTBFs of 1 - 1.2 million hours. Why would anybody want to halve the MTBF of their drive by getting an SE drive just to save $30?
Because, according to the WD literature for the TLER functionality of their RAID Edition drives, they are designed not to attempt 'heroic recovery' for longer than 7 seconds on the assumption that RAID at a higher level will use ECC to recover any errors. RE drives are not intended for use in plain stripe sets, or as single drives.
Oh, and I have 3 80G SE drives from 2002 which still run fine.
It's naïve to assume that the relative reliabilty of each vendor's drives remain static compared with each other. The 'top dog' position rotates irregularly, as does the 'PoS' position. Currently, I'd say that Seagate are top for reliability, with WD close behind with their "special edition" models (and each is backed by 5 years and 3 years warranty, respectively). I'd say that, right now, anecdotally, Maxtor are in the 'PoS' position; everyone I know whose lost a drive recently was using a Maxtor. OTOH, I have a 120M Maxtor drive from 1993 which still runs and a 1.3G WD from 1995 which was troublesome by 1997 (but still runs if you tap it right on power-up). I also have a 1.2G Quantum from 1992, a 4.3G Fujitsu from 1997, a 6G IBM from 1998, a 30G Maxtor from 1999, 3*80G WD SEs from 2002, a 20G 2.5" Hitachi from 2002 and 2*200G Seagates and a 60G 2.5" Hitachi from 2004 which still run with no problems.
The important thing is to research the reliability of the vendors on the market at the time of purchase, and not to rely upon either positive or negative reputation.
Next most important is to keep your drives cool, on a clean power supply, and away from vibrations.
Make sure to future-proof. Even if you don't want a dual-core Athlon 64 today, get a chipset that would support it. This will cut down your chipset choices further.
Thinking about the possible future upgrades is, from my past experience, not worth the money. I have never ever upgraded just the CPU (even on high-end servers[*]). By the time you decide that the present CPU is too slow the new CPUs will use better memory interfaces, which means at least the memory upgrade as well, but the new memory interface usually means the new mainboard infrastructure or new CPU socket.
Sometimes it works out, sometimes it doesn't. For example, I have an Asus P2B (i440BX-based) that I bought in August 1998 with a PII-266 which has since been upgraded to a Celeron 500, and subsequently a PIII 450 (when I found a pull for 5GBP at a fair). I avoided the problem with memory upgrades by buying PC100 memory at the outset as it was only a few GBP more, and avoided problems with dodgy budget PC66 memory which was doing the rounds back then (I just use Crucial now, but still buy the fastest memory a board can use, even if the CPU won't use it at full speed).
OTOH, my most recent boards (Gigabyte GA-8PE667 Ultras - i845PE-based) were supposed to be able to take CPUs with upto 18x667MHz=12GHz clockspeeds, but in actual fact, the fastest they can take is 3.06GHz Northwoods, which aren't really a cost-effective upgrade from the P4 2.40Bs I have right now. If I'd waited a while, though, I could have picked up i865PE-based boards that could take CPUs with 800MHz FSB, such as Prescott (which is still only marginally useful, but at least it is an upgrade in some circumstances).
OO.org on the other hand doesn't work 100% the vast majority of the time. Anything more than basic text documents simply doesn't translate well. You'll get the full text alright, just not spaced and formatted correctly. If you don't mind things like docs constantly changing from 2 to 3 pages then fine use OO.org to work with Word docs. Otherwise it should be avoided.
If you're running on a non-Windows OS, do you have the Microsoft fonts loaded? If not, you should try that as a first step to solve the problem you describe above.
Also, if you change your printer driver in Windows and the resolution changes, Office will reformat your documents anyway. OO.org won't do that; what's on a single page on one person's computer will stay on a single page on anyone else's (though using PDF is the most reliable way to guarantee WYSIWYG anywhere).
Slashdot Mirror
Yes, and the components that I've seen listed on the spec sheets of the machines sold by both Aldi and Lidl are pretty good - Seagate discs, Samsung memory, ASUS motherboards, Sony optical drives, etc. You'd be hard pressed to put together a machine of similar spec yourself, and that's excluding the software (which, given this is /. may actually have negative value) and warranty it all for 3 or 5 years as both Aldi and Lidl do. About the only concern with these machines is the quality of the phone support, but again, given this is /. ....
