BTW, am I the only one who thinks that securelevels stink?
IMHO it would be a better idea to be able to select the features securelevels imply individually. That way, one could still use the securelevel settings in/etc/rc.conf by just making/etc/rc setting a group of individual "securesettings".
I mean, just because I happen to like rewriting my firewall rules doesn't mean I want anybody to be able to write to kmem, or to remove noschg!
Um, the size a system takes is not that directly related to security. You can stuff quite a lot of remote exploits in 65MB.
If you want an insecure system, Linux already does more for you that it's distributions being bloated. If you chose Linux for security relevant systems - and I consider pure desktop use as one of the most dangerous ones - you should probably think of better using a system made by people who care about security more than they care about cloning the most insecure operating system available.
I started using Linux with SuSE (no surprise, it's by far the most used Linux distro in Germany). The "configuration scheme" was what made me switch.
Not only I never understood why they attempt to get the disadvantages from both BSD- and SYSV-Init, obviously SuSEConfig is intended to familiarize former Windows lusers (like me back then) with the new environment in that it randomly destroys your work. It is frankly the crappiest piece of shitty software I ever met - and hey, I develop shitty software for a living!
The one thing it did that made me look for alternatives to SuSE was it happily fixing my sendmail.cf. I just had gotten it right after several days of work (I really was a newbie right then, remember). Then I dared to install some random RPM using YaST. Of course, installing something as important as an additional set of fortune cookies means that you have to run SuSEConfig, and scince SuSEConfig is quite dumb, it changes everything in some random way - including, of course,/etc/whatnot/sendmail.cf. A nice example for the bad quality of closed source software.
However, this made me try other systems - first Debian for a while, with short interjections of RedHat and Mandrake (come on - you don't want to use a system made by people responsible for that/etc/motd, do you?). When Debian once ate my package database, rendering apt unusable, I looked at the BSDs.
So, at the end of the day, now I'm using FreeBSD and OpenBSD depending on my mood, and learned a lot on the way (I really was surprised that "Have a lot of fun" wasn't somehow compiled in the kernel...). I guess I somehow have to thank those bavarian morons for producing the worst software mankind ever encountered...
Sorry, but while this argument is quite commonly used to propagate the GPL, I think it's plain wrong. Nobody can make any Mono component, or any other Free-and-not-copylefted code, non-free. There is no "taking away our freedom".
All that is possible is to release derived work without a Free license. It's not as if the original Free code would go away. You do not lose anything - you just do not get more. That's quite a difference.
Everytime I think of Microsoft getting the benefit of Free Os in networking code, I feel happy, because the net would be a even more ugly and insecure place if those idiots would try to do it themselves.
If they wanted to be taken serious they also shouldn't have printed this 5 years old SGML article. Ok, it may be edited to reflect some changes, but when? 3 Years ago? SGML is history, and XML is more than SGML with some obscure features removed right now, it's a whole new tool-chain (and both Docbook and TEI are long available for XML, btw.).
As well, I really liked the distro article, because it mentions some smaller ditros I never heard of. It was fun reading it, some months ago on www.distrowatch.com.
Sorry, but this magazine doesn't do anything good to the Free Software community. It just lets us look boring and unprofessional.
Wow, that "C is here to stay" aticle must be the most content-free piece of text I ever read. Come on: C is good because some guys wanted to use a PDP some 30 Years ago, and because Perl is mainly used for text processing?
I see that there are areas where C may still be useful, like bare-metal hardware access, but the rest is purely historical accident. OK, there are lots of C code in use. There are also lots of COBOL programms. However, there are also languages (basically all except C, and by inheritance C++) where there was more progress in the last decades than finding funny new ways to get root by exploiting new classes of bugs (first buffer overflows, then format string errors...)
What is it that there are so many C advocates? I just don't get it...
Yes - on Linux I can log in remotely and kill(1) it.
And of course, just because Wine clones a fucked up API it doesn't mean they have to clone the implementation bugs, too.
