BTW: there's an even a bigger difference in the red channel.
They could have crappy monitors. OR the monitors aren't adjusted correctly. On mine 294d4a is darker and greener.
Which brings us back to what the story is about. Even if the people can't _perceive_ the difference at he/she wants them to see the same things on their monitors.
It's not such a dumb request. But I still don't see the point yet. Since if it's for a product, is that colour thing going to go all the way to the customer end? I suppose it could save time making actual samples and flying them here and there.
I agree. It's like having access to inner private parameters of an object/class that's normally not exposed publicly.
It is likely that people actually know exactly how many objects they see e.g. walk into a room and know instantly how many chairs there are.
However that info is abstracted away under layers and layers of abstraction - e.g. one, two, many, dozens, hundreds, thousands, enough, not enough - after all most people spot desks without chairs quite quickly, this is not necessarily such an easy thing.
These abstractions were probably very useful millenia ago. There's no point knowing that there are 23312 wildebeest on the plain if you don't even have a number system, much less express it to someone who doesn't. Just have to give the appropriate grunt(s) or clicks.
And these abstractions probably allow us to avoid the detail and focus on the big picture - the guy has problems going to the supermarket or the beach.
Even if you can count the number of hairs on a lioness instantly, I doubt the lioness bothers remembering how many bites of meat you make up - it's probably "enough for me and cubs, or need one more".
I suspect some of these people would be troubled and have difficult working if you gave them 1001 bolts and only 1000 nuts and told them to fasten stuff together. Most normal people won't even notice till the last one, and then they'll just shrug and go whatever...
Modern software can easily count how many light and dark pixels there are in picture. It has difficulty seeing how many chairs. But soon programs will count X chairs of type A, Y chairs of type B etc, but the next step then is "few, many, enough, not enough".
Guardian article (and Savant) only says european record which appears accurate.
Slashdot story should probably say holds "a record" not "the record", which in typical Slashdot context would either imply world record or USA record;).
Just because at a particular time something (e.g. a brain for instance) isn't useful doesn't mean it's not worth keeping it around.
Doh.
Anyway typically if all power is lost a modern Navy ship wouldn't be going anywhere fast (hopefully not anyway...), so you'd wait for conditions to improve or power to be restored. If the situation warrants it, then you might get on a lifeboat with a paddle and the sextant (and GPS unit if you have one;) ).
Software Buildings/Cars/etc design design alpha prelim design/models v1.0 functional model v2.0 blueprint v3.0 prototype v3.1 production v4.0 production with more features.
The trouble with software is the blueprint and models actually "work" AND the customer is often willing to buy it already (or easily convinced to;) ).
The other trouble with software is making those models is just as expensive as it is to make the "real thing" ("final" version).
The production cost of making X copies of software is pretty low, in contrast to the cost of making X copies of bridges and cars.
How'd you: 1) Stop really fast. 2) Coast to a stop. 3) maintain the same speed 4) Rest your foot temporarily without stopping really fast. 5) Achieve precision deceleration (for going through a corner at the right speed).
I think you still need an explicit STOP control as well as an explicit GO control. Implicit STOP is not good enough.
Yah. I wonder if mechanics have ever asked people (their friends;) ) to shift up, shift down, shift up, shift down, turn left, turn right, press accelerator, press brake, and then start...
"A single (S)ATA HDD requires a single (S)ATA connector. So in addition to that $100 drive you'll need a $5,000 box to put it in and to translate those several slow connections into Fibre or something else fast enough to make this idea reasonable"
I don't see your point at all. With LTO-3 your USD5K only gets you the drive, you still need to BUY a server to stick that drive to, and I'm sure you have the same problem of "slow connections" whatever that means (don't know what you're getting at).
Whereas so far most modern x86 servers have ATA connectors. If hot swap SATA becomes commonplace then you've already got your box - it's your server - I already see servers with hot swap SATA. So even though one LTO drive has faster transfer rates than one ATA HDD, you can stick 5 HDDs to 5 servers and back them up all at the same time, take the HDDs out and put them into archival storage. Whereas with the LTO solution - the 5 servers have to backup through that one server with the LTO3 drive.
