The evolution argument is disproven by Schneier himself; how could he be thinking about it if we hadn't already evolved to make it possible?
Schneiere isn't humanity, he's just Schniere.
Anyway, I think you're trying to take his comments too far. It seems to me Schneire's ideas are really more of a way of thinking about why people are bad at assessing risk rather than a predictive theory that can be picked apart and examined.
The correct spelling of Bruce Sniers's name is based on a constantly changing one time pad that hasn't been created yet.
Any large company is going to have special needs users who cannot make the switch, but these will often be the ones who have created the situation in the first place and the company may be better off without them.
Spoken like a true BOFH. In real life, I see both stupid solutions implemented by clueless desk jockeys because they don't know better, and weird but moderately ingenious solutions to problems directly created by a draconian IT department who has no incentive what so ever to increase user productivity. My experience is that the latter is actually more common than the former. And yet I have never, ever, seen an IT department acknowledge that end users implemented stuff that looks weird because the IT department made it impossible to do it in a reasonable way. Strange that, huh?
Quite, Ive even been part of a draconian IT dept. The other departments went off and cooked up their own multimillion pound IT fiascos. A synthesis was achieved by actually having proper requirements/costs discussions with the 'customer' depts (and moving a slightly controlling manager away from the IT dept).
Accounts software is a special case. Accounts departments often end up so conflicted by the incompatible goals of cost control and usable software that they often give up one (or both) of these requirements, whereas most other 2-department situations results in a discussion of costs vs benefits.
I think this is why many small companies fail, as they gave up cost-control or software useability for the sake of not having to do any research into solutions to their actual requirements.
And they'll say, OK, we'll try Crossover Office and see how it goes.
Unless they are so insane that they have managed to become beholden to to Office 2007 data formats already.
Any large company is going to have special needs users who cannot make the switch, but these will often be the ones who have created the situation in the first place and the company may be better off without them.
there's your problem. as I wrote in my original message.:)
Truecrypt.
I recently moved to truecrypt for my backup external disk (for a host of reasons, but primarily portability). urpmi truecrypt failed, some reading led me to urpmi realcrypt, which failed. So I went to the source. Which required kernel-source (and not the stripped version either). So thats a 200Mb download. On each PC. Followed by a compile taking 20 minutes.
On the first PC it took about 1 hour to get from urpmi truecrypt to installed, and required me to exercise extreme cleverness to understand the problems and work around them. Subsequent installs took 1/2 hour. It took a couple of minutes to go from the truecrypt download link to installed on windows. It would have taken a couple of minutes on Linux if an autopackage version was available on the truecrypt site.
I lightly investigated why urpmi realcrypt failed, and it appears it cos I dont have cooker as a repository. I dont want cooker as a repository, because I want a stable system.
Ive had similar issues, Wesnoth before it got packaged. Recently it was a 64bit version of partimage, only a patched version not available in Mandriva repositories worked, so it was the usual urpmi_failure->research why->get_source->compile->get_dependancies->compile->more_deps->compile->etc when an autopackage would require less hassle and fewer braincells.
urpmi is great when it works. When it fails, it can take alot of research, and some serious bash/sysadmin skills to get something going. I am convinced assisting developers to convert to autopackage is the way forward for the minor distros.
BTW. Seriously, drop the attitude. "Thank you for raising this issue, we appreciate your concerns, etc" may be unctuous and insincere, but never, ever, put down potential customers in a public forum.
Why do you want one? What's wrong with the package management system?
Nothing, since autopackage came along.
Its just a shame that no-one appears to use it, leading to issues like your current truecrypt woes, or the Unreal 2004 installer requiring tweaks to get UT2004 running on Mandriva.
BTW, you must try alot harder to be patient with those you suspect are trolls. Many of them are genuinely frustrated users, and a Mandriva representative coming across all 'spiky' and 'French' doesnt show your company in a good light. You really must treat them kindly, saying 'there, there, mummy kiss it better', right up until they pull off the mask to reveal they are Steve Ballmer.
If you choose the tagnames wisely, this can be true.
Unfortunately, I suffer from a colleage who favours single letter tag and attribute names:-(
Oh, and I use Natural Docs to create documentation and use (start code)... (end code) to document the XML within inches of the place where the XML is generated or parsed.
The paper is still in the preview part of natures website, so I couldn't get to the full text.
David M. Amodio has done similar trials, where responses are timed (otherwise, the subject could just take their time and get a 100% hitrate). Heres an example where latencies are measured. I have assumed a very similar method was used in both experiments. http://scan.oxfordjournals.org/cgi/content/full/1/1/26
actually, according to tfa liberals are better thinkers.
imho old persons become conservative just because of decline of cognitive functions due to old age.