If Wine were really good, it would be possible that Windows Apps would indeed run more stable than on Windows itself. Then again, it is likely that Wine indroduces it's own bugs (Just guessing - I never used it).
Funny - I never saw a page listing "BSD-compatible licenses", nor did I realize BSDers acting like the spanish inquisition when someone doesn't play to their rules, like some others do.
Linux (even the RedHat distro) has the GPL protecting it. Even AOL/TW's big lawyers can't break it. Why is it such a bad thing??
Maybe because you're wrong.
First, nobody knows if lawyers (or judges, they still exists, you know) could "break" the GPL. Right now, we only know nobody tried it yet.
Furthermore, the important part of Red Hat are not protected by the GPL. Neither their name and credibility, nor their customer base is GPLed. (In fact, I don't even know if all their software is - AFAIK SuSEs Yast is closed source, e.g.)
Unless they block everything except port 80 (and maybe 25, 110, 443 and a few others), you could always just use another port. The only problem is to agree on a well-known port fast enough to switch every so often.
Maybe we could distribute the portnumber-of-the-week for Gnutella on Freenet, and vice versa:)
The only chance if we don't want to look for another service twice a week is not to depend on central servers and, even more important, on single companies that can simply decide not to support Linux any more or to suddendly charge per download etc.
There are a number of Free alternatives, like Gnutella, Freenet or the FastTrack-Clone giFT. Use them. Make them work. Make them get big. Donate billions of dollars for advertising:)
This is nothing. The Deutsche Telekom once sued a company because they produced a magenta-colored book, and another one because their name was "T-Media". The letter t and the color magenta are part of their trademark.
The "root can do everything" is, frankly, a very stupid idea. That's
why people try to get rid of it.
There is no reason that, just because a web server needs to bind to
a port <1024, he has to be able to write to/etc/shadow. There are
capabilities to solve this, but right now they are not an option - not
only they aren't portable, even in Linux, you have to use kernel
patches to really use their power - and applying random patches to
your kernel generally is not what you want to do when you care about
security. (<flaimbait>So is running Linux,
btw.</flaimbait>:)
"Linux - an alternative to Microsoft's windows"
"the Linux ELF format"
Don't care much about the virus - People so stupid to be as surprised
that there is an alternative to Windows that they forget that there
may be even more OSes besides Windows and that Linux thingie
seem to be greater a thread.
So you think because their newspeak matches the definition of the RDF
data model, so they are right? This data model is just a directed
labled graph. They could have as well have patented the idea of a set,
and sue Oracle for using it in their databases.
Slashdot Mirror
Try following links before you post them. Active Directory is no fs.
Because they will decide who your peers are.
IMHO it would be a better idea to be able to select the features securelevels imply individually. That way, one could still use the securelevel settings in /etc/rc.conf by just making /etc/rc setting a group of individual "securesettings".
I mean, just because I happen to like rewriting my firewall rules doesn't mean I want anybody to be able to write to kmem, or to remove noschg!Icons? cat /etc/motd, and you will love the icons...
Oh my god, please not everyone!
If you want an insecure system, Linux already does more for you that it's distributions being bloated. If you chose Linux for security relevant systems - and I consider pure desktop use as one of the most dangerous ones - you should probably think of better using a system made by people who care about security more than they care about cloning the most insecure operating system available.
Not only I never understood why they attempt to get the disadvantages from both BSD- and SYSV-Init, obviously SuSEConfig is intended to familiarize former Windows lusers (like me back then) with the new environment in that it randomly destroys your work. It is frankly the crappiest piece of shitty software I ever met - and hey, I develop shitty software for a living!
The one thing it did that made me look for alternatives to SuSE was it happily fixing my sendmail.cf. I just had gotten it right after several days of work (I really was a newbie right then, remember). Then I dared to install some random RPM using YaST. Of course, installing something as important as an additional set of fortune cookies means that you have to run SuSEConfig, and scince SuSEConfig is quite dumb, it changes everything in some random way - including, of course, /etc/whatnot/sendmail.cf. A nice example for the bad quality of closed source software.
