It probably depends on the version of install that is used to install binaries. Some packages come with their own so there is no sure way to tell. Also this only happens (probably) when you are replacing a library with one with the same version (filename). If you're upgrading a library to a newer version this problem should not arise.
So far (on 5/7 firewalls), updating the ssl libraries caused ssh to kick out. This is very much unlike upgrading ssh, where the currently running sessions would stay active and you just kill off the 'parent' sshd process and restart sshd to upgrade.
Library upgrades may break running programs depending on the underlying OS (I noticed this on Solaris). It all depends on whether the existing library get overwritten or gets replaced (depends on cp or install used). This probably only happens if the library version number isn't changed.
A workaround would be to move the existing library aside before you do make install. (e.g. mv libssl.so.0.9.6 libssl.so.0.9.6-OLD)
Unless you use biometrics (I don't generally leave my fingers on my desk when I go to lunch), the stupid-factor will always play a part.
You may not leave your fingers, but you leave a hell of a lot of fingerprints. Fingerprints are easy to gather. Retina scans are much harder but do not adress other issues of using biometric data for authentication.
You cannot trust biometric data to be secret. You can't use it to replace passwords because you can't chage it and I'm afraid people put too much value on biometric data.
To me the distinction between "something you have", "something you know" and "something you are" has always been vague. In most practical places they can be reduced to eachother. A hand can be replaicated artificially (are->have). A onetime password system kan be described in terms of an initial vector (have->known). A password you can write down (know->have). Etc.
My Sony TV has this feature. If a press a button on the remote the picture freezes and when I press it again live images continue (images in between are lost).
Of cource it is no use because the delay is too big and you'll miss part of the stuff you're watching.
So the solution seems simple to me. Design a simple algoritm for effectively protecting your copyrighted essay or something. After that no one can legaly reverse engineer it and it can be used for other purposes. Right?
I found a caching nameserver that remembers an A record for microsoft.com is 207.46.230.218 (www is not cached). traceroute to this ip fails as well as to the authoritive nameservers for this domain.
At the moment a lot of development is going into a replacement for RADIUS. An IETF working-group is developing the DIAMETER protocol. RADIUS has a lot of weaknesses especialy in terms of security (transport of RADIUS data over a network is practicly in the clear (except passwords)). If you choose a package be sure that is ready for DIAMETER.
(www.diameter.org)
Mozilla 0.6: split from Netscape Communicator 5.0, milestone M18, same as Netscape 6, evironment variable MOZILLA5_HOME... When will there be some consistent version numbering for this? (it's worse than java or even solaris)
Otherwise: great software, use it at home all the time (M18). The rendering is cool. Love it how it rebuilds the page while loading.
Problem is I need about 600Mb to build it from source... and then it fails due to some weird configuration thingy in a Makefile/configure script/whatever.
Allso has the same problem that staroffice has: it tries to do all things you ever wanted (and a lot more you don't ever want) in a single application. Do one thing and do it well: make ghecko a seporate library! (go galeon go!)
there are advertisements at the top of every damn page of/.
Hey now that you mention it, there is a small box on the top of my page. I guss junkbuster destroyed that little piece of advertising.
I'm still waiting for a browser that finds advertising in a page and replaces that with it's own. That would be realy fun (the legal battles alone). DIE advertisers DIE!
We change our environment to suit are needs, as most animals do;
Most animals change their environment, but there are not a lot of animals that make it inhabitable, even for themselves. I beleive goats, given the chanche, will eat EVERYTHING they can find getting too stuffed one day and not having anything to eat another day. (it could be another animal)
The point is humans may be animals changing their environment, but we wouln't be the first species to go extinkt for destroying the things we need to live.
The current problems with toplevel domains are nog problems of DNS. DNS was meant to be used to give organisations a name on a global network. The problem with domain names today is that they are used to label content of a specific service on a network, namely HTTP.
HTTP uses hostnames as a basis to describe infomation and are now allmost part of the content. This sceme does not scale very well since you cannot possibly determin by a hostname what content may be on it or vice versa. The problem is not with DNS, it still works for the purpose intended. It is with URL's. They are based on hostnames and that is showing it's limitations. A scheme used in NNTP (news) is better (but by no means perfect).
New toplevel domains will not fix this problem, because the problem is not DNS, it is HTTP and that is what should be fixed.