No, according to TFA, liberals are faster and more reliable at differentiating between the letter M and the letter W in a timed experiment.
I think we may safely extrapolate, and say that we expect this to be true for all differentiation between objects,
I.e. liberals could tell the difference between:
a man, a woman, and cop when propositioning someone in the toilets.
a deer and the vice-president when out shooting.
a WMD and a big fat nothing.
having sexual relations, and just mislaying a cigar.
As for going so far as to say conservatives are slow and stupid. Well, there are many chains to be yanked, but saying that would would be neither fair, nor scientific.
I'm curious to know how they could screw up a simple thing like database security to the point where some windows laptop on their network can just connect and do the above.
Look at the response headers from those two websites. The site is setup by the usual MCSEs who showed management a pretty webpage without actually having anything robust underneath.
I'd imagine that the support costs alone would be enough to want to ignore Linux.....
In reality, they probably didnt do any sums at all. Some bloke in an office somewhere scratched his arse, farted, and decided that he didnt want to learn a new anything new this year. So he cracked off an email to his boss making out how difficult it would be to write Linux drivers, and how they'd get cracked and the company IP would get exposed if they did.
At no point did anybody:
Actually do any research.
Boot a Linux system and see what these things look like.
Even become aware that they could get Linux drivers written for free.
Ask their customers if Linux support was desirable.
Alas, we are still in a world where the managers of these companies have no communications with the customer, market research is just chatting in the pub, usability testing consists of ticking the 'feature implemented' box, and market knowledge is all about the knowing the customers you have, rather than the customers you dont have.
I view ogg support as a clue that the designers of the device have given some thought to what the device should be capable of, and arnt just ticking the boxes that marketing want ticked.
Other clues are:
Battery life. In particular, lying about it. A sign that the company is dishonest, and have lied about other features.
No products in the lineup that take AAA or AA batteries. A sign that designers only develop features suggested by marketing.
Proprietary connectors. Unless theres some good reason for it, like Apples dock.
Sudden changes in technical direction from the company. Archos and iRiver did this, and their product suite turned to shit overnight.
Flash ridden website with no device specifications. A sign that the company are selling an image and do not understand the product.
I go for UMS devices because then I can manage my player from anywhere without having to use special software. I can drop a podcast onto my player at work, or when I'm away.
train your stormtroopers so they can hit a man sized target at 100 ft distance
Fixed. Ground troops replaced with an airforce that bombs everything flat.
don't have your war droids depend on a centralized node that, when destroyed, would disable the whole army
Fixed Droids now have autonomous rape, pillage and Abu Gharib modes.
make sure there are no vents leading directly to your death star's reactor, no matter how hard or unlikely to hit they are
ToDo In fact, still fixing the requirements so that our preferred bidders win the contracts.
fun as it may be, and sure as you may be that he's a complete bastard, don't send a father to torture his daughter and duel his son. They might end up working together against you. Also, if you've decided to replace him with his son, don't tell it to his face.
Fixed In fact, Daddy Bush was completely happy to see his son take over.
don't make yourself hated by whole populations in the first place. Destroying whole planets just to show you can, is actually pretty bad PR. It's bad for your tax income too. Noone will rise in rebellion or send suicide bombers against you for just treating them right and creating employment.
Not a problem We need rebelious planets, otherwise the Military Industrial complex on the home planets wont have any outlets for its products.
make sure the doors, especially prison doors or doors to critical command rooms, can't be opened by shooting the control panel. And generally, security means everything should fail in the way that is the least of a security problem. Losing electricity should cause the door bolts to lock the door (e.g., they're on springs that push them to the locked state, and you need current to pull them open), not unlock it.
Fixed We now keep the whole population under surveillance. Criminals are forced to work in government, obviating the need for prisons.
for that matter, and according to the same principle, a damaged reactor should tend to shut down, not blow up. There's a reason 20'th century nuclear reactors need current to keep the moderator rods out, and get to shut down if they lose that current
ToDo To busy securing the oil supply to worry about that stuff.
control consoles don't have much of a reason to explode when the ship takes a hit in some point half a mile away. You may need that console again, and trained specialist officers that operate them are expensive to replace too
Not a problem Are you kidding? The monkeys we pay to watch those screens are even more expendable than the ground troopers.
invest in some shielding technology, or at least armour. The Mitsubishi A6M Zero fared poorer than you'd think with only speed and maneuverability as its only defenses, and got shot by airplanes which could take a whole clip and keep flying. The TIE fighter is just repeating an existing mistake. Don't do it.