However, this made me try other systems - first Debian for a while, with short interjections of RedHat and Mandrake (come on - you don't want to use a system made by people responsible for that /etc/motd, do you?). When Debian once ate my package database, rendering apt unusable, I looked at the BSDs.
So, at the end of the day, now I'm using FreeBSD and OpenBSD depending on my mood, and learned a lot on the way (I really was surprised that "Have a lot of fun" wasn't somehow compiled in the kernel...). I guess I somehow have to thank those bavarian morons for producing the worst software mankind ever encountered...
Sorry, but while this argument is quite commonly used to propagate the GPL, I think it's plain wrong. Nobody can make any Mono component, or any other Free-and-not-copylefted code, non-free. There is no "taking away our freedom".
All that is possible is to release derived work without a Free license. It's not as if the original Free code would go away. You do not lose anything - you just do not get more. That's quite a difference.
Unfortunatly, they usually do it themselves.
As well, I really liked the distro article, because it mentions some smaller ditros I never heard of. It was fun reading it, some months ago on www.distrowatch.com.
Sorry, but this magazine doesn't do anything good to the Free Software community. It just lets us look boring and unprofessional.
I see that there are areas where C may still be useful, like bare-metal hardware access, but the rest is purely historical accident. OK, there are lots of C code in use. There are also lots of COBOL programms. However, there are also languages (basically all except C, and by inheritance C++) where there was more progress in the last decades than finding funny new ways to get root by exploiting new classes of bugs (first buffer overflows, then format string errors...)
What is it that there are so many C advocates? I just don't get it...
Yes - on Linux I can log in remotely and kill(1) it. And of course, just because Wine clones a fucked up API it doesn't mean they have to clone the implementation bugs, too. If Wine were really good, it would be possible that Windows Apps would indeed run more stable than on Windows itself. Then again, it is likely that Wine indroduces it's own bugs (Just guessing - I never used it).
Funny - I never saw a page listing "BSD-compatible licenses", nor did I realize BSDers acting like the spanish inquisition when someone doesn't play to their rules, like some others do.
Try to have a look at Emacspeak. Perhaps that's what you want.
Maybe because you're wrong.
First, nobody knows if lawyers (or judges, they still exists, you know) could "break" the GPL. Right now, we only know nobody tried it yet.
Furthermore, the important part of Red Hat are not protected by the GPL. Neither their name and credibility, nor their customer base is GPLed. (In fact, I don't even know if all their software is - AFAIK SuSEs Yast is closed source, e.g.)
Maybe we could distribute the portnumber-of-the-week for Gnutella on Freenet, and vice versa :)
There are a number of Free alternatives, like Gnutella, Freenet or the FastTrack-Clone giFT. Use them. Make them work. Make them get big. Donate billions of dollars for advertising :)
Hardly news, guys.
Maybe that's why there is no CMYK in the Gimp
And vice versa, of course.
The "root can do everything" is, frankly, a very stupid idea. That's why people try to get rid of it.
There is no reason that, just because a web server needs to bind to a port <1024, he has to be able to write to /etc/shadow. There are
capabilities to solve this, but right now they are not an option - not
only they aren't portable, even in Linux, you have to use kernel
patches to really use their power - and applying random patches to
your kernel generally is not what you want to do when you care about
security. (<flaimbait>So is running Linux,
btw.</flaimbait> :)
"Linux - an alternative to Microsoft's windows"
"the Linux ELF format"
Don't care much about the virus - People so stupid to be as surprised that there is an alternative to Windows that they forget that there may be even more OSes besides Windows and that Linux thingie seem to be greater a thread.
Funny, XEmacs is one of the first things I install on a windows box I am supposed to get work done on.
On the other hand - how well is Eclipse suited for use on a text-only terminal?
So you think because their newspeak matches the definition of the RDF data model, so they are right? This data model is just a directed labled graph. They could have as well have patented the idea of a set, and sue Oracle for using it in their databases.