Re:Imminent death of Usenet Again!
on
Is Usenet Dying?
·
· Score: 1
Yup, not really "copied" because it was the same developers.
http://notes.net/whatisnotes has the history of Lotus Notes and it's orgin from "PLATO Group Notes". PLATO Notes dates from 1973, Lotus Notes from 1985, so this work was being done in parallel to the standard Internet protocols. --
Anyone else notice that Slashdot has a JimJag? Is this the same JimJag that kept A/UX alive for so many years? (If so, Thanks! You certainly made my life easier.)
As an editorial, A/UX was the greatest thing Apple came out with until OSX Server. If they would have run with that ball, they would have had their 'modern' OS ten years ago.
When the "Copeland" efforts failed, Be was in negotiations with Apple to be purchased.
Maybe they didn't start the company with the business plan of being purchased by Apple (although, with Gasse there, that's doubtful), but when Apple came to them, they were willing to talk.
I agree that Apple got a better deal with NeXT, although they probably could have a product to market quicker with Be.
(PS -- Gasse is responsible for some of the worst decisions that Apple ever made, such as the look-and-feel lawsuits, refusing to license clones, and the insistence that Macs would use proprietary networking. It's a good thing that they didn't bring him back.) --
I take it from your statement that Linux on 68K Macs isn't going well. Too bad, because the Quadra 950 is a far better machine than most 486s -- Fast SCSI, supports at least 128MB RAM(maybe more), 24-bit video, etc. (I used to run A/UX on one of these, and it was a rock solid file server.)
--
Re:Drag into trash is *still* a bad UI design.
on
PPCLinux.Apple.Com
·
· Score: 1
}} they could have picked anything as the "secure attention key"!
Ctrl+Alt+Delete is recognized by the hardware (which is why it works during the BIOS POST sequence). That supposedly makes it harder to trap the SAK and capture someone's password with a trojan. --
Re:Drag into trash is *still* a bad UI design.
on
PPCLinux.Apple.Com
·
· Score: 1
But even "Put Away" is a disaster. What does it mean? When appiled to a file, it undoes a previous move operation, but when applied to a disk, it unmounts it.
To further the confusion, MacOS also includes an "Eject" command, which ejects the disk, but does not unmount it. Which means the MacOS will eventually force you (with a blocking dialog box) to reinsert the disk. Of course at this point, the user is thinking "I already ejected that disk, why do I have to put it back in?!?"
A good solution would be an Unmount icon on the desktop for drag operations, or just make "Eject" = "Put Away"= Unmount, because the days of multi-floppy systems are long over. Of couse, Apple has heard the complaints, and hasn't fixed it for 15 years. --
Re:Usenet Dying? What about Gopher?
on
Is Usenet Dying?
·
· Score: 1
The Center of GopherSpace (gopher://gopher.tc.umn.edu/ - Slashdot won't buy a gopher HREF) is amazingly still on line, so it ain't dead yet.
And, thanks to Microsoft's monopolistic practices, every Windows user has a Gopher client right on their desktop.
(Maybe someone should start a Gopher Warez Network -- it would be better than that Hotline crap.)
Signed, --umn.edu vet who got his start with gopher --
Re:Imminent death of Usenet Again!
on
Is Usenet Dying?
·
· Score: 1
Just as a historical footnote, there used to be a Lotus Notes (that other thing) Internet that provided mail, various newsfeeds, and discussion forums between Lotus Notes sites.
Similar in concept to Usenet or WWW news sites, and now pretty much entirely dead except for the stuff hosted by Lotus. --
Go back and read his example -- AMD isn't currently selling a competitor to the Intel Xeon. Dell can't afford to lose the server business with their highest profit margins (as well as "end-to-end" sales.)
Besides, switching from 100% Intel (with associated discounts) to AMD carries a bunch of re-engineering and support costs, as well as the risk of chipset shortages and so on. (essentially the same boat they're in with Intel.)
However, any smart business doesn't rely on a single supplier. (But, historically, PC companies haven't been so smart.) So, like Compaq and Gateway, I would expect Dell to slowly introduce AMD-based computers into their lineup, starting at the low-end and hoping that Intel doesn't get too pissy. --
Correct. When Compaq bought Digital, they were looking at the services division, and they probably didn't really have any idea what they were getting into. However, once looking at the books, they realized that the midrange business (Unix and VMS on Alpha) either was or could be enormously profitable.
