Right. And if you consider entire world, then only 0.0000x% are unhappy. Is that what you're implying? You simply cannot consider non-voters. Anyway you slice it, a lot of people are unhappy. And if you look at this you will see even more reason to be upset. It's not Iraq, or economy that dominated this election. It's moral values. For some reason, the pesky problem of abortion and gay mariage (or civil union or whatever you want to call it) is way more important then everything else including lives of thousands of people.
Granted, SP2 does a lot. But it's hardly a "patch" based on it's size. So, to say that "I dont think Linux could come close to releasing a patch of this magnitude" is the same as saying "I don't think Linux could release 2.6 kernel".
The lockpicking talk is scheduled for Saturday morning
The lockpicking talk is given by Matt Blaze.
I was lucky to attend his talk at my university.
The paper he presented is avalable here.
Not only was his talk educational, it was very entertaining. He actually brough some "tools of trade", and had an interactive demonstration of master key picking. It was fun.
Would you rather have a user process wait for all pages to be brought into memory, or would you rather have your kernel wait until a page is brought into the cache from the disk?
Kernel preemtiveness can't solve everything. If your system call has to read a page from disk, it might have to wait until the page is brought in from disk. During this time the kernel could be locked.
Your application, on the other hand, will be put onto wait queue until data is available, allowing other applications to run.
Of course, it's a known fact that DB people generaly do not like kernel people:) -- primarily because of page caching. All of these nice read aheads are abslutely useless for DB system. But, do not despair! 2.6 kernel has a much improved read ahead algorithm.
It might be a joke to you, but my family found out about this because my dads factory got a big order for geiger counters to be made in great hurry. He did not know exactly what had happened, but he know something was not right.
The first official acknowledgement was, if I recall correctly, on May 3rd -- full week after the disaster.
Be thankfull the commend did not look like this:
' Not sure why this works for my test data.
' Probably should come back and re-write this
' if we have time before the product ships.
This might not work as expected. Since SA assignes -8.0 score for Habeas rules, and default configuration for autolearning ham is -5.0, SA would have learnt that Habeas headers are associated with ham messages. As a result, Bayesian scores for _any_ message with these headers will score very low (50-60% probability even for the spammiest spam).
I had to manually train SA by feeding it habeas headers and training as spam, until habeas headers were associated with enough spam AND ham messages.
Totall BS.
Two issues here:
1. Technical... Security researchers seem to thing that the only way to ensure security is not by obscurity, but by keeping algorithms and implementations open. As the article points out, releasing the source for the system, enabled a scientist from AU to find a serious flow. While it's unrealistic to expect everybody to go over every line of code, it is however very likely that somebody will find a problem.
2. Trust. People trust these ancient voting machines because they know how they work. They also think that there are enough checks and balances to ensure that the system functions and is not rigged. The old voting machines do not work very well in extremelly close elections where results can be contested due to inherent error level of manual counting. Why should this trust be any different for electronic voting system? Why should the new electronic machines be a complete black box, whose results cannot be understood or challanged?
In short,"open source doesn't make it right" is a good statement in a lot of situations, in this particular situation, open source is the ONLY way to make it right.
Are you sure you'll have enough storage?.NET libray you install with dev studio is around 3 GB. The truth is: if Microsoft wants to prevent other search engines from indexing them, they can and will do it.
That is why when you decide to add patches to qmail, you only add patches that had be verified , are known to be good, and had been in use for couple of years. There are plenty terrible patches available -- if you apply those, it just means that you don't know what you're doing. However, some of the best patches where written by very smart people (who published qmail books, for example). Qmail is secure because it was _designed_ from the ground up to be secure. Anybody who knows enough C program and about qmail internals can add patches to qmail without comprimising it's security because qmail _makes it easy_ to be secure.
Slashdot Mirror
Right. And if you consider entire world, then only 0.0000x% are unhappy. Is that what you're implying?
You simply cannot consider non-voters. Anyway you slice it, a lot of people are unhappy. And if you look at this you will see even more reason to be upset.
It's not Iraq, or economy that dominated this election. It's moral values. For some reason, the pesky problem of abortion and gay mariage (or civil union or whatever you want to call it) is way more important then everything else including lives of thousands of people.
Granted, SP2 does a lot. But it's hardly a "patch" based on it's size. So, to say that "I dont think Linux could come close to releasing a patch of this magnitude" is the same as saying "I don't think Linux could release 2.6 kernel".
