So firefox has the same addons (literally) like chrome, the same limited possiblity to tune the UI, a loss of many good firefox-only addons and the thunderbird developers need to maintain xul themself?
Looks like firefox will get a lot less important and thunderbird may die.
That's not because of pgp, enigmail,... But because of the lack of interest. Build a fancy app and people use it. Bundle encryption and you get encrypted messages. But they do not care about the encryption.
That's not fully true. pgp is a system, which can extend any message system with security. This does not depend on the protocols in any way. SMTP, POP3, IMAP, MIME... do not care about the message body itself. You can pgp encrypt your facebook message or paste your encrypted message in a pastebin, you can even print it. The medium does not care, as long as you get the ciphertext back into the pgp program.
Let's face it, users stop using e-mail anyway. Many apps which required e-mail for signup now work with a phone number alone. So let's stop forcing e-mail for every bit of communication. Use XMPP. The user likes the chat interface anyway, encryption with OMEMO (which has forward secrecy, which isn't possible with e-mail) is secure and apps like Conversations work like a charm hiding all the details with a Trust-on-first-use model, which is enough for 99% of the users.
The other thing: USE XMPP.
There are a lot of shiny new easy crypto messengers. And many of them work real good, some seem to be pretty secure. But almost all of them use a central server or when they let you configure the server, all your friends need to be at the same server. To be independend we need a federated system. We have such a system. The System is called XMPP, has good specifications and a lot of opensource server and client softwares and a infrastructure of many free servers around the world, which are already available. Stop creating new crypto messenger protocols. USE XMPP!
And... you mixed up RSA and AES. RSA just relies on prime factorization being NP-hard and P!=NP. AES is the encryption with some unknown constants (which are still not proven as dangerous btw.)
So maybe the user is right about not knowing the details of this... when even you, who are a bit more advanced, get it wrong.
Because using a CA system for E-Mail encryption just gives a loooong list of CAs the chance to fake certificates so an attacker can read the mail. you're no longer defending against one attacker or needing to trust one provider, but you need to trust a list of providers. A list, which is not choosen by you, but by the recipient which is fooled by the attacker. And the usual recipient (probably including you) did not change the list of accepted CAs. So actually the mail client programmer decides who can fake certificates for your mailadress.
App was a short for application, but not often used. And MacOS always called applications applications, which is why apple had no big problem to use App for mobile applications without confusing people.
violence at least HAS an useful definition and you can tell people that speech is not violence and they should use the correct words. They might not accept it, but then you know, that they are weird. But most of the new terms are free to be used as appropriate in the moment and to be used as something else later.
So, the misunderstanding is, that the context to the pure information does not come from a well-known name, but from the content of the information and your own research on the topic, if you're interested.
You can believe something, just because the president said it, but maybe you better google yourself. Some *good* ressources on the topic.
AC is the purest form of information. You judge the content instead of the name. You need to verify each claim yourself and will know much more, than when you blindly trust a known name.
Leave Toxicity to the chemistry and use real words to describe what you want to accuse people of. Terms like toxic behaviour and hate speech are cool, because there is no clear definition, which means you can redefine them each time you use them. If somebody refutes a claim, you tell them, that the word was used in another context than what he refuted.
add a root node to a queue get an element from the queue when you visited the element, stop else process the element and add all linked elements to the queue mark all graph links (child directorys and symlinks) as visisted.
complexity of a directory traversal O(n log n) where n is the tree depth complexity of "is marked": We use a tree. insertion is O(log m), lookup is O(log m) for the number of graph edges. The total complexity is then O(n log n + n log m) (which is another way to express n*log(max(n,m))), which is for normal directory structures (number of links proportional to number of directories) O(n log n)
did i forget something in this algorithm? I did not test it, yet;-).
"A series of hacks" Nope, these aren't hacks, they are simple settings.
And there is a much easier way. Download the f*cking ESR-Release, unzip it and use it instead of the normal release. This even works coming from a newer version (nobody will guarantee you a smooth downgrade, but normally it works without major problems).