If a server doesn't have a hotswap SATA maybe USB caddy HDD drives might not be too bad a solution at 30MB/sec.
"HDDs are really, really expensive solutions for any sort of archving, or for off-site storage. I can put a pack of 10 LTO-3 tapes in a box and ship it around the world for $20. "
You can ship HDDs around too.
And while you can ship your 10 LTO-3 tapes, if you need to read them you may wish to make sure there's a USD5K LTO-3 drive at the destination. Whereas finding a machine that can read SATA/ATA drives isn't as hard.
HDDs really expensive? Go do the figures yourself. AFAIK LTO3 tapes are 400GB for USD140. That's USD0.35/GB (I'm ignoring the compressed figures).
But they don't include a USD5000 drive.
Whereas ATA HDDs are 200GB for USD100. That's USD0.50/GB. They come with drive included.
As I mentioned in my previous post you need to have more than 83 tapes _per_ LTO3 drive, only then it starts getting cheaper than 200GB HDDs.
(AFAIK you only need 25 sets for the 6 week days, 4 weeks, 12 months, 3 years thing. 29 sets if 7 years.)
BUT if you start to need more than one LTO3 _drive_ before you hit 83 tapes then the LTO3 solution is more expensive. Remember - if you have one LTO3 drive at point A, and turns out you need another drive at point B (where you're shipping them to), then you now need more than 166 tapes, to be cheaper than HDDs.
Don't forget with 166 HDDs you can just stick them to 166 servers if you have that many servers. You don't have to look for a drive, since the "backup media" is a drive.
Also, I'm willing to bet the HDD prices are going to drop. With this scheme you don't get locked in to an expensive drive and media. Whereas if LTO3 media or drives become more expensive you're screwed. You still have to buy LTO-3 media and drives.
So far tape sure still seems like a con. Unless you're going to have hundreds of tapes per drive.
The tape drives would probably start chewing up tapes long before you have hundreds of tapes. So far the tape drive in my prev workplace started chewing up tapes.
Linux is fine for servers and people who need a POSIX/Unix style environment.
But really it isn't that good as a desktop IMO. At my office I have SuSE 9.1 (KDE) on a Dell P4 3GHz with 2GB RAM, and the display is a GeForce 5200.
For one the display/UI sure is sluggish compared to my machine at home (athlon XP 2500+ 512MB RAM with GF Ti4200). Yah it's X. But erm, X was this sluggish in the 90s on 200-500MHz machines with S3 video cards. This is a 3GHz machine with a GF5200. Looks like I have to install gnome - hopefully that's much better.
Worse: somehow at least twice I managed to get the display to not update- end up with a blank/corrupted screen. Once when I was using VMWare (going to full screen etc). Another time when I was testing out the screen savers. The apps are still there running, but I can't see them, or get back to them - the screen doesn't refresh properly. This in just the first week or two of using SuSE 9.1.
Sure the Linux kernel and other close subsystems are fairly stable. But it looks like the GUI stuff needs a LOT of work.
The kernel and other low level stability etc doesn't help me when I have a few GUI applications full of unsaved data which I can't get to though I know they are still there running.
Perhaps I should have installed and used VNC. But hey, that sure doesn't show it's ready.
Yes I know how to use Yast online update, and I did. That brings me to my next issue: somehow during the kernel update it buggered things up so I couldn't boot. Yeah I used XFS (was going to stick with the default - reiserfs but my boss said use XFS), (I did install the XFS updates - you have to, in order to install for 9.1 - kinda a bad sign IMO).
Fortunately I knew enough to boot from the SuSE install disk and copy the old kernel rpms, then chroot and rpm install them. Then it booted. The kernel update somehow worked when I tried again. Weird. This is kinda "windows"-ish don't you think?
I'm not sure why all this happened. Maybe it's the vmware stuff.
On SuSE 9.1's KDE the copy and paste thing sure sucks (it's better than the 90s though anyone remember the bad old days?). Can't copy a selection from Konsole using any decent key combination (e.g. involving ctrl)- have to use the frigging mouse. Also, the icons are rather indistinctive - all the K apps seem to use the same blooming shades. Makes it hard to notice stuff - like I've got a new message etc. They need a bit more work.
