Here's some good information that you'er going to hate about Win2k that I've run across recently. I used to think just like you until about 3 weeks ago when everything went to shit.
1) I can find no way to directly manipulate the MTU (and no, I will not install IIS to see if it provides the ability). The MTU on the Win2k box is really small and doing anything on the net that requires a large packet barfs completely. The Win98 box has no problems doing the same exact thing.
2) Remove your PCI network card (3com here), boot the machine, and put it back in. When you get the driver selection screen, the only thing you get is a listing of ISA cards. What you have to do, is disable the device, reboot, enable the device, reboot, and then delete the device, and "scan for hardware changes". Then you will actually see the PCI cards. The drivers are nowhere to be found on the CD, and 3com does not release them becuase "they are included with the win2k distribution".
3) No 440BX AGP-PCI bridge drivers. I have suffered serious graphic performance because of this. The AGP-CPU bridge drivers are there, but not hte PCI one. Of course, like 3com, intel has no intention of releasing anything remote close to a driver or INF file.
Win2k is nice, but is has a plethora of bugs, and it's completely new (rewritten kernel based on NT4 SPEC, not code, like a cleanroom technique) which means there's even more to pop up. I would still carry a copy of Win98/NT4 on your system if I were you.
The trick is to use network storage - novell is great for this task, because hte client actually writes the profiles to the server on logout and restores them on login. (i'm sure something similar could be whacked together using samba)
This way, you can just ghost or dd the "new system" and slap in on the clients, and everything is the same as long as said client exists on each client. The user should be informed before hand that whatever they left on the client drive is GONE.
The biggest problem that I've had in similar situations is just that, because hte users were not properly trained to store everything on the server. If that's done correctly, then everything comes out fine and there are no problems, really.
The hardest part is getting the management to understand what a network drive is.. For some reaosn the regular employees seem to get it fine.:)
Just to add to this, if you were using "some other OS", even some of the other BSD/GPL systems, you wouldn't even know about these bugs until they were fixed, and probably wouldn't even know that kernel 2.4 was coming out soon at all.
The advantage is that this list is posted so that people that can write kernel code know what is on the plate so they can help if they want.
Oh yeah, that and my boss asked me what that "silly serial port thingy on the back of the UPS was", while we were repairing hte filesystem for an NT server for a bank this week.
If you know what you're doing with a current linux system you should already KNOW what works and what doesn't. So, the only research should be applied to those things that are alien to you (you'd be surprised how many NC's don't know what RAID is, but put mirroring on their NT servers all the time), or are relatively "cutting-edge" products (that is, they haven't been tested in the market for long). Personally it's better to be paranoid in this area and "go with what you know".
And in my experience I have not had a NC client ever get angry at me for doing 3 hours of research when putting together an expensive system. I charge $75/hr at the place I work now (soon to be leaving), so $225 is nothing when you're putting together a system that gets up to nearly 6 figures on occasion (including clients).
And if you have hired staff or a contracted consultant, THEY are your support. Most consultants fix their self-created mistakes and handle RMA's for free, and of course you don't have to pay extra for in house work. Personally I have had more problems with book-reading nimrods at the help desks for the software apps that I support (in some cases, the clients have paid over $10,000 for such support) than I have had trying to figure out the problem myself.
That said, the whole point of the word "professional" is that they rely on no support for their field of expertise. Unfortunately a lot of people get away with calling themselves "professionals" when they are so far from it it's disturbing.
For instance, 1 ex-employee that I worked with claimed NT, Novell and 98 experience. Worked there over a year, and I went on a call with him one day, simple modem internet connection BS - he used the Internet Setup Wizard that starts up when you first start Internut Exploder to configure the dialup networking. Then he proceeded to ask me what a DNS address and a SMTP server was.
Another ex-employee got a decetly paid job (40k/yr) hinging on his Linux experience - which totaled one client for about 3 weeks. He was convinced that the "non-contiguous" message that accompanies fsck when the filesystem is fragmented comes up when fsck "checks the disk after I power down". I asked him how he powered the machine down, and he said "I just turn it off, Linux's filesystem is so rock solid it'll come right back up". I promptly setup a BIOS password on my machine, as much as I hate to do that. 2 weeks later his machine was sitll at the office with FSCK saying "FSCK FAILED - REBOOT NOW"... Which so many other sysadmins have unfortuantely seen due to power outages.
