They should wait a known period after contacting distribution people. That way red hat etc. have a known time to test the fix. They don't want to release an untested product, even if that means leaving their systems vulnerable.
No, they should patch their binary releases, give redhat, suse, and everyone on their accredited list (show you're a vendor and care vaguely about security and you get on) the source patches, but hold off making the source patches public for a week. That way redhat and everyone know where they stand, if they can't get it out in a week it's their fault, but if they can there's no need for their users to be exposed to extra vulnerability.
Once the source code for the patch is out there, any script kiddie will have an exploit. This leaves distribution users who don't want to install an upgrade from source (which will lose their support) high and dry. If you save releasing the exploit publicly until every vendor has a patch ready, then people can patch before the vulnerability becomes widely known. Real crackers can reverse engineer the vendor patches, or have someone on the inside - but then real crackers probably found the vulnerability before any of the "good guys" knew about it.
Some people do buy for the firmware or drivers though. I know people who buy nvidia graphics cards solely because the drivers are better. I've heard wireless cards recommended simply because you have more control over them in various ways, all of which are firmware-dependent. As far as the manufacturers are concerned, if it's a possible advantage they want to hold on to it. And can you blame them?
I can't remember the details, it was a long time ago, but I do remember following all the instructions I got from the actual program, and it not being able to find the partition.
Not really. Bittorrent was designed from scratch to make it hard to use for illegal purposes - trackers are there so there's an obvious target to sue if what's being downloaded is infringing.
Unnecessary choice is bad, because it's making everything nonstandard. Choice is good, yes - but everyone having their own language is not good, a standard language is better. Perl gives you many ways to do things. This makes it easier to write, but far harder to read, because you need to know all the different ways. In the long run this tends to bite you.
I find python more fun, because it looks cleaner. It's got more structure. I'm a mathematician and my aesthetics may not be "normal", but to me something near-geometric is preferrable to the "tangle" I seem to get when programming perl. As far as perl dying is concerned, how is that community going? Is it shrinking? Is it aging in almost real time? If either of those are true, perl is dying.
Why? They're just doing what they've been taught. They've been taught that anyone can use a computer (that's what all the MS advertising says). They've been taught that anyone can use the internet (that's what all the AOL advertising says). They've been taught that if they don't click yes at dialog boxes, things don't work (A very similar security warning is seen two or three times when logging in to hotmail. So users learn they're not serious) How can you blame them for simply doing what they've been taught?
But why doesn't the installer tell you that? There's no web browser on the install cd, is one expected to do each stage and then reboot to the old OS to see what to do next?
Did you pay or already have an account (meaning you found it before)? Because if not I don't think there's any way you can find something good in 13 seconds.
Of course I tried it. Boot, ooh, kernel messages, nice, hmm, hda is wd0, that's...interesting... right, NO I definitely don't want to use the whole disk. Right, cfdisk - not found, oh well, I can use fdisk if I need to. Read the help, add partition, primary, change type to BSD, set active, write, quit. Run install. Can't find the partition. What? I just made it. fdisk again to check, yep, the partition is there, why can't install see it. Run install again. Still not found. Bah. I give up.
It won't be. It won't even be popular non-mainstream. A program that requires reading the docs to use is a flawed program. If you're doing something really complex, sure, one should expect to read the documentation, but installing alongside an existing partition is not complex.
I know the basics, but prior to this thread I didn't even know about the existence of the disklabel program. The installer gives you no clues at all. A "you probably want to do this" message would make the world of difference without much effort.
I don't mind the lack of graphics, but the installer doesn't seem to give any pointers about how to install. It tells me I need to make a primary partition for bsd and mark it as active, so I do that. Then I try to run install and it can't find a partition. I appreciate that reading the documentation is important, but it should be doable at least for basic common configurations (not just the entire disk) without needing the doc.
Doesn't that mean you have to trust not only the makers of the protocol, but also absolutely anyone, since anyone can run a server and give themselves an ID on their own server? Making the ID worthless?
The reason they released it was it was all they had. They didn't have time to test and integrate a new version. It's the firefox people's fault for having the vulnerabilities, I think they're just trying to distract from their own failings.
Slashdot Mirror
They should wait a known period after contacting distribution people. That way red hat etc. have a known time to test the fix. They don't want to release an untested product, even if that means leaving their systems vulnerable.