The funny thing is, they originally thought DEC could be assimilated into Compaq's traditional Wintel-follower PC business, but then found out in order to really get that big midrange profit margin, Compaq had to essentially become Digital Equipment. This has led to a bunch of unresolved management turmoil and strategy shifts (reminiscent of DEC in the dark days), and probably explains why Compaq's current server equipment is so nice, and their desktop PCs are so crappy. --
There are too few IRQs, too few base addresses, and too few DMA channels not to mention DMA can only access the lower 16 Megs of memory.
Yes, a valid bitch, but only if you are speaking of 1999's PC. Now that Microsoft and Intel have effectively killed ISA (the politically correct name for the IBM PC AT design from 1984), the IRQ and DMA problems have been solved.
As for including 8088 CPU instructions, the burden is on you to prove that they significantly add to the cost and complexity of a modern x86 processor. My assertation is that the PC hardware prices don't lie -- even with 8086/80286 compatibility, x86 is still the best bang for the buck.
Another way to think about this is that breaking backwards compatibility fractures the market, which reduces economies of scale, which increases the price. (But I'm somone who runs 8088 DOS programs on my Pentium II, so take it for what it's worth.)
This argument can be extended to validate x86-based PDAs (look ma, no legacy ISA stuff!). If a software developer can ship the same or similar package on desktops and handhelds, it will increase economies of scale, and the number of handheld applications will explode.
(It should be noted that the Mac rumor sites used to frequently toss around the idea that Apple would ship a Newton iMate-like handheld device, except that it would be built on standard Mac PPC hardware and run a modified MacOS. Excellent idea, in my book.) --
Yes, maybe the great Carmack will save Linux users from the dozens of undocumented winmodem chipsets. I'd rather have him work on more worthy products, like video drivers, myself.
Or maybe people will just go out and buy real modems. --
The reason Apple wouldn't allow it is because the current situation allows Apple to continue to gouge their customers by forcing them to buy Apple hardware.
True. But another good reason is that supporting an OS on commodity PC hardware is expensive.
If they did ship an x86 Mac, people would continually bitch that there wasn't drivers for their nVidia cards, their winmodems, their oddball SCSI cards, their software-driven printers, their old scanners, ATA-66, weird PCMCIA chipsets and so on.
The inability to supply every possible driver on the x86 platform is a huge concern -- it's hampered the success of Solaris x86, OS/2, BeOS, and pretty much every other OS except Windows and Linux (and even Linux is always going to have it's winmodem problems.)
Besides, the article is about "Darwin" not MacOS proper. Darwin is the Mach kernel and the BSD layer. It's been open sourced (= no direct revenue for Apple), and a x86 source tree already exists with a bunch of drivers from the OpenStep/Intel era. --
For most 8080/Z80 CP/M users, the 8088 IBM PC or a machine running CP/M-86 was a way that they could run their existing software, only faster.
All of the major CP/M-80 packages (WordStar, VisiCalc, etc) were ported to the 8086 almost immediately. Plus, new, supposedly better software like WordPerfect and Lotus 123 was coming out only for the 8086. Plus the IBM PC had other niceities like a standard disk format and color graphics.
(People hit the maximum sheet size in Visicalc on 64K 8080 machines pretty quickly. Even with 512K or 640K, people also were bumping against Lotus's maximum sheet size early on. Lotus finally had to drive a specification for expanded memory to solve the problem. In short, the user applications were driving the upgrades in those days, not the games or the server stuff.) --
Windows Media player has no bearing on the success or failure of Apple, of Linux, or even of Windows.
They create a Mac version (perpetually in beta) so they can tell content providers that they cover 99 percent of the end user market (like Quicktime and Real do). Linux/Unix support would only matter in certain markets where Unix support is important, such as academia or engineering. --
Well, they didn't have any trouble porting to windows
QuickTime for Windows contains a good chunk of the MacOS API. So, they pretty much had to create an entire Macintosh-compatiblity library to get QuickTime to run. It was a big job.
(Some non-media MacOS programs have been ported to Windows using Quicktime.) --
So, the obvious question: Is there any way to get the Solaris DPS libraries running under Linux? Or is ghostscript good enough that it's not worth it?
I know about ibcs (?), but had trouble getting it to do anything with Solaris x86 apps.
(Yup, solaris stuff is not free software, but on the otherhand, the licence from Sun for personal solaris is cheeep, and it would be nice to make use of the software I paid for.) --
Then again, you could argue that Quake * is a special case, because in all likelyhood the game would be successful enough to be eventually ported to every platform under the sun. So in that case, it's cheaper to design it cross-platform up front.