Sorry for the book, but if you are one of these people, GO BACK AND GET YOUR HIGH SCHOOL DEGREE and POSSIBLY ATTEND COLLEGE IN SOMETHING OTHER THAN ANYTHING THAT RUNS ON ELECTRICITY.
Sorry for the rant, but I've had an interesting day. Anyone seen falling down?:)
If you took a good look at the links, they were both written in 1999, once in October and the second in December... There are at probably 10 posters to slashdot (whatever happened to sengan?:), it's not like there are 300 of them each writing their own stories, they just pick up links. This only goes to show how little they read their own site. (or make the effort to look before they post, I thought that search feature was there for a reason)
And I'm not done - all this information is kept in a RDBMS, if they don't already have a key for the target urls in the article they damn well should, this would make it 10 times easier to search for things like this, and the database itself could be told not to accept duplicate entries of that type, transparent of the perl code that undoubtedly controls the database.
So IMO, considering the fact that they are corporate backed now, (ie, they have cash and programmers to do BS like this) this should have been on the plate already. I mean, I could understand it when it was just Rob and Hemos and such, but they just wasting their bandwidth and productivity by not implementing these simple, very uncomplex things. And for those who scream backward compatibility, adding this to a universal database takes about 5-10 lines of coding using DBI under perl. Besides, then they could integrate a field into the title bar that actually references JUST the target url so those of us who actually want to read the text know right where to go everytime. That's just good UI design.
That said, yes, I would agree that remembering an article from 1997 is rather absurd, and I would flame something like this if it didn't happen so damned often.
It's important to understand this, that someone will have to pay the programmers. In the past, it was the hardware houses who were hiring programmers to write specialized applications for their hardware.
Nowadays it ranges from places like redhat, who pay their programmers with the money they reap off of support, to microsoft (we obviously know how they pay their programmers), to places like IBM and some of hte more obscure UNIX vendors (Data General for instance) who are still doing it the old way.
IMO, the old way is the best way to do it, but that almost directly collides with open source. However, if systems were open enough to where the programmers working at these software houses were actually writing software to release as source, this would workout almost perfectly, and we would get the benefits of specialized hardware (such as the S/390 which was a topic of discussion this week) and the benefit of being able to modify the source code.
I don't see this happening in the future, but it does allow flexing of the idealistic part of the brain, eh?
Anyone who receieved a Linux Journal in the last few days will notice there is a HUGE article about the S/390 and the linux port done by both IBM and hte Open Source community.
It also goes to mention that in the IBM port, there are a couple of core level modules (i'm not much with a S/390 so not sure what they are) that are object code only, IOW, not open source.
I have no problem with this other than the fact that Debian, which is mentioned as having a port in the works (the writer of the article is a core debian maintainer), prides itself in being a "Free Software Only" distribution. I'd really like to hear some comments on why that is or should be any different for this case.
The big advantage from what I can tell is not even runnign linux as the core OS, but running it under the very powerful VM mechanisms in the S/390 (the article explains how the VM is actually tied into the HARDWARE, which just plain rules) allowed the writer of the full free software kernel port (which is not finished, IBM did a private port and announced it this month) to START OVER *41,000* COPIES OF LINUX + THE APACHE WEBSERVER! Good LORD!
So don't kid yourself, with native power like that, no one is going to even bother running linux standalone on one of these things, not to mention there is much cheaper, adequate hardware that will run linux by itself - a S/390 is a very high end machine and chances are it's going to be a cold day in hell before the suits let you throw your "Free OS" directly on the machine.
But that doesn't matter, from the article I gathered it's going to be much more popular in a VM.:)
Sorry for the bluntness, but this guy needs to wake up and realize he's not in the 1950's anymore...
This is the same kind of propoganda that was spreading in the late 1950's, when computing with electronic devices actually became a viable option for the future... People started talking about automation, and then the paranoid McCarthy-ist types started talking about how robots were going to take over the world, etc, blah, etc, blah.
Then came lost in space and buck rogers, and well, people laughed at the idea.
I find it hard to believe that this man contains any level of technical aptitude higher than starting the copy of word that he used to write this "e-book" (doesn't anyone call them text files anymore?), considering the observations he makes.