No, they should patch their binary releases, give redhat, suse, and everyone on their accredited list (show you're a vendor and care vaguely about security and you get on) the source patches, but hold off making the source patches public for a week. That way redhat and everyone know where they stand, if they can't get it out in a week it's their fault, but if they can there's no need for their users to be exposed to extra vulnerability.
Once the source code for the patch is out there, any script kiddie will have an exploit. This leaves distribution users who don't want to install an upgrade from source (which will lose their support) high and dry. If you save releasing the exploit publicly until every vendor has a patch ready, then people can patch before the vulnerability becomes widely known. Real crackers can reverse engineer the vendor patches, or have someone on the inside - but then real crackers probably found the vulnerability before any of the "good guys" knew about it.
Some people do buy for the firmware or drivers though. I know people who buy nvidia graphics cards solely because the drivers are better. I've heard wireless cards recommended simply because you have more control over them in various ways, all of which are firmware-dependent. As far as the manufacturers are concerned, if it's a possible advantage they want to hold on to it. And can you blame them?
Crapflooding doesn't make speech impossible. The right to free speech is far more important than the right to be listened to.
I can't remember the details, it was a long time ago, but I do remember following all the instructions I got from the actual program, and it not being able to find the partition.
You have to open the .definitions_of_irony_and_sarcasm file
Not really. Bittorrent was designed from scratch to make it hard to use for illegal purposes - trackers are there so there's an obvious target to sue if what's being downloaded is infringing.
Unnecessary choice is bad, because it's making everything nonstandard. Choice is good, yes - but everyone having their own language is not good, a standard language is better. Perl gives you many ways to do things. This makes it easier to write, but far harder to read, because you need to know all the different ways. In the long run this tends to bite you.
I find python more fun, because it looks cleaner. It's got more structure. I'm a mathematician and my aesthetics may not be "normal", but to me something near-geometric is preferrable to the "tangle" I seem to get when programming perl. As far as perl dying is concerned, how is that community going? Is it shrinking? Is it aging in almost real time? If either of those are true, perl is dying.
Why doesn't the installer tell you you're supposed to run the disklabel program
No, the reason bittorrent is faster is that it stops leeching, but without requiring you to wait in line for ages to get files.
Don't they fill it in when registering windows?
Because the installer doesn't tell you how.
Why? They're just doing what they've been taught. They've been taught that anyone can use a computer (that's what all the MS advertising says). They've been taught that anyone can use the internet (that's what all the AOL advertising says). They've been taught that if they don't click yes at dialog boxes, things don't work (A very similar security warning is seen two or three times when logging in to hotmail. So users learn they're not serious) How can you blame them for simply doing what they've been taught?
But why doesn't the installer tell you that? There's no web browser on the install cd, is one expected to do each stage and then reboot to the old OS to see what to do next?
No, this is the first time an extraterrestrial satellite has taken a picture of another.
Did you pay or already have an account (meaning you found it before)? Because if not I don't think there's any way you can find something good in 13 seconds.
Lol, I know about him. I have a classmate who bears an uncanny resemblance to him.
Of course I tried it. Boot, ooh, kernel messages, nice, hmm, hda is wd0, that's...interesting... right, NO I definitely don't want to use the whole disk. Right, cfdisk - not found, oh well, I can use fdisk if I need to. Read the help, add partition, primary, change type to BSD, set active, write, quit. Run install. Can't find the partition. What? I just made it. fdisk again to check, yep, the partition is there, why can't install see it. Run install again. Still not found. Bah. I give up.
It won't be. It won't even be popular non-mainstream. A program that requires reading the docs to use is a flawed program. If you're doing something really complex, sure, one should expect to read the documentation, but installing alongside an existing partition is not complex.
I know the basics, but prior to this thread I didn't even know about the existence of the disklabel program. The installer gives you no clues at all. A "you probably want to do this" message would make the world of difference without much effort.
I don't mind the lack of graphics, but the installer doesn't seem to give any pointers about how to install. It tells me I need to make a primary partition for bsd and mark it as active, so I do that. Then I try to run install and it can't find a partition. I appreciate that reading the documentation is important, but it should be doable at least for basic common configurations (not just the entire disk) without needing the doc.
Doesn't that mean you have to trust not only the makers of the protocol, but also absolutely anyone, since anyone can run a server and give themselves an ID on their own server? Making the ID worthless?
The reason they released it was it was all they had. They didn't have time to test and integrate a new version. It's the firefox people's fault for having the vulnerabilities, I think they're just trying to distract from their own failings.