Note that this logic doesn't hold for Generic KnockOff Shootumup for Windows and Playstation, which (like most games) would be lucky to break even on dev costs. But it does give people writing to Direct3D something to chew over. --
You are ignoring that they are right. Something on the order of 0.5% of PC users actually use Linux. And you know what? Probably only 10 or 20 percent of THAT number would actually be willing to pay for software, let alone a game.
I think you are underestimating the Linux market. First of all, most PC users and modern PCs are locked up at work. Now this might have been fine for after hours Starcraft or Quake, games like Quake III are pretty much no-go on work machines with their cheapo video systems.
A huge percentage of computers that are in homes are outdated, or in the hands of those not interested in games or not competant enough to install anything anyway.
So, what do you have left? A small percentage of high powered machines in the hands of knowledgeable users with money and time on their hands -- in short, just the kind of people who might be running Linux part time.
The thing to note here is that the market for games is very different than the market for spreadsheets, "QuickBooks", or "ACT!". (And, given that Linux is an Open Source Unix, the bet is more 'late adopters' are interested in the Unixness than the Open Sourceness.)
Now, I don't think the Linux game market is huge, but it is big enough to attract more vendor attention than any other commercial software market (short of RDBMS), in fact it is almost getting to the point where Linux is competitive with MacOS, and that's a huge milestone. --
Right on, despite the retail costs and disk footprint of a MS|Corel|Lotus Office, from a corporate standpoint, there's enormous cost savings in deploying a solution that meets 90% of your user base's needs, even though each particular user might only use 10% of the functionality.
Imagine trying to decide which users need slideshow software, or 3-D charting software, or revision control software. Or worse, the nightmare of a swarm of techs running around installing this stuff after someone mailed out a slideshow, 3D chart, or revision controlled document. It's easier to waste $20 worth of disk space for each user and forget about it.
(I was there back in the DOS days of smaller, more focused tools, and it was an administration disaster, not to mention the licence surveys. Linux OSes solve this problem by giving you the more focused tools, but also giving you lots and lots of them by default, which again leads to lots of disk space usage.) --
I agree that simpler and componentized equals more secure and easier to maintain.
Microsoft deserves to be ripped on this -- Index Server, complete with huge security hole, gets installed and enabled by default on every IIS server since version 3.0. (Even though it would only take someone 10 seconds of pointing and clicking to enable it, if in fact they really needed it.)
Whether or not it's part of the "OS" is a muddy issue. Microsoft likes to call anything that comes inside of the shiny box that says "Windows" part of the OS, and whatever happens (good or bad), "Windows" takes the credit.
Linux users, on the other hand, like to point at lpd and wuftpd and even though they came in the shiny "RedHat" box, and were enabled by default, the spin is to say "That's just a user application, not part of the (holy) Linux kernel, and therefore is not a serious problem (and won't get posted to slashdot, etc)." Not what someone running a RedHat box wants to hear. --
Good idea -- I've seen new threads where the first 50 posts have had at least 50 mod points expended on them. (A few 5s, 3s, a bunch of 2s lots of -1s).
What happens to the next 300 {good, bad, trugly} posts? Nothing! The moderators blow their load in the first 30 minutes. --
It was possible in the control file of a print job to specify arguments to sendmail. By careful manipulation of control and data files, this could cause sendmail to be executed with a user-specified configuration file. This could lead very easily to a root compromise.
A security bug was found in userhelper; the bug can be exploited to provide local users with root access
Users who had csh/tcsh as their login shell could be vulnerable to having arbitrary shell code run by their shell on login.
By opening a large number of connections to the log daemon, the user could make the system unresponsive.
A bug in the processing of NXT records can theoretically allow a remote attacker to gain access to the DNS server as the user running bind (by default, root).
With ypserv, local administrators in the NIS domain could possibly inject password tables
Remote and local intruders may be able exploit these vulnerabilities to execute arbitrary code as the user running the ftpd daemon, usually root.
Since screen is not setuid root, this means that it leaves the ptys with insecure permissions. The updated packages restore the Unix98 pty support.
(Sure, these aren't kernel bugs, but neither are the Windows2000 problems mentioned in the linked article. To be fair, there's plenty of good reading at http://www.microsoft.com/technet/security/current. asp?ID=4&Parent=1, however I'm not going to dismiss what there as unimportant because they aren't kernel bugs.) --
Slashdot Mirror
Yup, not really "copied" because it was the same developers.
http://notes.net/whatisnotes has the history of Lotus Notes and it's orgin from "PLATO Group Notes". PLATO Notes dates from 1973, Lotus Notes from 1985, so this work was being done in parallel to the standard Internet protocols.