And on the other hand, the ESR will only delay the change. ESR is similiar to stopping upgrading firefox, but guarantees you security patches until the next ESR superseedes the current one.
A tree has nodes and leafs. A leaf is where the algorithm terminates. A node has children, which can be followed. A search visits each node once, where links between nodes may alter when a node is visited first. When you visit a node, you mark it as visited and add the children (and links) to a queue or stack (depths first / breath first search). I do not see any need to visit a node twice, as you get no new information on the second visit.
That's a termination condition: If you ever visit a node, which was visited before, stop.
You do not follow the symlink 1000 times and then abort. You follow it one time and the next time you see a mark "followed it" and stop. Without error as successful termination of a directory traversal.
You think such a link would be an invalid condition, but it is actually not. And it isn't even a special case to the algorithm, which has the invariant "always take a node, which wasn't visisted yet until there are no such nodes".
Nope. You need to get the computer scientists point of view: Have a termination condition. And this is usually NOT a counter, but a condition, which needs to be fulfilled. i.e. you visited every node of a tree. This takes 1 mio steps? Yep, then it does. Your counter is arbitrary and/or hard to calculate and some measure for ppl who don't trust their code to have the right condition. Which is made easier by recursion, because functions tend to be more understable, as they are similiar to their math equivalents.
The only thing to keep in mind: Don't recalculate values. Things like adding fib(n) and fib(n+1) shouldn't Take 2n+1 steps.
Slashdot Mirror
It does. Look up, what's the unique wrkey, firefox uses to get anti-phishing-data from google.
So firefox has the same addons (literally) like chrome, the same limited possiblity to tune the UI, a loss of many good firefox-only addons and the thunderbird developers need to maintain xul themself?
Looks like firefox will get a lot less important and thunderbird may die.
Blackberry OS is mostly dead, tizen and so on never really started and other custom OS run on phones which are not called smartphones.
When you decide to get an insurance, you can decide to get a backup. the problem firms are those, who don't think about the problem at all.
But we HAVE a better system, and that's PGP. It's not like CA vs. unencrypted, but CA vs. WoT/own verification vs. unencrypted.
That's not because of pgp, enigmail, ...
But because of the lack of interest.
Build a fancy app and people use it. Bundle encryption and you get encrypted messages. But they do not care about the encryption.
That's not fully true. pgp is a system, which can extend any message system with security. This does not depend on the protocols in any way. SMTP, POP3, IMAP, MIME ... do not care about the message body itself. You can pgp encrypt your facebook message or paste your encrypted message in a pastebin, you can even print it. The medium does not care, as long as you get the ciphertext back into the pgp program.
Maybe it's time to stop teaching e-mail to users.
Let's face it, users stop using e-mail anyway. Many apps which required e-mail for signup now work with a phone number alone.
So let's stop forcing e-mail for every bit of communication.
Use XMPP. The user likes the chat interface anyway, encryption with OMEMO (which has forward secrecy, which isn't possible with e-mail) is secure and apps like Conversations work like a charm hiding all the details with a Trust-on-first-use model, which is enough for 99% of the users.
The other thing: USE XMPP.
There are a lot of shiny new easy crypto messengers. And many of them work real good, some seem to be pretty secure. But almost all of them use a central server or when they let you configure the server, all your friends need to be at the same server.
To be independend we need a federated system. We have such a system. The System is called XMPP, has good specifications and a lot of opensource server and client softwares and a infrastructure of many free servers around the world, which are already available.
Stop creating new crypto messenger protocols. USE XMPP!
And ... you mixed up RSA and AES.
RSA just relies on prime factorization being NP-hard and P!=NP.
AES is the encryption with some unknown constants (which are still not proven as dangerous btw.)
So maybe the user is right about not knowing the details of this ... when even you, who are a bit more advanced, get it wrong.