Sure looks that by the time the Linux Desktop is "ready", it'll be just as unreliable as Win2K/WinXP, or worse, and probably just as crappy.
BTW, mozilla is using 126MB of RAM... X is using 76MB. Fortunately I have 2GBs of RAM.
However, on windows 2K and XP I've managed to hit a limit where I can't open any more apps - no more IE windows, even though I have RAM left... If you want to know - the windows task bar actually does give you a scrollbar if you have lots of task buttons.
Haven't run into that problem on Linux yet:).
KDE is the default for SuSE 9.1. Sure you can probably customize X and the GUI, avoid the indistinct icons and other stuff. But the real skill AND point is in picking the "right" _defaults_.
Who cares if you have access to 1000 themes. If the defaults suck, then the GUI sucks.
Joe Schmoe is going to be using the defaults. And Helpdesk is going to be assuming the defaults.
1) You can't write to the HDD if it's on a shelf. It's removed from the computer after it's used, just like any other removable media.
2) You can't (and wouldn't want to) write protect the media you are backing up to. If the trojan screws up data whilst you are backing it's the same whether you're backing up to HDD or some other media - e.g. floppy or DVD+R or CD-R. If you can't fill a 200GB HDD, then just use smaller and cheaper HDDs. It's just like using lower capacity media.
Your argument is valid for restoring/reading data, but if you aren't able to make sure your systems don't corrupt your backups during a restore then you should consider getting hardware write blocking devices - go search on some forensic products -(e.g. Fastbloc, Nowrite) there are a few companies around. Expect to pay some money for that.
So far I haven't had problems with trojans or viruses on my computers, so such products aren't worth it for me.
Uh, my removable media are _harddisks_. I've got my HDDs in caddies/trays (with fans) so I can slide them out easily.
Doing the figures, HDDs aren't really that expensive compared to other media especially when you factor in the performance, reliability, stability and convenience. Buy one or two 200GB HDDs, backup everything (two or more copies just to be sure:) ), then store/archive the HDDs.
The stability of data on magnetic disks is pretty good. The only problem usually is the electronics failing or the mechanical stuff failing. But if you're not using/abusing the drives, the shelf life is pretty good.
When the time comes to migrate the data off the obsolete 200GB drives so you can still read the data, it's a lot faster and simpler to copy the data off the 200GB drives than it is to copy from other removable media. The transfer rates of optical or tape media are pretty bad. OK DVD stuff isn't that bad, but writing isn't that fast... I personally believe tapes and tape drives are a big con-job nowadays.
It'll be a lot better if hotswap SATA drives and their caddies become common and cheap. Then backing up wouldn't require a power down.
1) The virtual machines are not written by the same people writing the apps. While you might be a better programmer than those writing virtual machines, as Bugtraq shows, the vast bulk of programmers aren't very good at writing C or C++ programs.
2) Even if there is a vulnerability in the virtual machine, the attacker does not usually have direct access to the virtual machine - the attacker can only go through the app, and the vulnerability might not be able to be exploited.
3) Often a bug in the VM would be found fairly quickly once it gets used by everyone because it's at a low level - it's like having your CPU not do something correctly. Only the really obscure functionality would be untouched - but if they're mostly untouched they're less likely to be problems in practice.
4) If there's a security bug in the VM, you only have to fix one thing- the VM and your apps should still work. In contrast fixing C buffer overflows etc for dozens of different programs is not that easy. Multiply that by tons of different apps by thousands of programmers and you should see the magnitude of the problem.
5) It's an example of specialisation and delegation. Go look up how many vulnerabilities there are in the Perl/Java/Python VMs. There haven't been many announced over the years. The few people who are good at writing VMs write VMs. Those that are crap at programming like me, write in Perl etc and so don't have to deal with keeping track of pointers, allocating enough memory etc, regularly screwing things up so that "an attacker can execute arbitrary code of the attacker's choice".
IMO in general women (or perhaps just the louder/whinier ones) tend to view obstacles as problems whereas in general men view obstacles as challenges or even something fun...