--
Well, I mentioned it
Anyone else notice that Slashdot has a JimJag? Is this the same JimJag that kept A/UX alive for so many years? (If so, Thanks! You certainly made my life easier.)
As an editorial, A/UX was the greatest thing Apple came out with until OSX Server. If they would have run with that ball, they would have had their 'modern' OS ten years ago.
--
When the "Copeland" efforts failed, Be was in negotiations with Apple to be purchased.
Maybe they didn't start the company with the business plan of being purchased by Apple (although, with Gasse there, that's doubtful), but when Apple came to them, they were willing to talk.
I agree that Apple got a better deal with NeXT, although they probably could have a product to market quicker with Be.
(PS -- Gasse is responsible for some of the worst decisions that Apple ever made, such as the look-and-feel lawsuits, refusing to license clones, and the insistence that Macs would use proprietary networking. It's a good thing that they didn't bring him back.)
--
I take it from your statement that Linux on 68K Macs isn't going well. Too bad, because the Quadra 950 is a far better machine than most 486s -- Fast SCSI, supports at least 128MB RAM(maybe more), 24-bit video, etc. (I used to run A/UX on one of these, and it was a rock solid file server.)
--
}} they could have picked anything as the "secure attention key"!
Ctrl+Alt+Delete is recognized by the hardware (which is why it works during the BIOS POST sequence). That supposedly makes it harder to trap the SAK and capture someone's password with a trojan.
--
But even "Put Away" is a disaster. What does it mean? When appiled to a file, it undoes a previous move operation, but when applied to a disk, it unmounts it.
To further the confusion, MacOS also includes an "Eject" command, which ejects the disk, but does not unmount it. Which means the MacOS will eventually force you (with a blocking dialog box) to reinsert the disk. Of course at this point, the user is thinking "I already ejected that disk, why do I have to put it back in?!?"
A good solution would be an Unmount icon on the desktop for drag operations, or just make "Eject" = "Put Away"= Unmount, because the days of multi-floppy systems are long over. Of couse, Apple has heard the complaints, and hasn't fixed it for 15 years.
--
The Center of GopherSpace (gopher://gopher.tc.umn.edu/ - Slashdot won't buy a gopher HREF) is amazingly still on line, so it ain't dead yet.
And, thanks to Microsoft's monopolistic practices, every Windows user has a Gopher client right on their desktop.
(Maybe someone should start a Gopher Warez Network -- it would be better than that Hotline crap.)
Signed,
--umn.edu vet who got his start with gopher
--
Just as a historical footnote, there used to be a Lotus Notes (that other thing) Internet that provided mail, various newsfeeds, and discussion forums between Lotus Notes sites.
Similar in concept to Usenet or WWW news sites, and now pretty much entirely dead except for the stuff hosted by Lotus.
--
Go back and read his example -- AMD isn't currently selling a competitor to the Intel Xeon. Dell can't afford to lose the server business with their highest profit margins (as well as "end-to-end" sales.)
Besides, switching from 100% Intel (with associated discounts) to AMD carries a bunch of re-engineering and support costs, as well as the risk of chipset shortages and so on. (essentially the same boat they're in with Intel.)
However, any smart business doesn't rely on a single supplier. (But, historically, PC companies haven't been so smart.) So, like Compaq and Gateway, I would expect Dell to slowly introduce AMD-based computers into their lineup, starting at the low-end and hoping that Intel doesn't get too pissy.
--
Correct. When Compaq bought Digital, they were looking at the services division, and they probably didn't really have any idea what they were getting into. However, once looking at the books, they realized that the midrange business (Unix and VMS on Alpha) either was or could be enormously profitable.
The funny thing is, they originally thought DEC could be assimilated into Compaq's traditional Wintel-follower PC business, but then found out in order to really get that big midrange profit margin, Compaq had to essentially become Digital Equipment. This has led to a bunch of unresolved management turmoil and strategy shifts (reminiscent of DEC in the dark days), and probably explains why Compaq's current server equipment is so nice, and their desktop PCs are so crappy.
--
There are too few IRQs, too few base addresses, and too few DMA channels not to mention DMA can only access the lower 16 Megs of memory.