Part of the problem with voting for a single person with so much power is that people vote for a pretty face insteadof good ideas. Most campains (of what I follow about it) are about getting the candidate to look good (look sharp, nice smile, quick response to questions, etc) rather than te content of the message.
I'm from Holland (so I probably don't have any right to bitch about this subject, but anyway) and we have a queen here. She wasn't elected so we don't have to choose a nice face. The only thing we have here are government and parlement. We can vote for ideas insteadof a face. Lately campains seem to go more about image here too, but I'm glad it hasn't turned into a tie-wearing-cometition yet. Our primeministers aren't the best looking guys around and I hope it stays that way.
Another positive aspect of having something other that a single happy face in power is that whatever the people vote, they get. If 30% votes for party A and 50% for party B (much more diverse than that over here) A gets 30% of the power (in government and parlement) and B gets 50%. So it's not like the winner gets it all and the looser has to fall.:) This ensures that somewhat balanced decisions are made where everybody's interests are taken into account.
Insurance should be about spreading the costs of a ceirtain risk. Wether this is illness, disaster, accident, whatever. We all pay an equal amount of money so we all are ensured. (that's my idea of assurance anyway)
Insurance is about making money! Insurance companies make a lot of money and try to make more by collectiong as much money as possible and try to pay out as little as possible.
Meking you pay more dependant of factors you cannot influence is wrong (moraly that is, but it seems that most companies never headrd of that). This move is only one step in the abolishment of insurance. If you are expected to have a desease and have to pay the cost of the cure in ensurance in advance, insurance becomes you saving for your deseae, in stead of everybody saving for the cure of all deseases. When this happens you can better save your money in a bank (they take a smaller cut out of your money).
I don't understand why health insurance can be a buissness and I don't think it should. Government health insurance is not realy an answer either, but people making lots of money of ill people does not seem right to me. And this stuff only makes it worse!
Carnivore is GOOD! It finaly shows people that not useing encryption for personal messages and data is tupid! I often use a computer system from the university network and I know there is a good chance someone (sniffer) is listening in. So I use ssh all the time. PGP (GPG) is not easy enough yet, so I'm sorry to say I don't use it much, then again, I don't send private things via email.
If this does not give a huge boost to crypto usage around the globe, I don't know what will.
What about the rain in the pictures containing the "guard"? Los Alamos looks like a sunny place, except when you are photographing a guard?
It probably depends on the version of install that is used to install binaries. Some packages come with their own so there is no sure way to tell. Also this only happens (probably) when you are replacing a library with one with the same version (filename). If you're upgrading a library to a newer version this problem should not arise.
So far (on 5/7 firewalls), updating the ssl libraries caused ssh to kick out. This is very much unlike upgrading ssh, where the currently running sessions would stay active and you just kill off the 'parent' sshd process and restart sshd to upgrade.
Library upgrades may break running programs depending on the underlying OS (I noticed this on Solaris). It all depends on whether the existing library get overwritten or gets replaced (depends on cp or install used). This probably only happens if the library version number isn't changed.
A workaround would be to move the existing library aside before you do make install. (e.g. mv libssl.so.0.9.6 libssl.so.0.9.6-OLD)
Unless you use biometrics (I don't generally leave my fingers on my desk when I go to lunch), the stupid-factor will always play a part.
You may not leave your fingers, but you leave a hell of a lot of fingerprints. Fingerprints are easy to gather. Retina scans are much harder but do not adress other issues of using biometric data for authentication.
You cannot trust biometric data to be secret. You can't use it to replace passwords because you can't chage it and I'm afraid people put too much value on biometric data.
To me the distinction between "something you have", "something you know" and "something you are" has always been vague. In most practical places they can be reduced to eachother. A hand can be replaicated artificially (are->have). A onetime password system kan be described in terms of an initial vector (have->known). A password you can write down (know->have). Etc.
My Sony TV has this feature. If a press a button on the remote the picture freezes and when I press it again live images continue (images in between are lost).
Of cource it is no use because the delay is too big and you'll miss part of the stuff you're watching.
So the solution seems simple to me. Design a simple algoritm for effectively protecting your copyrighted essay or something. After that no one can legaly reverse engineer it and it can be used for other purposes. Right?
I found a caching nameserver that remembers an A record for microsoft.com is 207.46.230.218 (www is not cached). traceroute to this ip fails as well as to the authoritive nameservers for this domain.