Interestingly, a German friend told me a few years back that Medion/Aldi had secured 11.6% of the PC market in Germany.
Pretty similar: look into the bsdcomp, deflate, predictor1 options in pppd and the CONFIG_PPP_BSDCOMP and CONFIG_PPP_DEFLATE kernel options. Of course, the hard thing - at least in the UK - is finding an ISP who supports either of those options. I've never experienced that, so I've always had to rely upon my modems' built-in compression. With serial port tuning (i.e. 230400baud), I was able to achieve 20Kbyte/s on a V.90 modem downloading files consisting entirely of 0x00 or 0xff. In reality 5-7Kbyte/s was much more realistic for ASCII text.
Sort of: electricity is currently (ha!) about four times more expensive per kWh than gas in the UK. Presumably this is down to a) conversion losses at the power station b) transmission losses c) value - electricity can be used for more purposes in a typical home than gas. If you're heating your home with electricity, you're effectively doing chemical->thermal->[kinetic->]electricity->thermal , rather than just chemical->thermal, and the overall energy efficiency will be accordingly lower.
Seconded. Good information security should ideally be transparent, and with a bit of work on the part of the people implementing it, often can be. Sometimes, it's even possible for the good security to facilitate working practices that wouldn't have previously been considered possible.
The last time this happened, I disconnected the PSU from the mains and the motherboard and reluctantly started stripping it for parts. Whilst removing the AGP graphics card, the motherboard-mounted LED to warn that the motherboard still had power started flickering on and off (eek!)
I hurriedly started yanking cables from the back, and I'm pretty sure the LED went off when I yanked one of the cables from my externally-powered USB hubs. I reconnected everything back up, and it started first time. My hubs are no longer externally-powered, and I've had no problems since, touch wood.
Hence it's best practice to configure your RAID (be it hardware or software) to use slightly less than the full capacity of a drive, in case you can't get an exact match if/when you need to get a replacement.
But the best backdoors are the ones that are plausibly deniable, so that your potential targets don't get er... spooked and switch from using your software (presumably with other backdoors) to something more secure.
I would love to see some kind of accountability system that did not gut opensource software.
Such a law could be drafted to explicitly exclude any software that came with full, unobfuscated source code.
Whether such a law will or would be drafted that way is another matter entirely.
*Fewer chances to play with new tech
That depends. In my experience, hardware budgets are lumpy - sometimes you can get shiny, expensive new hardware easy as pie, sometimes you need to make do with the hand-me-down desktop machine. Individual software packages are generally pretty easy to get bought, but 'enterprise' software can be very difficult, unless there's a good discount available for educational users (e.g. Microsoft does well here, since it's in their interest to get the next generation of graduates hooked on their stuff). Maintenance/support is generally extremely hard to get bought, which further limits the choice of 'enterprise' software you can obtain.
On the other hand, if you want to play with new Free/Open Source software, you'll have much more chances to do so in education, than in the private sector, simply because at times you'll be expected to achieve miracles with nothing more than some string, bubble gum, sticky tape, your ingenuity, a hand-me-down PC and a fast Internet link. If you like working with FOSS, sometimes having too much money sloshing around is a disadvantage. :-)
Obviously, something will need to be present in order to emulate the rest of the system. What I do hope VT (and AMD's Pacifica) will improve is the CPU virtualization subsystems within VMware and friends, that currently need to provide equivalent mechanisms by various software techniques. Your point froim your blog that just because VT/Pacifica provides hardware mechanisms, doesn't mean they'll be any faster than existing software techniques is entirely valid, of course. I trust VMware developers to benchmark all the options and use VT, software, or some hybrid blend as appropriate. Also, I don't overestimate the overall effect VT/Pacifica will have on complete system virtualization; I'm well aware that VMware's CPU virtualization is already very efficient.
Don't get me wrong; I like VMware, and have done since I was one of the early beta testers. I've plugged it to goodness-knows-how-many organizations since then (without commission!), but I look forward to it delivering even better performance by taking advantage of any useful architectural enhancements as they are provided by new CPUs. I hear that VMware has a product in the pipeline for the x86 Macs, which would make them an extremely interersting proposition for anyone who has to work with and support multiple OSs.