And considering that AI, after... decades of work is just finally simulating a 2 year old with no ability to learn at a normal human pace, well, I'm willing to bet that me, being in my mid twenties is probably going to have grandchildren that are dead before this is even a possiblity.
In other words, this makes a convincing movie, but not a convincing world. Break out the poodle skirts.
But however, it certainly doens't hurt. I'd be willing to bet that their admins have their shit together.
The point of "no security by obscurity" is that you don't want to waste all your time setting up firewalls that do nothing but block ports - you want logging, and (real) filtering, you want to make sure the firewall is not crackable itself.
Not disclosing your internal network setup to the world is a good idea if you ask me.... It's another thing to move telnet to port 666 and ftp to 28 and then, if they can find the webserver at port 8086 they can download a custom services file listing all of their port moves so they can actually use the server. Now that is obscurity.
However for the stroke of this dicussion it's good that CmdrTaco posted it... Altbough you'd never find any backbone providers talking like that.:)
now, my libc5 programs seem to wrok fine on my debian system, when the core system lib is glibc 2.1...
However on my slackware 4 system... I have to use libc5. Heaven forbid I might have to compile source. Even better, the system just might not have the tool that I need to fix the situation, because, you know UNIX isn't known for having stuff like that. Right? And heaven forbid I couldn't compile a new libc.
IIRC, the ELF binary format hasn't changed in a long time, something that was created by SUN MICROSYSTEMS, IIRC.
Now the libc's have changed, but there's nothing keeping you from creating statically linked binaries (Which 99% of windows binaries are) that you can use in just about any system. A simple "ldd binary" will also tell you what you need for the system, should you be curious.
Anyone who wants to make an installer can easily write this stuff into their program/script for use.
LSB is pointless as long as we have working package formats that are well supported, and alien. I don't think that most people realize how similar the setup is on most Linux distributions, comapared to, oh, Data General Unix or SCO or XENIX... Coming from a Linux background, these systems are so backassed it's not even funny. (Why the hell is shutdown in/etc?)
However, just like 99% of all Linux distro's, these machines share a good portion of the paths and locations of binaries.
SCO, however, with it's wonderful forethought, refuses to include which or locate. (both of which are on a DG/UX system, along with many other GNU utilities or clones)
Just to make my point clear. differing libc versions do not equal an unstable system. Nothing, i mean NOTHING is keeping these vendors from entering the arena, especially if they know how to create static libraries (which are compiled as default under gcc), and link them to their binary.
ELF is the binary standard. IBCS (IIRC) is the binary standard for IPC for X86 SVR4 binaries. The last time anything but ELF was used was back in kernel 1.2.something, which was about....6 years ago. Not to mention you can compile a module for linux that still uses the old a.out format. Get a clue.
As for the LSB, if the smart people stick to/usr/local or/opt, then I would have a hell of an easier time adminstering systems. Last I checked both of these exist on most systems. (I just symlink/opt to/usr/local if it's not there:)
HE DID WORK IT OUT ALREADY, as he stated in his article, he signed a release to let Be use it in the OS despite the restrictions placed by the GPL.
As copyright holder, he is allowed to do that. Remember, the GPL is copyleft, with the intentional meaning that it is an addition to copyright, not a replacement.
You know, I see this is as a good thing that Be and Bruce worked things out so quickly and decently, and that there was no legal BS that was needed.
After all, isn't this what we're always complaining about on here? The legal system getting in our way? Is it TOO hard to recognize the fact that some people are above having to resort to such petty measures, and for christ's sake, figure it out themselves?
You be the judge, but I think the fact that Bruce saw a useful application for his program in another operating system, had no resentment towards the company (who obviously had made a mistake considering the way that library is constructed), and even allowed them to continue using the code.
This is the way that things SHOULD be done, IMO. The posting was simply a description of what was going on.... After all, what if Bruce didn't catch this, and some.... "news service" got ahold of it first and raised a big ruckus *AHEM* before anything could get resolved? Things may have turned out differently.
Just a side note: MS DOS did nothing to do anything for or to DR-DOS... It's not "technically" possible to have both DOS's running at the same time.