I don't even think that guys playing games go "I wish I had muscles like him etc". In fact during the game they are more likely to subconsciously go "This is _me_! This is me kicking ass! Yeah!! Eat that!". And don't forget "I 0wn3d J00!".
I wonder whether most women actually see it as a problem or it's just the whiny ones.
Rakfisk doesn't seem far from the way lots of people make fish sauce (e.g. Ancient Romans, Vietnamese etc).
Doesn't even look like it gets that much of a chance to rot unlike Surstromming, or involve saturation in a pretty _toxic_ chemical (e.g. lutefisk in lye aka NaOH).
But maybe you've tasted lutefisk and rakfisk and the latter actually tastes worse?
"What you really need is a device with its own computing power, such as an iButton. You then have software which sends a challenge from the server to the iButton, calculates a hash, then calculates another hash on that hash using standard password techniques."
Yeah, but don't forget the most important thing - usually the old system is there because it did have some value in the first place.
First do no harm.
And so far we're still based on DNA after all these years. Maybe we're overdue for a rewrite, but so far I'm not too sure about the problem definition...
It seems the objectives of that project were quite easy to _define_. AND this sort of problem has been successfully done many times before - just not as fast:).
Whereas in many of these big IT projects it is unclear what the customer is trying to achieve. So many parties with different goals, some probably conflicting - though they may not know it yet (and at least some would want to milk the customer as much without the customer mooing or kicking;) ). So it is probably hard to define what they want to their satisfaction.
They often will have no idea on many important decisions, so a short-term profit minded vendor (which = most) will just pick the easiest to do which complies to what the customer _said_ they wanted and not bring the issue up - educating them will just take too long. And they'll have plenty of conflicting ideas on the relatively less technically important decisions = long meeting over what colour some thingy should be (sure it's important but it doesn't need to involve so many people - it'll just waste even more manhours).
Imagine if there's really only one pedal in your car and you press it once to activate the clutch, twice quickly in the same spot to accelerate, and three times quickly to brake.
Doubleclicking is the hard stuff- it requires _coordination_ in time and space. Clicking two different spots twice rapidly that are far apart does not trigger a doubleclick. For newbies the effort of clicking two spots rapidly often causes them to move the mouse greatly and thus not trigger a doubleclick. You can often visibly see the strain and stress in their faces and bodies when they try to doubleclick - often gritting their teeth or something...
Worse - nowadays most browsers a single click is what it takes to "execute/launch/activate" a link. Whereas you need a double click for other areas (word processing etc)... Now that's confusing.
If people had settled for one button for just select/deselect (without activating) and another button for activate, perhaps things might actually be easier.
I'm not sure about the drag and drop thing too... Perhaps clicking and holding down might be hard for some people.
Intelligence is not just knowing _absolute_ semantic meaning - e.g. that a cow is a cow and grass is grass. And being able to group grass with other grasses and cows with other bovine animals.
It's being able to understand the statement that cow is to grass in a similar way that balleen whales are to krill.
And then now knowing something about krill from that even if you didn't know what krill was at all.
It's not just knowing the "absolute value" of the meaning - or even that these two objects are linked or close in the same area ( which seems to be the level which most current AI are at).
It's more like kind of knowing the "vector/direction" they are linked, and being able to organize other objects that are related in similar ways in a similar vector. Thus you can learn about things by analogies and metaphors AND even create new things with those.
Would pump out more BS but I have to go for dinner:)
Slashdot Mirror
BTW: there's an even a bigger difference in the red channel.
They could have crappy monitors. OR the monitors aren't adjusted correctly. On mine 294d4a is darker and greener.
Which brings us back to what the story is about. Even if the people can't _perceive_ the difference at he/she wants them to see the same things on their monitors.
It's not such a dumb request. But I still don't see the point yet. Since if it's for a product, is that colour thing going to go all the way to the customer end? I suppose it could save time making actual samples and flying them here and there.
Yah. Maybe the animals do, and the humans don't.
That's why the animals ran...
I agree. It's like having access to inner private parameters of an object/class that's normally not exposed publicly.