I'm warming to the idea of keeping these two devices seperate. :-)
Ideally, I'd like both the PDA/webpad to be in the same physical case as the phone, since I'm more likely to keep the PDA with me at all times, but that doesn't really seem to be a practical option right now... Suggestions welcome, though, if anyone's got any!
Even better, make your /boot or / partition the only active (aka bootable) partition, and put GRUB or LILO on the bootsector of that partition, leaving Windows' MBR code in the MBR. The boot sequence will then be BIOS->Windows MBR->Bootsector of Active Partition (GRUB/LILO)->kernel or bootsector of C:->/sbin/init or NTLOADR.
Possibly Giovanni Gentile, described by Mussolini as 'the philosopher of Fascism' (and who better to judge?) who said:
(this quote is frequently attributed to Mussolini himself, and publiceye.org questions whether either Mussolini or Gentile ever said it).
Not really; if a boycott can get nVidia to change their policies and behaviour, it would set an example to many other companies, because nVidia is viewed, rightly or wrongly, as something of a leader. Certainly, they were one of the first hardware manufacturers to go down the binary-only-driver route for Linux. It can be interpreted that ATI's recent actions in also going binary-only (rather than their previous policy of providing useful documentation to the XFree/Xorg project) are simply following nVidia's lead.
Be very careful who you vote for if you do so; the UK tried that in 1997 in order to get rid of the Conservative government of the previous 18 years.
I had the same idea back around 1989 when I was about 15. Luckily for my criminal record, I never did get my program to print EAN-13 barcodes working completely right...
Because, according to the WD literature for the TLER functionality of their RAID Edition drives, they are designed not to attempt 'heroic recovery' for longer than 7 seconds on the assumption that RAID at a higher level will use ECC to recover any errors. RE drives are not intended for use in plain stripe sets, or as single drives.
Oh, and I have 3 80G SE drives from 2002 which still run fine.
The important thing is to research the reliability of the vendors on the market at the time of purchase, and not to rely upon either positive or negative reputation.
Next most important is to keep your drives cool, on a clean power supply, and away from vibrations.
Thinking about the possible future upgrades is, from my past experience, not worth the money. I have never ever upgraded just the CPU (even on high-end servers[*]). By the time you decide that the present CPU is too slow the new CPUs will use better memory interfaces, which means at least the memory upgrade as well, but the new memory interface usually means the new mainboard infrastructure or new CPU socket.
Sometimes it works out, sometimes it doesn't. For example, I have an Asus P2B (i440BX-based) that I bought in August 1998 with a PII-266 which has since been upgraded to a Celeron 500, and subsequently a PIII 450 (when I found a pull for 5GBP at a fair). I avoided the problem with memory upgrades by buying PC100 memory at the outset as it was only a few GBP more, and avoided problems with dodgy budget PC66 memory which was doing the rounds back then (I just use Crucial now, but still buy the fastest memory a board can use, even if the CPU won't use it at full speed).
OTOH, my most recent boards (Gigabyte GA-8PE667 Ultras - i845PE-based) were supposed to be able to take CPUs with upto 18x667MHz=12GHz clockspeeds, but in actual fact, the fastest they can take is 3.06GHz Northwoods, which aren't really a cost-effective upgrade from the P4 2.40Bs I have right now. If I'd waited a while, though, I could have picked up i865PE-based boards that could take CPUs with 800MHz FSB, such as Prescott (which is still only marginally useful, but at least it is an upgrade in some circumstances).
Thankfully, now no longer trading. According to a hand-written note on the door of my local 'Time' branch, they even stiffed their retail employees.
If you're running on a non-Windows OS, do you have the Microsoft fonts loaded? If not, you should try that as a first step to solve the problem you describe above.
Also, if you change your printer driver in Windows and the resolution changes, Office will reformat your documents anyway. OO.org won't do that; what's on a single page on one person's computer will stay on a single page on anyone else's (though using PDF is the most reliable way to guarantee WYSIWYG anywhere).