However, WFW 3.11 (I believe that's the version) would not RUN under DR-DOS, reporting an error with the DOS version, hwoever, version 5 of MS-DOS still ran fine under it (in other words, nothing had changed to facilitate this).
Half of the problem with credibility in some of these writings here is the rumors that they heard from john doe in sanitional engineering that they use to make points in the posts.
Just a tip for some of you, you can have the most credible message/story/thesis but if you blow it by throwing in a bent nail to secure the coffin all you have is 6 pieces of shitty wood.
You know, I figured that CEO's with millions (billions?) of dollars wouldn't fall into something so petty as peer pressure (or use it as a cheap excuse)....
Well it's good to see you've had your 12 cups today.:)
The bitterness could certiainly take a breather, but here's some criticism on your comments.
The net has changed for everyone universally despite what you think. Advertisements are the biggest change that I can think of at the moment, even the smallest places you go to nowadays are serving banner ads that profile and exploit personal information.
The internet has both enhanced and degraded the privacy of individuals... It's important to understand that concept. Unless you stick to the local area network, you're giving it away in one way or another.
There's also something to be said about arguing that your opinion is right and another's is wrong, especially when that's what you're arguing about.:)
Devin thought he was a 3133t h4X0R, and frequented the BBS's I ran... he said he was writing a virus that would align all the lasers in your monitor to produce enough power to melt through the screen, and consequently, through your head.
Now, I know this is bullshit, but seeing this got me thinking... Foregoing that a virus was possible to do this, I'm just curious, would a concentrated mass of monitor lasers produce enough heat to even burn THROUGH the screen in the first place? Maybe some of you EE geeks can help me out on this, I'm very curious.:)
Slashdot Mirror
Here's some good information that you'er going to hate about Win2k that I've run across recently. I used to think just like you until about 3 weeks ago when everything went to shit.
1) I can find no way to directly manipulate the MTU (and no, I will not install IIS to see if it provides the ability). The MTU on the Win2k box is really small and doing anything on the net that requires a large packet barfs completely. The Win98 box has no problems doing the same exact thing.
2) Remove your PCI network card (3com here), boot the machine, and put it back in. When you get the driver selection screen, the only thing you get is a listing of ISA cards. What you have to do, is disable the device, reboot, enable the device, reboot, and then delete the device, and "scan for hardware changes". Then you will actually see the PCI cards. The drivers are nowhere to be found on the CD, and 3com does not release them becuase "they are included with the win2k distribution".
3) No 440BX AGP-PCI bridge drivers. I have suffered serious graphic performance because of this. The AGP-CPU bridge drivers are there, but not hte PCI one. Of course, like 3com, intel has no intention of releasing anything remote close to a driver or INF file.
Win2k is nice, but is has a plethora of bugs, and it's completely new (rewritten kernel based on NT4 SPEC, not code, like a cleanroom technique) which means there's even more to pop up. I would still carry a copy of Win98/NT4 on your system if I were you.
-Erik-
The trick is to use network storage - novell is great for this task, because hte client actually writes the profiles to the server on logout and restores them on login. (i'm sure something similar could be whacked together using samba)
:)
This way, you can just ghost or dd the "new system" and slap in on the clients, and everything is the same as long as said client exists on each client. The user should be informed before hand that whatever they left on the client drive is GONE.
The biggest problem that I've had in similar situations is just that, because hte users were not properly trained to store everything on the server. If that's done correctly, then everything comes out fine and there are no problems, really.
The hardest part is getting the management to understand what a network drive is.. For some reaosn the regular employees seem to get it fine.
Just to add to this, if you were using "some other OS", even some of the other BSD/GPL systems, you wouldn't even know about these bugs until they were fixed, and probably wouldn't even know that kernel 2.4 was coming out soon at all.
The advantage is that this list is posted so that people that can write kernel code know what is on the plate so they can help if they want.
-Erik-
Oh yeah, that and my boss asked me what that "silly serial port thingy on the back of the UPS was", while we were repairing hte filesystem for an NT server for a bank this week.
I have a wonderful job.
-Erik-
Err.. Research?
:)
If you know what you're doing with a current linux system you should already KNOW what works and what doesn't. So, the only research should be applied to those things that are alien to you (you'd be surprised how many NC's don't know what RAID is, but put mirroring on their NT servers all the time), or are relatively "cutting-edge" products (that is, they haven't been tested in the market for long). Personally it's better to be paranoid in this area and "go with what you know".