Not a problem The best defense is a good offense. All ships now equiped with really big bombs for use against villagers, so we are really, really offensive.
And generally, read the evil overlord's list already.
Local cycle path where they installed gates across it. Now removed, thankfully.
Bracknell (UK) cyclepath terminated 10 meters before a very busy roundabout. A couple of fatalities every year on that one.
Any non-standard component or interfaces, particularly lithium batteries and the almost infinite variety of power jacks on mobile phones.
The human body. Why is it so weak and fragile, and whats with all this hair? I swear, the designer was either incompetent or just making stuff up as he went along.
Not really. Its not brilliant, but its no worse that other kludged together monstrosities like perl or XSL.
Add in the fact that there is no direct equivalent to control the running of programs in *nix and windows, and JCL (plus JES and the spool) seems far preferable to kludging cron and bash scripts together to get regular tasks going with the output saved on a spool for review or deletion.
While we are still channelling the spirit of DNA, why not build robots that want to be blown up, and are perfectly capable of saying so for themselves?
Slashdot Mirror
The evolution argument is disproven by Schneier himself; how could he be thinking about it if we hadn't already evolved to make it possible?
Schneiere isn't humanity, he's just Schniere.
Anyway, I think you're trying to take his comments too far. It seems to me Schneire's ideas are really more of a way of thinking about why people are bad at assessing risk rather than a predictive theory that can be picked apart and examined.
The correct spelling of Bruce Sniers's name is based on a constantly changing one time pad that hasn't been created yet.
Spoken like a true BOFH. In real life, I see both stupid solutions implemented by clueless desk jockeys because they don't know better, and weird but moderately ingenious solutions to problems directly created by a draconian IT department who has no incentive what so ever to increase user productivity. My experience is that the latter is actually more common than the former. And yet I have never, ever, seen an IT department acknowledge that end users implemented stuff that looks weird because the IT department made it impossible to do it in a reasonable way. Strange that, huh?
Quite, Ive even been part of a draconian IT dept. The other departments went off and cooked up their own multimillion pound IT fiascos. A synthesis was achieved by actually having proper requirements/costs discussions with the 'customer' depts (and moving a slightly controlling manager away from the IT dept).
Accounts software is a special case. Accounts departments often end up so conflicted by the incompatible goals of cost control and usable software that they often give up one (or both) of these requirements, whereas most other 2-department situations results in a discussion of costs vs benefits.
I think this is why many small companies fail, as they gave up cost-control or software useability for the sake of not having to do any research into solutions to their actual requirements.
And they'll say, OK, we'll try Crossover Office and see how it goes.
Unless they are so insane that they have managed to become beholden to to Office 2007 data formats already.
Any large company is going to have special needs users who cannot make the switch, but these will often be the ones who have created the situation in the first place and the company may be better off without them.
"Unreal 2004 installer"
there's your problem. as I wrote in my original message.
Truecrypt.
I recently moved to truecrypt for my backup external disk (for a host of reasons, but primarily portability). urpmi truecrypt failed, some reading led me to urpmi realcrypt, which failed. So I went to the source. Which required kernel-source (and not the stripped version either). So thats a 200Mb download. On each PC. Followed by a compile taking 20 minutes.
On the first PC it took about 1 hour to get from urpmi truecrypt to installed, and required me to exercise extreme cleverness to understand the problems and work around them. Subsequent installs took 1/2 hour. It took a couple of minutes to go from the truecrypt download link to installed on windows. It would have taken a couple of minutes on Linux if an autopackage version was available on the truecrypt site.
I lightly investigated why urpmi realcrypt failed, and it appears it cos I dont have cooker as a repository. I dont want cooker as a repository, because I want a stable system.
Ive had similar issues, Wesnoth before it got packaged. Recently it was a 64bit version of partimage, only a patched version not available in Mandriva repositories worked, so it was the usual urpmi_failure->research why->get_source->compile->get_dependancies->compile->more_deps->compile->etc when an autopackage would require less hassle and fewer braincells.
urpmi is great when it works. When it fails, it can take alot of research, and some serious bash/sysadmin skills to get something going. I am convinced assisting developers to convert to autopackage is the way forward for the minor distros.
BTW. Seriously, drop the attitude. "Thank you for raising this issue, we appreciate your concerns, etc" may be unctuous and insincere, but never, ever, put down potential customers in a public forum.
Nothing, since autopackage came along.
Its just a shame that no-one appears to use it, leading to issues like your current truecrypt woes, or the Unreal 2004 installer requiring tweaks to get UT2004 running on Mandriva.