Yes, a valid bitch, but only if you are speaking of 1999's PC. Now that Microsoft and Intel have effectively killed ISA (the politically correct name for the IBM PC AT design from 1984), the IRQ and DMA problems have been solved.
As for including 8088 CPU instructions, the burden is on you to prove that they significantly add to the cost and complexity of a modern x86 processor. My assertation is that the PC hardware prices don't lie -- even with 8086/80286 compatibility, x86 is still the best bang for the buck.
Another way to think about this is that breaking backwards compatibility fractures the market, which reduces economies of scale, which increases the price. (But I'm somone who runs 8088 DOS programs on my Pentium II, so take it for what it's worth.)
This argument can be extended to validate x86-based PDAs (look ma, no legacy ISA stuff!). If a software developer can ship the same or similar package on desktops and handhelds, it will increase economies of scale, and the number of handheld applications will explode.
(It should be noted that the Mac rumor sites used to frequently toss around the idea that Apple would ship a Newton iMate-like handheld device, except that it would be built on standard Mac PPC hardware and run a modified MacOS. Excellent idea, in my book.)
--
Correct, the systems weren't compatible.
The only 'upgrade' I remember was the DEC Rainbow, which had both a Z-80 and a 8086, and could run both versions of CP/M (and possibly MS-DOS).
--
Yes, maybe the great Carmack will save Linux users from the dozens of undocumented winmodem chipsets. I'd rather have him work on more worthy products, like video drivers, myself.
Or maybe people will just go out and buy real modems.
--
The reason Apple wouldn't allow it is because the current situation allows Apple to continue to gouge their customers by forcing them to buy Apple hardware.
True. But another good reason is that supporting an OS on commodity PC hardware is expensive.
If they did ship an x86 Mac, people would continually bitch that there wasn't drivers for their nVidia cards, their winmodems, their oddball SCSI cards, their software-driven printers, their old scanners, ATA-66, weird PCMCIA chipsets and so on.
The inability to supply every possible driver on the x86 platform is a huge concern -- it's hampered the success of Solaris x86, OS/2, BeOS, and pretty much every other OS except Windows and Linux (and even Linux is always going to have it's winmodem problems.)
Besides, the article is about "Darwin" not MacOS proper. Darwin is the Mach kernel and the BSD layer. It's been open sourced (= no direct revenue for Apple), and a x86 source tree already exists with a bunch of drivers from the OpenStep/Intel era.
--
For most 8080/Z80 CP/M users, the 8088 IBM PC or a machine running CP/M-86 was a way that they could run their existing software, only faster.
All of the major CP/M-80 packages (WordStar, VisiCalc, etc) were ported to the 8086 almost immediately. Plus, new, supposedly better software like WordPerfect and Lotus 123 was coming out only for the 8086. Plus the IBM PC had other niceities like a standard disk format and color graphics.
(People hit the maximum sheet size in Visicalc on 64K 8080 machines pretty quickly. Even with 512K or 640K, people also were bumping against Lotus's maximum sheet size early on. Lotus finally had to drive a specification for expanded memory to solve the problem. In short, the user applications were driving the upgrades in those days, not the games or the server stuff.)
--
Certainly not if Intel prices IA64 like DEC/Compaq prices the Alpha.
If, on the other hand, IA64 systems are within 30% of the cost of IA32 systems, IA64 could catch on.
--
Windows Media player has no bearing on the success or failure of Apple, of Linux, or even of Windows.
They create a Mac version (perpetually in beta) so they can tell content providers that they cover 99 percent of the end user market (like Quicktime and Real do). Linux/Unix support would only matter in certain markets where Unix support is important, such as academia or engineering.
--
Well, they didn't have any trouble porting to windows
QuickTime for Windows contains a good chunk of the MacOS API. So, they pretty much had to create an entire Macintosh-compatiblity library to get QuickTime to run. It was a big job.
(Some non-media MacOS programs have been ported to Windows using Quicktime.)
--
So, the obvious question: Is there any way to get the Solaris DPS libraries running under Linux? Or is ghostscript good enough that it's not worth it?
I know about ibcs (?), but had trouble getting it to do anything with Solaris x86 apps.
(Yup, solaris stuff is not free software, but on the otherhand, the licence from Sun for personal solaris is cheeep, and it would be nice to make use of the software I paid for.)
--
Then again, you could argue that Quake * is a special case, because in all likelyhood the game would be successful enough to be eventually ported to every platform under the sun. So in that case, it's cheaper to design it cross-platform up front.