And in my experience I have not had a NC client ever get angry at me for doing 3 hours of research when putting together an expensive system. I charge $75/hr at the place I work now (soon to be leaving), so $225 is nothing when you're putting together a system that gets up to nearly 6 figures on occasion (including clients).
And if you have hired staff or a contracted consultant, THEY are your support. Most consultants fix their self-created mistakes and handle RMA's for free, and of course you don't have to pay extra for in house work. Personally I have had more problems with book-reading nimrods at the help desks for the software apps that I support (in some cases, the clients have paid over $10,000 for such support) than I have had trying to figure out the problem myself.
That said, the whole point of the word "professional" is that they rely on no support for their field of expertise. Unfortunately a lot of people get away with calling themselves "professionals" when they are so far from it it's disturbing.
For instance, 1 ex-employee that I worked with claimed NT, Novell and 98 experience. Worked there over a year, and I went on a call with him one day, simple modem internet connection BS - he used the Internet Setup Wizard that starts up when you first start Internut Exploder to configure the dialup networking. Then he proceeded to ask me what a DNS address and a SMTP server was.
Another ex-employee got a decetly paid job (40k/yr) hinging on his Linux experience - which totaled one client for about 3 weeks. He was convinced that the "non-contiguous" message that accompanies fsck when the filesystem is fragmented comes up when fsck "checks the disk after I power down". I asked him how he powered the machine down, and he said "I just turn it off, Linux's filesystem is so rock solid it'll come right back up". I promptly setup a BIOS password on my machine, as much as I hate to do that. 2 weeks later his machine was sitll at the office with FSCK saying "FSCK FAILED - REBOOT NOW"... Which so many other sysadmins have unfortuantely seen due to power outages.
Sorry for the book, but if you are one of these people, GO BACK AND GET YOUR HIGH SCHOOL DEGREE and POSSIBLY ATTEND COLLEGE IN SOMETHING OTHER THAN ANYTHING THAT RUNS ON ELECTRICITY.
Sorry for the rant, but I've had an interesting day. Anyone seen falling down?
-Erik, AKA D-Fens-
Hehe, I meant that I WOULDN'T flame something like this if it didn't happen so often. :)
-Erik-
If you took a good look at the links, they were both written in 1999, once in October and the second in December... There are at probably 10 posters to slashdot (whatever happened to sengan? :), it's not like there are 300 of them each writing their own stories, they just pick up links. This only goes to show how little they read their own site. (or make the effort to look before they post, I thought that search feature was there for a reason)
:)
And I'm not done - all this information is kept in a RDBMS, if they don't already have a key for the target urls in the article they damn well should, this would make it 10 times easier to search for things like this, and the database itself could be told not to accept duplicate entries of that type, transparent of the perl code that undoubtedly controls the database.
So IMO, considering the fact that they are corporate backed now, (ie, they have cash and programmers to do BS like this) this should have been on the plate already. I mean, I could understand it when it was just Rob and Hemos and such, but they just wasting their bandwidth and productivity by not implementing these simple, very uncomplex things. And for those who scream backward compatibility, adding this to a universal database takes about 5-10 lines of coding using DBI under perl. Besides, then they could integrate a field into the title bar that actually references JUST the target url so those of us who actually want to read the text know right where to go everytime. That's just good UI design.
That said, yes, I would agree that remembering an article from 1997 is rather absurd, and I would flame something like this if it didn't happen so damned often.
Ahem... Andover need a perl coder?
-Erik-
It's important to understand this, that someone will have to pay the programmers. In the past, it was the hardware houses who were hiring programmers to write specialized applications for their hardware.
Nowadays it ranges from places like redhat, who pay their programmers with the money they reap off of support, to microsoft (we obviously know how they pay their programmers), to places like IBM and some of hte more obscure UNIX vendors (Data General for instance) who are still doing it the old way.
IMO, the old way is the best way to do it, but that almost directly collides with open source. However, if systems were open enough to where the programmers working at these software houses were actually writing software to release as source, this would workout almost perfectly, and we would get the benefits of specialized hardware (such as the S/390 which was a topic of discussion this week) and the benefit of being able to modify the source code.