BTW, you must try alot harder to be patient with those you suspect are trolls. Many of them are genuinely frustrated users, and a Mandriva representative coming across all 'spiky' and 'French' doesnt show your company in a good light. You really must treat them kindly, saying 'there, there, mummy kiss it better', right up until they pull off the mask to reveal they are Steve Ballmer.
If you choose the tagnames wisely, this can be true.
:-(
... (end code) to document the XML within inches of the place where the XML is generated or parsed.
Unfortunately, I suffer from a colleage who favours single letter tag and attribute names
Oh, and I use Natural Docs to create documentation and use (start code)
More likely, start by playing the "Guess the Webserver" game.
Compare with the likes of Bank of India, Monster.com, USAjobs.gov, myspace.com and other recent security incidents.
Do you see a pattern emerging?
Well spotted.
The paper is still in the preview part of natures website, so I couldn't get to the full text.
David M. Amodio has done similar trials, where responses are timed (otherwise, the subject could just take their time and get a 100% hitrate). Heres an example where latencies are measured. I have assumed a very similar method was used in both experiments.
http://scan.oxfordjournals.org/cgi/content/full/1/1/26
imho old persons become conservative just because of decline of cognitive functions due to old age.
No, according to TFA, liberals are faster and more reliable at differentiating between the letter M and the letter W in a timed experiment.
I think we may safely extrapolate, and say that we expect this to be true for all differentiation between objects,
I.e. liberals could tell the difference between:
As for going so far as to say conservatives are slow and stupid. Well, there are many chains to be yanked, but saying that would would be neither fair, nor scientific.
Look at the response headers from those two websites. The site is setup by the usual MCSEs who showed management a pretty webpage without actually having anything robust underneath.
So who do you imagine thinks SCO has any value?
In reality, they probably didnt do any sums at all. Some bloke in an office somewhere scratched his arse, farted, and decided that he didnt want to learn a new anything new this year. So he cracked off an email to his boss making out how difficult it would be to write Linux drivers, and how they'd get cracked and the company IP would get exposed if they did.
At no point did anybody:
Alas, we are still in a world where the managers of these companies have no communications with the customer, market research is just chatting in the pub, usability testing consists of ticking the 'feature implemented' box, and market knowledge is all about the knowing the customers you have, rather than the customers you dont have.
At least it doesn't interfere with bathroom breaks.
Well, it hasn't since I had the catheter put in.
Ditto that.
I view ogg support as a clue that the designers of the device have given some thought to what the device should be capable of, and arnt just ticking the boxes that marketing want ticked.
Other clues are:
I go for UMS devices because then I can manage my player from anywhere without having to use special software. I can drop a podcast onto my player at work, or when I'm away.
Fixed. Ground troops replaced with an airforce that bombs everything flat.
Fixed Droids now have autonomous rape, pillage and Abu Gharib modes.
ToDo In fact, still fixing the requirements so that our preferred bidders win the contracts.
Fixed In fact, Daddy Bush was completely happy to see his son take over.
Not a problem We need rebelious planets, otherwise the Military Industrial complex on the home planets wont have any outlets for its products.
Fixed We now keep the whole population under surveillance. Criminals are forced to work in government, obviating the need for prisons.
ToDo To busy securing the oil supply to worry about that stuff.
Not a problem Are you kidding? The monkeys we pay to watch those screens are even more expendable than the ground troopers.
Not a problem The best defense is a good offense. All ships now equiped with really big bombs for use against villagers, so we are really, really offensive.
Read it? We wrote it!
Karl, Dick and Donald.
bad designs
But my personal ones are:
Me too, and I am nearly five.
Not really. Its not brilliant, but its no worse that other kludged together monstrosities like perl or XSL.
Add in the fact that there is no direct equivalent to control the running of programs in *nix and windows, and JCL (plus JES and the spool) seems far preferable to kludging cron and bash scripts together to get regular tasks going with the output saved on a spool for review or deletion.
Not really.
Other zOS skills are more useful, particularly the Sysprogging skills like JCL, console commands, rexx, TSO and ISPF.
The Linux skills to have in this area are the System Admin ones like bash scripting, setting up Apache and other server stuff, using ssh, etc.
Why, only 2 hours ago I was heard to say:
"That's like drawing a picture of a car, and saying 'This is how it will feel to drive'"
Not sure it worked, though.
Heres an earlier I.e.
While we are still channelling the spirit of DNA, why not build robots that want to be blown up, and are perfectly capable of saying so for themselves?
I think you'll find that the oil is English. It was us, after all, who stole it from Norway.
Well, at least it limits the number of items they can steal from you like that.
You mean we need to forest over France as well?
Oh well, if we must.