It is likely that people actually know exactly how many objects they see e.g. walk into a room and know instantly how many chairs there are.
However that info is abstracted away under layers and layers of abstraction - e.g. one, two, many, dozens, hundreds, thousands, enough, not enough - after all most people spot desks without chairs quite quickly, this is not necessarily such an easy thing.
These abstractions were probably very useful millenia ago. There's no point knowing that there are 23312 wildebeest on the plain if you don't even have a number system, much less express it to someone who doesn't. Just have to give the appropriate grunt(s) or clicks.
And these abstractions probably allow us to avoid the detail and focus on the big picture - the guy has problems going to the supermarket or the beach.
Even if you can count the number of hairs on a lioness instantly, I doubt the lioness bothers remembering how many bites of meat you make up - it's probably "enough for me and cubs, or need one more".
I suspect some of these people would be troubled and have difficult working if you gave them 1001 bolts and only 1000 nuts and told them to fasten stuff together. Most normal people won't even notice till the last one, and then they'll just shrug and go whatever...
Modern software can easily count how many light and dark pixels there are in picture. It has difficulty seeing how many chairs. But soon programs will count X chairs of type A, Y chairs of type B etc, but the next step then is "few, many, enough, not enough".
Guardian article (and Savant) only says european record which appears accurate.
;).
Slashdot story should probably say holds "a record" not "the record", which in typical Slashdot context would either imply world record or USA record
Sure thing. Topple the closet backwards and get in your lead lined coffin^H^H^H^H^H^Hcloset and wait for a week or two.
Your reception should then be out of this world.
More bugs/security issues?
Helps when visibility improves?
;) ).
Just because at a particular time something (e.g. a brain for instance) isn't useful doesn't mean it's not worth keeping it around.
Doh.
Anyway typically if all power is lost a modern Navy ship wouldn't be going anywhere fast (hopefully not anyway...), so you'd wait for conditions to improve or power to be restored. If the situation warrants it, then you might get on a lifeboat with a paddle and the sextant (and GPS unit if you have one
Did he say it was true?
AFAIK far most jokes on Slashdot aren't marked as such, unless the poster thinks people won't get it.
Uh, software is already treated differently.
;) ).
Something like this:
Software Buildings/Cars/etc
design design
alpha prelim design/models
v1.0 functional model
v2.0 blueprint
v3.0 prototype
v3.1 production
v4.0 production with more features.
The trouble with software is the blueprint and models actually "work" AND the customer is often willing to buy it already (or easily convinced to
The other trouble with software is making those models is just as expensive as it is to make the "real thing" ("final" version).
The production cost of making X copies of software is pretty low, in contrast to the cost of making X copies of bridges and cars.
How'd you:
1) Stop really fast.
2) Coast to a stop.
3) maintain the same speed
4) Rest your foot temporarily without stopping really fast.
5) Achieve precision deceleration (for going through a corner at the right speed).
I think you still need an explicit STOP control as well as an explicit GO control. Implicit STOP is not good enough.
Yah. I wonder if mechanics have ever asked people (their friends ;) ) to shift up, shift down, shift up, shift down, turn left, turn right, press accelerator, press brake, and then start...
"A single (S)ATA HDD requires a single (S)ATA connector. So in addition to that $100 drive you'll need a $5,000 box to put it in and to translate those several slow connections into Fibre or something else fast enough to make this idea reasonable"
I don't see your point at all. With LTO-3 your USD5K only gets you the drive, you still need to BUY a server to stick that drive to, and I'm sure you have the same problem of "slow connections" whatever that means (don't know what you're getting at).
Whereas so far most modern x86 servers have ATA connectors. If hot swap SATA becomes commonplace then you've already got your box - it's your server - I already see servers with hot swap SATA. So even though one LTO drive has faster transfer rates than one ATA HDD, you can stick 5 HDDs to 5 servers and back them up all at the same time, take the HDDs out and put them into archival storage. Whereas with the LTO solution - the 5 servers have to backup through that one server with the LTO3 drive.
If a server doesn't have a hotswap SATA maybe USB caddy HDD drives might not be too bad a solution at 30MB/sec.