I don't see this happening in the future, but it does allow flexing of the idealistic part of the brain, eh?
-Erik-
Anyone who receieved a Linux Journal in the last few days will notice there is a HUGE article about the S/390 and the linux port done by both IBM and hte Open Source community.
:)
It also goes to mention that in the IBM port, there are a couple of core level modules (i'm not much with a S/390 so not sure what they are) that are object code only, IOW, not open source.
I have no problem with this other than the fact that Debian, which is mentioned as having a port in the works (the writer of the article is a core debian maintainer), prides itself in being a "Free Software Only" distribution. I'd really like to hear some comments on why that is or should be any different for this case.
The big advantage from what I can tell is not even runnign linux as the core OS, but running it under the very powerful VM mechanisms in the S/390 (the article explains how the VM is actually tied into the HARDWARE, which just plain rules) allowed the writer of the full free software kernel port (which is not finished, IBM did a private port and announced it this month) to START OVER *41,000* COPIES OF LINUX + THE APACHE WEBSERVER! Good LORD!
So don't kid yourself, with native power like that, no one is going to even bother running linux standalone on one of these things, not to mention there is much cheaper, adequate hardware that will run linux by itself - a S/390 is a very high end machine and chances are it's going to be a cold day in hell before the suits let you throw your "Free OS" directly on the machine.
But that doesn't matter, from the article I gathered it's going to be much more popular in a VM.
-Erik-
Yeah, God IS a complicated and intelligent being... :)
:)
j/k, please no emails from the cult awareness network thank you
-Erik-
Sorry for the bluntness, but this guy needs to wake up and realize he's not in the 1950's anymore...
... decades of work is just finally simulating a 2 year old with no ability to learn at a normal human pace, well, I'm willing to bet that me, being in my mid twenties is probably going to have grandchildren that are dead before this is even a possiblity.
This is the same kind of propoganda that was spreading in the late 1950's, when computing with electronic devices actually became a viable option for the future... People started talking about automation, and then the paranoid McCarthy-ist types started talking about how robots were going to take over the world, etc, blah, etc, blah.
Then came lost in space and buck rogers, and well, people laughed at the idea.
I find it hard to believe that this man contains any level of technical aptitude higher than starting the copy of word that he used to write this "e-book" (doesn't anyone call them text files anymore?), considering the observations he makes.
And considering that AI, after
In other words, this makes a convincing movie, but not a convincing world. Break out the poodle skirts.
-Erik-
But however, it certainly doens't hurt. I'd be willing to bet that their admins have their shit together.
:)
The point of "no security by obscurity" is that you don't want to waste all your time setting up firewalls that do nothing but block ports - you want logging, and (real) filtering, you want to make sure the firewall is not crackable itself.
Not disclosing your internal network setup to the world is a good idea if you ask me.... It's another thing to move telnet to port 666 and ftp to 28 and then, if they can find the webserver at port 8086 they can download a custom services file listing all of their port moves so they can actually use the server. Now that is obscurity.
However for the stroke of this dicussion it's good that CmdrTaco posted it... Altbough you'd never find any backbone providers talking like that.
-Erik-
You want to know what's sad?
This couldn't be more true. I know two people that have NIKE tatoos on varying parts of their bodies. (I live in portland, home of evil NIKE)
If this doesn't validate what Blue says, I don't know what does.
-Erik-
Try running SCO binaries on a XENIX system. :)
... I have to use libc5. Heaven forbid I might have to compile source. Even better, the system just might not have the tool that I need to fix the situation, because, you know UNIX isn't known for having stuff like that. Right? And heaven forbid I couldn't compile a new libc.
/etc?)
... .6 years ago. Not to mention you can compile a module for linux that still uses the old a.out format. Get a clue.
/usr/local or /opt, then I would have a hell of an easier time adminstering systems. Last I checked both of these exist on most systems. (I just symlink /opt to /usr/local if it's not there :)
They are the same OS, after all, right?
now, my libc5 programs seem to wrok fine on my debian system, when the core system lib is glibc 2.1...
However on my slackware 4 system
IIRC, the ELF binary format hasn't changed in a long time, something that was created by SUN MICROSYSTEMS, IIRC.