It's not a question of how fast Mozilla developers will patch the system. It's a question on how fast will the end user upgrade the software.
My bet is that since there is windows update service for Mozilla, a _lot_ of vulnerable mozillas will still be out there.
Mozilla probably needs a "nagging" service to remind people to upgrade if newer version of mozilla is available.
The lockpicking talk is scheduled for Saturday morning
The lockpicking talk is given by Matt Blaze. I was lucky to attend his talk at my university. The paper he presented is avalable here.
Not only was his talk educational, it was very entertaining. He actually brough some "tools of trade", and had an interactive demonstration of master key picking. It was fun.
Watch out! RIAA might be coming after you.
Of course they would have to go over all material provided by FSF to claim (and prove) that FSF missed something.
Not that SCO needs proof anyway.
Would you rather have a user process wait for all pages to be brought into memory, or would you rather
:) -- primarily because of page caching. All of these nice read aheads are abslutely useless for DB system. But, do not despair! 2.6 kernel has a much improved read ahead algorithm.
have your kernel wait until a page is brought into the cache from the disk?
Kernel preemtiveness can't solve everything. If your system call has to read a page from disk, it might have to wait until the page is brought in from disk. During this time the kernel could be locked.
Your application, on the other hand, will be put onto wait queue until data is available, allowing other applications to run.
Of course, it's a known fact that DB people generaly do not like kernel people
This does not mean that perl is an ugly nasty language. It might mean that the person might not b e aware of perls capabilities.
Here is a one liner in perl:
perl -ne 'print unless(++$seen{$_} > 1)' quote*.txt
It might be a joke to you, but my family found out about this because my dads factory got a big order for geiger counters to be made in great hurry. He did not know exactly what had happened, but he know something was not right.
The first official acknowledgement was, if I recall correctly, on May 3rd -- full week after the disaster.
Be thankfull the commend did not look like this:
' Not sure why this works for my test data.
' Probably should come back and re-write this
' if we have time before the product ships.
From now on, KDE reviews should be posted by Gnome users and vise versa. This is likelly to generate more balanced and useful reviews
This actually reminded me of Nikita Khrushchev famous shoe poinding insident at UN:
Darl, removing his show:
THERE IS NO FREE LUNCH!!!
This might not work as expected. Since SA assignes -8.0 score for Habeas rules, and default configuration for autolearning ham is -5.0, SA would have learnt that Habeas headers are associated with ham messages. As a result, Bayesian scores for _any_ message with these headers will score very low (50-60% probability even for the spammiest spam).
I had to manually train SA by feeding it habeas headers and training as spam, until habeas headers were associated with enough spam AND ham messages.
This balanced things out.
You should take a loot at
http://cr.yp.to/im2000.html
You forgot:
Ignorance is strength
You really do have to be ignorant to buy "MS is about choice.
Actually, television watches you in Orwells 1984.
Totall BS.
Two issues here:
1. Technical... Security researchers seem to thing that the only way to ensure security is not by obscurity, but by keeping algorithms and implementations open. As the article points out, releasing the source for the system, enabled a scientist from AU to find a serious flow. While it's unrealistic to expect everybody to go over every line of code, it is however very likely that somebody will find a problem.
2. Trust. People trust these ancient voting machines because they know how they work. They also think that there are enough checks and balances to ensure that the system functions and is not rigged. The old voting machines do not work very well in extremelly close elections where results can be contested due to inherent error level of manual counting. Why should this trust be any different for electronic voting system? Why should the new electronic machines be a complete black box, whose results cannot be understood or challanged?
In short,"open source doesn't make it right" is a good statement in a lot of situations, in this particular situation, open source is the ONLY way to make it right.
One of the slashdotters sigs would be a nice answer to this:
vi VS emacs wars are pointless...
vi is superior
Are you sure you'll have enough storage? .NET libray you install with dev studio is around 3 GB.
The truth is: if Microsoft wants to prevent other search engines from indexing them, they can and will do it.
That is why when you decide to add patches to qmail, you only add patches that had be verified , are known to be good, and had been in use for couple of years.
There are plenty terrible patches available -- if you apply those, it just means that you don't know what you're doing. However, some of the best patches where written by very smart people (who published qmail books, for example).
Qmail is secure because it was _designed_ from the ground up to be secure. Anybody who knows enough C program and about qmail internals can add patches to qmail without comprimising it's security because qmail _makes it easy_ to be secure.