"HDDs are really, really expensive solutions for any sort of archving, or for off-site storage. I can put a pack of 10 LTO-3 tapes in a box and ship it around the world for $20. "
You can ship HDDs around too.
And while you can ship your 10 LTO-3 tapes, if you need to read them you may wish to make sure there's a USD5K LTO-3 drive at the destination. Whereas finding a machine that can read SATA/ATA drives isn't as hard.
HDDs really expensive? Go do the figures yourself. AFAIK LTO3 tapes are 400GB for USD140. That's USD0.35/GB (I'm ignoring the compressed figures).
But they don't include a USD5000 drive.
Whereas ATA HDDs are 200GB for USD100. That's USD0.50/GB. They come with drive included.
As I mentioned in my previous post you need to have more than 83 tapes _per_ LTO3 drive, only then it starts getting cheaper than 200GB HDDs.
(AFAIK you only need 25 sets for the 6 week days, 4 weeks, 12 months, 3 years thing. 29 sets if 7 years.)
BUT if you start to need more than one LTO3 _drive_ before you hit 83 tapes then the LTO3 solution is more expensive. Remember - if you have one LTO3 drive at point A, and turns out you need another drive at point B (where you're shipping them to), then you now need more than 166 tapes, to be cheaper than HDDs.
Don't forget with 166 HDDs you can just stick them to 166 servers if you have that many servers. You don't have to look for a drive, since the "backup media" is a drive.
Also, I'm willing to bet the HDD prices are going to drop. With this scheme you don't get locked in to an expensive drive and media. Whereas if LTO3 media or drives become more expensive you're screwed. You still have to buy LTO-3 media and drives.
So far tape sure still seems like a con. Unless you're going to have hundreds of tapes per drive.
The tape drives would probably start chewing up tapes long before you have hundreds of tapes. So far the tape drive in my prev workplace started chewing up tapes.
Linux is fine for servers and people who need a POSIX/Unix style environment.
:).
But really it isn't that good as a desktop IMO. At my office I have SuSE 9.1 (KDE) on a Dell P4 3GHz with 2GB RAM, and the display is a GeForce 5200.
For one the display/UI sure is sluggish compared to my machine at home (athlon XP 2500+ 512MB RAM with GF Ti4200). Yah it's X. But erm, X was this sluggish in the 90s on 200-500MHz machines with S3 video cards. This is a 3GHz machine with a GF5200. Looks like I have to install gnome - hopefully that's much better.
Worse: somehow at least twice I managed to get the display to not update- end up with a blank/corrupted screen. Once when I was using VMWare (going to full screen etc). Another time when I was testing out the screen savers. The apps are still there running, but I can't see them, or get back to them - the screen doesn't refresh properly. This in just the first week or two of using SuSE 9.1.
Sure the Linux kernel and other close subsystems are fairly stable. But it looks like the GUI stuff needs a LOT of work.
The kernel and other low level stability etc doesn't help me when I have a few GUI applications full of unsaved data which I can't get to though I know they are still there running.
Perhaps I should have installed and used VNC. But hey, that sure doesn't show it's ready.
Yes I know how to use Yast online update, and I did. That brings me to my next issue: somehow during the kernel update it buggered things up so I couldn't boot. Yeah I used XFS (was going to stick with the default - reiserfs but my boss said use XFS), (I did install the XFS updates - you have to, in order to install for 9.1 - kinda a bad sign IMO).
Fortunately I knew enough to boot from the SuSE install disk and copy the old kernel rpms, then chroot and rpm install them. Then it booted. The kernel update somehow worked when I tried again. Weird. This is kinda "windows"-ish don't you think?
I'm not sure why all this happened. Maybe it's the vmware stuff.
On SuSE 9.1's KDE the copy and paste thing sure sucks (it's better than the 90s though anyone remember the bad old days?). Can't copy a selection from Konsole using any decent key combination (e.g. involving ctrl)- have to use the frigging mouse. Also, the icons are rather indistinctive - all the K apps seem to use the same blooming shades. Makes it hard to notice stuff - like I've got a new message etc. They need a bit more work.