Note that this logic doesn't hold for Generic KnockOff Shootumup for Windows and Playstation, which (like most games) would be lucky to break even on dev costs. But it does give people writing to Direct3D something to chew over.
--
You are ignoring that they are right. Something on the order of 0.5% of PC users actually use Linux. And you know what? Probably only 10 or 20 percent of THAT number would actually be willing to pay for software, let alone a game.
I think you are underestimating the Linux market. First of all, most PC users and modern PCs are locked up at work. Now this might have been fine for after hours Starcraft or Quake, games like Quake III are pretty much no-go on work machines with their cheapo video systems.
A huge percentage of computers that are in homes are outdated, or in the hands of those not interested in games or not competant enough to install anything anyway.
So, what do you have left? A small percentage of high powered machines in the hands of knowledgeable users with money and time on their hands -- in short, just the kind of people who might be running Linux part time.
The thing to note here is that the market for games is very different than the market for spreadsheets, "QuickBooks", or "ACT!". (And, given that Linux is an Open Source Unix, the bet is more 'late adopters' are interested in the Unixness than the Open Sourceness.)
Now, I don't think the Linux game market is huge, but it is big enough to attract more vendor attention than any other commercial software market (short of RDBMS), in fact it is almost getting to the point where Linux is competitive with MacOS, and that's a huge milestone.
--
Right on, despite the retail costs and disk footprint of a MS|Corel|Lotus Office, from a corporate standpoint, there's enormous cost savings in deploying a solution that meets 90% of your user base's needs, even though each particular user might only use 10% of the functionality.
Imagine trying to decide which users need slideshow software, or 3-D charting software, or revision control software. Or worse, the nightmare of a swarm of techs running around installing this stuff after someone mailed out a slideshow, 3D chart, or revision controlled document. It's easier to waste $20 worth of disk space for each user and forget about it.
(I was there back in the DOS days of smaller, more focused tools, and it was an administration disaster, not to mention the licence surveys. Linux OSes solve this problem by giving you the more focused tools, but also giving you lots and lots of them by default, which again leads to lots of disk space usage.)
--
I agree that simpler and componentized equals more secure and easier to maintain.
Microsoft deserves to be ripped on this -- Index Server, complete with huge security hole, gets installed and enabled by default on every IIS server since version 3.0. (Even though it would only take someone 10 seconds of pointing and clicking to enable it, if in fact they really needed it.)
Whether or not it's part of the "OS" is a muddy issue. Microsoft likes to call anything that comes inside of the shiny box that says "Windows" part of the OS, and whatever happens (good or bad), "Windows" takes the credit.
Linux users, on the other hand, like to point at lpd and wuftpd and even though they came in the shiny "RedHat" box, and were enabled by default, the spin is to say "That's just a user application, not part of the (holy) Linux kernel, and therefore is not a serious problem (and won't get posted to slashdot, etc)." Not what someone running a RedHat box wants to hear.
--
Good idea -- I've seen new threads where the first 50 posts have had at least 50 mod points expended on them. (A few 5s, 3s, a bunch of 2s lots of -1s).
What happens to the next 300 {good, bad, trugly} posts? Nothing! The moderators blow their load in the first 30 minutes.
--
You say: nope no major security hole bugs here
. asp?ID=4&Parent=1, however I'm not going to dismiss what there as unimportant because they aren't kernel bugs.)
RedHat says:
It was possible in the control file of a print job to specify arguments to sendmail. By careful manipulation of control and data files, this could cause sendmail to be executed with a user-specified configuration file. This could lead very easily to a root compromise.
A security bug was found in userhelper; the bug can be exploited to provide local users with root access
Users who had csh/tcsh as their login shell could be vulnerable to having arbitrary shell code run by their shell on login.
By opening a large number of connections to the log daemon, the user could make the system unresponsive.
A bug in the processing of NXT records can theoretically allow a remote attacker to gain access to the DNS server as the user running bind (by default, root).
With ypserv, local administrators in the NIS domain could possibly inject password tables
Remote and local intruders may be able exploit these vulnerabilities to execute arbitrary code as the user running the ftpd daemon, usually root.
Since screen is not setuid root, this means that it leaves the ptys with insecure permissions. The updated packages restore the Unix98 pty support.
(Sure, these aren't kernel bugs, but neither are the Windows2000 problems mentioned in the linked article. To be fair, there's plenty of good reading at http://www.microsoft.com/technet/security/current
--