I found no such reference in rfc 2865! Only a MD5 hash over the password. Is that some extention?
At the moment a lot of development is going into a replacement for RADIUS. An IETF working-group is developing the DIAMETER protocol. RADIUS has a lot of weaknesses especialy in terms of security (transport of RADIUS data over a network is practicly in the clear (except passwords)). If you choose a package be sure that is ready for DIAMETER. (www.diameter.org)
Otherwise: great software, use it at home all the time (M18). The rendering is cool. Love it how it rebuilds the page while loading.
Problem is I need about 600Mb to build it from source... and then it fails due to some weird configuration thingy in a Makefile/configure script/whatever.
Allso has the same problem that staroffice has: it tries to do all things you ever wanted (and a lot more you don't ever want) in a single application. Do one thing and do it well: make ghecko a seporate library! (go galeon go!)
Hey now that you mention it, there is a small box on the top of my page. I guss junkbuster destroyed that little piece of advertising.
I'm still waiting for a browser that finds advertising in a page and replaces that with it's own. That would be realy fun (the legal battles alone).
DIE advertisers DIE!
Most animals change their environment, but there are not a lot of animals that make it inhabitable, even for themselves. I beleive goats, given the chanche, will eat EVERYTHING they can find getting too stuffed one day and not having anything to eat another day. (it could be another animal)
The point is humans may be animals changing their environment, but we wouln't be the first species to go extinkt for destroying the things we need to live.
The current problems with toplevel domains are nog problems of DNS. DNS was meant to be used to give organisations a name on a global network. The problem with domain names today is that they are used to label content of a specific service on a network, namely HTTP.
HTTP uses hostnames as a basis to describe infomation and are now allmost part of the content. This sceme does not scale very well since you cannot possibly determin by a hostname what content may be on it or vice versa. The problem is not with DNS, it still works for the purpose intended. It is with URL's. They are based on hostnames and that is showing it's limitations. A scheme used in NNTP (news) is better (but by no means perfect).
New toplevel domains will not fix this problem, because the problem is not DNS, it is HTTP and that is what should be fixed.
Part of the problem with voting for a single person with so much power is that people vote for a pretty face insteadof good ideas. Most campains (of what I follow about it) are about getting the candidate to look good (look sharp, nice smile, quick response to questions, etc) rather than te content of the message.
:) This ensures that somewhat balanced decisions are made where everybody's interests are taken into account.
I'm from Holland (so I probably don't have any right to bitch about this subject, but anyway) and we have a queen here. She wasn't elected so we don't have to choose a nice face. The only thing we have here are government and parlement. We can vote for ideas insteadof a face. Lately campains seem to go more about image here too, but I'm glad it hasn't turned into a tie-wearing-cometition yet. Our primeministers aren't the best looking guys around and I hope it stays that way.
Another positive aspect of having something other that a single happy face in power is that whatever the people vote, they get. If 30% votes for party A and 50% for party B (much more diverse than that over here) A gets 30% of the power (in government and parlement) and B gets 50%. So it's not like the winner gets it all and the looser has to fall.
Insurance is about making money! Insurance companies make a lot of money and try to make more by collectiong as much money as possible and try to pay out as little as possible.
Meking you pay more dependant of factors you cannot influence is wrong (moraly that is, but it seems that most companies never headrd of that). This move is only one step in the abolishment of insurance. If you are expected to have a desease and have to pay the cost of the cure in ensurance in advance, insurance becomes you saving for your deseae, in stead of everybody saving for the cure of all deseases. When this happens you can better save your money in a bank (they take a smaller cut out of your money).
I don't understand why health insurance can be a buissness and I don't think it should. Government health insurance is not realy an answer either, but people making lots of money of ill people does not seem right to me. And this stuff only makes it worse!
It has been while since I did the complexity cource, and I have not read the complete paper, but I found a small mistake (but it might be a big one).
quote:The running time of the algorithm is equal to O(n6), where n is the number of graph vertices.
n should be the number of graph vertices and graph edges together (think of Turing Machines).
Carnivore is GOOD! It finaly shows people that not useing encryption for personal messages and data is tupid! I often use a computer system from the university network and I know there is a good chance someone (sniffer) is listening in. So I use ssh all the time. PGP (GPG) is not easy enough yet, so I'm sorry to say I don't use it much, then again, I don't send private things via email.
If this does not give a huge boost to crypto usage around the globe, I don't know what will.