Sure looks that by the time the Linux Desktop is "ready", it'll be just as unreliable as Win2K/WinXP, or worse, and probably just as crappy.
BTW, mozilla is using 126MB of RAM... X is using 76MB. Fortunately I have 2GBs of RAM.
However, on windows 2K and XP I've managed to hit a limit where I can't open any more apps - no more IE windows, even though I have RAM left... If you want to know - the windows task bar actually does give you a scrollbar if you have lots of task buttons.
Haven't run into that problem on Linux yet
KDE is the default for SuSE 9.1. Sure you can probably customize X and the GUI, avoid the indistinct icons and other stuff. But the real skill AND point is in picking the "right" _defaults_.
Who cares if you have access to 1000 themes. If the defaults suck, then the GUI sucks.
Joe Schmoe is going to be using the defaults. And Helpdesk is going to be assuming the defaults.
The LTO-3 drives cost a fair bit tho right? USD5000+?
That pays for a LOT of 200GB HDDs. A single HDD can go up to 50-60MB/sec (uncompressed transfer rates) which isn't too bad.
I believe the LTO3 media is actually 400GB. 800GB is for compressed data (if you can compress it).
So the figures are:
LTO3 400/800GB media: USD140 for 400GB (real capacity)
LTO3 drive: USD5000
200GB ATA HDD (media + drive): USD100
You'll need to have more than 33200GB of LTO-3 backup data for it to be cheaper than 200GB HDDs. e.g. 83 LTO3 tapes (or 166 HDDs).
AFAIK you usually need only 20+ backup sets per machine...
1) You can't write to the HDD if it's on a shelf. It's removed from the computer after it's used, just like any other removable media.
2) You can't (and wouldn't want to) write protect the media you are backing up to. If the trojan screws up data whilst you are backing it's the same whether you're backing up to HDD or some other media - e.g. floppy or DVD+R or CD-R. If you can't fill a 200GB HDD, then just use smaller and cheaper HDDs. It's just like using lower capacity media.
Your argument is valid for restoring/reading data, but if you aren't able to make sure your systems don't corrupt your backups during a restore then you should consider getting hardware write blocking devices - go search on some forensic products -(e.g. Fastbloc, Nowrite) there are a few companies around. Expect to pay some money for that.
So far I haven't had problems with trojans or viruses on my computers, so such products aren't worth it for me.
It's not a dupe. It's a backup.
Seriously tho, this one is by NIST and was released AFTER that slashdot article. So it's not a dupe.
Uh, my removable media are _harddisks_. I've got my HDDs in caddies/trays (with fans) so I can slide them out easily.
:) ), then store/archive the HDDs.
Doing the figures, HDDs aren't really that expensive compared to other media especially when you factor in the performance, reliability, stability and convenience. Buy one or two 200GB HDDs, backup everything (two or more copies just to be sure
The stability of data on magnetic disks is pretty good. The only problem usually is the electronics failing or the mechanical stuff failing. But if you're not using/abusing the drives, the shelf life is pretty good.
When the time comes to migrate the data off the obsolete 200GB drives so you can still read the data, it's a lot faster and simpler to copy the data off the 200GB drives than it is to copy from other removable media. The transfer rates of optical or tape media are pretty bad. OK DVD stuff isn't that bad, but writing isn't that fast... I personally believe tapes and tape drives are a big con-job nowadays.
It'll be a lot better if hotswap SATA drives and their caddies become common and cheap. Then backing up wouldn't require a power down.
1) The virtual machines are not written by the same people writing the apps. While you might be a better programmer than those writing virtual machines, as Bugtraq shows, the vast bulk of programmers aren't very good at writing C or C++ programs.
2) Even if there is a vulnerability in the virtual machine, the attacker does not usually have direct access to the virtual machine - the attacker can only go through the app, and the vulnerability might not be able to be exploited.
3) Often a bug in the VM would be found fairly quickly once it gets used by everyone because it's at a low level - it's like having your CPU not do something correctly. Only the really obscure functionality would be untouched - but if they're mostly untouched they're less likely to be problems in practice.