Because using a CA system for E-Mail encryption just gives a loooong list of CAs the chance to fake certificates so an attacker can read the mail. you're no longer defending against one attacker or needing to trust one provider, but you need to trust a list of providers. A list, which is not choosen by you, but by the recipient which is fooled by the attacker. And the usual recipient (probably including you) did not change the list of accepted CAs. So actually the mail client programmer decides who can fake certificates for your mailadress.
App was a short for application, but not often used. And MacOS always called applications applications, which is why apple had no big problem to use App for mobile applications without confusing people.
violence at least HAS an useful definition and you can tell people that speech is not violence and they should use the correct words. They might not accept it, but then you know, that they are weird. But most of the new terms are free to be used as appropriate in the moment and to be used as something else later.
So, the misunderstanding is, that the context to the pure information does not come from a well-known name, but from the content of the information and your own research on the topic, if you're interested.
You can believe something, just because the president said it, but maybe you better google yourself. Some *good* ressources on the topic.
https://chrome.google.com/webs...
please stop the bullshit.
AC is the purest form of information. You judge the content instead of the name. You need to verify each claim yourself and will know much more, than when you blindly trust a known name.
Trolling is a art. You have lost.
Leave Toxicity to the chemistry and use real words to describe what you want to accuse people of. Terms like toxic behaviour and hate speech are cool, because there is no clear definition, which means you can redefine them each time you use them. If somebody refutes a claim, you tell them, that the word was used in another context than what he refuted.
Hmm, i think of the following algorithm:
add a root node to a queue
get an element from the queue
when you visited the element, stop
else process the element and add all linked elements to the queue
mark all graph links (child directorys and symlinks) as visisted.
complexity of a directory traversal O(n log n) where n is the tree depth
complexity of "is marked": We use a tree. insertion is O(log m), lookup is O(log m) for the number of graph edges.
The total complexity is then O(n log n + n log m) (which is another way to express n*log(max(n,m))), which is for normal directory structures (number of links proportional to number of directories) O(n log n)
did i forget something in this algorithm? I did not test it, yet ;-).
WD drives hold my hosts file and the access is very very fast, much faster than adblock!
"A series of hacks"
Nope, these aren't hacks, they are simple settings.
And there is a much easier way. Download the f*cking ESR-Release, unzip it and use it instead of the normal release. This even works coming from a newer version (nobody will guarantee you a smooth downgrade, but normally it works without major problems).
And on the other hand, the ESR will only delay the change. ESR is similiar to stopping upgrading firefox, but guarantees you security patches until the next ESR superseedes the current one.
Only if you don't know your data structures.
A tree has nodes and leafs. A leaf is where the algorithm terminates. A node has children, which can be followed.
A search visits each node once, where links between nodes may alter when a node is visited first. When you visit a node, you mark it as visited and add the children (and links) to a queue or stack (depths first / breath first search).
I do not see any need to visit a node twice, as you get no new information on the second visit.
That's a termination condition: If you ever visit a node, which was visited before, stop.
You do not follow the symlink 1000 times and then abort. You follow it one time and the next time you see a mark "followed it" and stop. Without error as successful termination of a directory traversal.
You think such a link would be an invalid condition, but it is actually not. And it isn't even a special case to the algorithm, which has the invariant "always take a node, which wasn't visisted yet until there are no such nodes".
Tail OPTIMIZATION means, that your compiler moves the recursive call to the end, not you. You may do, because it often adds clarity, though.
Nope. You need to get the computer scientists point of view: Have a termination condition. And this is usually NOT a counter, but a condition, which needs to be fulfilled. i.e. you visited every node of a tree. This takes 1 mio steps? Yep, then it does. Your counter is arbitrary and/or hard to calculate and some measure for ppl who don't trust their code to have the right condition. Which is made easier by recursion, because functions tend to be more understable, as they are similiar to their math equivalents.
The only thing to keep in mind: Don't recalculate values. Things like adding fib(n) and fib(n+1) shouldn't Take 2n+1 steps.