4) If there's a security bug in the VM, you only have to fix one thing- the VM and your apps should still work. In contrast fixing C buffer overflows etc for dozens of different programs is not that easy. Multiply that by tons of different apps by thousands of programmers and you should see the magnitude of the problem.
5) It's an example of specialisation and delegation. Go look up how many vulnerabilities there are in the Perl/Java/Python VMs. There haven't been many announced over the years. The few people who are good at writing VMs write VMs. Those that are crap at programming like me, write in Perl etc and so don't have to deal with keeping track of pointers, allocating enough memory etc, regularly screwing things up so that "an attacker can execute arbitrary code of the attacker's choice".
It seems to be more a confidence/security thing.
IMO in general women (or perhaps just the louder/whinier ones) tend to view obstacles as problems whereas in general men view obstacles as challenges or even something fun...
I don't even think that guys playing games go "I wish I had muscles like him etc". In fact during the game they are more likely to subconsciously go "This is _me_! This is me kicking ass! Yeah!! Eat that!". And don't forget "I 0wn3d J00!".
I wonder whether most women actually see it as a problem or it's just the whiny ones.
Rakfisk doesn't seem far from the way lots of people make fish sauce (e.g. Ancient Romans, Vietnamese etc).
Doesn't even look like it gets that much of a chance to rot unlike Surstromming, or involve saturation in a pretty _toxic_ chemical (e.g. lutefisk in lye aka NaOH).
But maybe you've tasted lutefisk and rakfisk and the latter actually tastes worse?
Sounds a bit like a smartcard.
Yeah, but don't forget the most important thing - usually the old system is there because it did have some value in the first place.
First do no harm.
And so far we're still based on DNA after all these years. Maybe we're overdue for a rewrite, but so far I'm not too sure about the problem definition...
It seems the objectives of that project were quite easy to _define_. AND this sort of problem has been successfully done many times before - just not as fast :).
;) ). So it is probably hard to define what they want to their satisfaction.
Whereas in many of these big IT projects it is unclear what the customer is trying to achieve. So many parties with different goals, some probably conflicting - though they may not know it yet (and at least some would want to milk the customer as much without the customer mooing or kicking
They often will have no idea on many important decisions, so a short-term profit minded vendor (which = most) will just pick the easiest to do which complies to what the customer _said_ they wanted and not bring the issue up - educating them will just take too long. And they'll have plenty of conflicting ideas on the relatively less technically important decisions = long meeting over what colour some thingy should be (sure it's important but it doesn't need to involve so many people - it'll just waste even more manhours).
Left and right clicking isn't that hard.
Imagine if there's really only one pedal in your car and you press it once to activate the clutch, twice quickly in the same spot to accelerate, and three times quickly to brake.
Doubleclicking is the hard stuff- it requires _coordination_ in time and space. Clicking two different spots twice rapidly that are far apart does not trigger a doubleclick. For newbies the effort of clicking two spots rapidly often causes them to move the mouse greatly and thus not trigger a doubleclick. You can often visibly see the strain and stress in their faces and bodies when they try to doubleclick - often gritting their teeth or something...
Worse - nowadays most browsers a single click is what it takes to "execute/launch/activate" a link. Whereas you need a double click for other areas (word processing etc)... Now that's confusing.
If people had settled for one button for just select/deselect (without activating) and another button for activate, perhaps things might actually be easier.
I'm not sure about the drag and drop thing too... Perhaps clicking and holding down might be hard for some people.
Intelligence is not just knowing _absolute_ semantic meaning - e.g. that a cow is a cow and grass is grass. And being able to group grass with other grasses and cows with other bovine animals.
:)
It's being able to understand the statement that cow is to grass in a similar way that balleen whales are to krill.
And then now knowing something about krill from that even if you didn't know what krill was at all.
It's not just knowing the "absolute value" of the meaning - or even that these two objects are linked or close in the same area ( which seems to be the level which most current AI are at).
It's more like kind of knowing the "vector/direction" they are linked, and being able to organize other objects that are related in similar ways in a similar vector. Thus you can learn about things by analogies and metaphors AND even create new things with those.
Would pump out more BS but I have to go for dinner