Now the libc's have changed, but there's nothing keeping you from creating statically linked binaries (Which 99% of windows binaries are) that you can use in just about any system. A simple "ldd binary" will also tell you what you need for the system, should you be curious.
Anyone who wants to make an installer can easily write this stuff into their program/script for use.
LSB is pointless as long as we have working package formats that are well supported, and alien. I don't think that most people realize how similar the setup is on most Linux distributions, comapared to, oh, Data General Unix or SCO or XENIX... Coming from a Linux background, these systems are so backassed it's not even funny. (Why the hell is shutdown in
However, just like 99% of all Linux distro's, these machines share a good portion of the paths and locations of binaries.
SCO, however, with it's wonderful forethought, refuses to include which or locate. (both of which are on a DG/UX system, along with many other GNU utilities or clones)
Just to make my point clear. differing libc versions do not equal an unstable system. Nothing, i mean NOTHING is keeping these vendors from entering the arena, especially if they know how to create static libraries (which are compiled as default under gcc), and link them to their binary.
ELF is the binary standard. IBCS (IIRC) is the binary standard for IPC for X86 SVR4 binaries. The last time anything but ELF was used was back in kernel 1.2.something, which was about
As for the LSB, if the smart people stick to
-Erik-
Not to mention all that paper and ink would cost them a LOT more.... Not to mention wear and tear on the printer that was used to copy it....
:)
Imagine the SAMBA source tree being printed out and sent... CDROM would be much easier
-Erik-
This is what I think a lot of you are missing....
HE DID WORK IT OUT ALREADY, as he stated in his article, he signed a release to let Be use it in the OS despite the restrictions placed by the GPL.
As copyright holder, he is allowed to do that. Remember, the GPL is copyleft, with the intentional meaning that it is an addition to copyright, not a replacement.
You know, I see this is as a good thing that Be and Bruce worked things out so quickly and decently, and that there was no legal BS that was needed.
After all, isn't this what we're always complaining about on here? The legal system getting in our way? Is it TOO hard to recognize the fact that some people are above having to resort to such petty measures, and for christ's sake, figure it out themselves?
You be the judge, but I think the fact that Bruce saw a useful application for his program in another operating system, had no resentment towards the company (who obviously had made a mistake considering the way that library is constructed), and even allowed them to continue using the code.
This is the way that things SHOULD be done, IMO. The posting was simply a description of what was going on.... After all, what if Bruce didn't catch this, and some.... "news service" got ahold of it first and raised a big ruckus *AHEM* before anything could get resolved? Things may have turned out differently.
-Erik-
Yeah... now if I could only apropos the answer to life....
-Erik-
Perhaps they have done a good job, and it's just wonderful windows doing what it does best :)
Erik
Just a side note: MS DOS did nothing to do anything for or to DR-DOS... It's not "technically" possible to have both DOS's running at the same time.
However, WFW 3.11 (I believe that's the version) would not RUN under DR-DOS, reporting an error with the DOS version, hwoever, version 5 of MS-DOS still ran fine under it (in other words, nothing had changed to facilitate this).
Half of the problem with credibility in some of these writings here is the rumors that they heard from john doe in sanitional engineering that they use to make points in the posts.
Just a tip for some of you, you can have the most credible message/story/thesis but if you blow it by throwing in a bent nail to secure the coffin all you have is 6 pieces of shitty wood.
It'll be an antique...
:)
just like standard oil signs.
You know, I figured that CEO's with millions (billions?) of dollars wouldn't fall into something so petty as peer pressure (or use it as a cheap excuse)....
:)
I guess I was wrong.
-Erik-
Well it's good to see you've had your 12 cups today. :)
:)
The bitterness could certiainly take a breather, but here's some criticism on your comments.
The net has changed for everyone universally despite what you think. Advertisements are the biggest change that I can think of at the moment, even the smallest places you go to nowadays are serving banner ads that profile and exploit personal information.
The internet has both enhanced and degraded the privacy of individuals... It's important to understand that concept. Unless you stick to the local area network, you're giving it away in one way or another.
There's also something to be said about arguing that your opinion is right and another's is wrong, especially when that's what you're arguing about.
mmm.... philsophy in the morning.. tasty
Long ago, there was a man named Devin.
:)
Devin thought he was a 3133t h4X0R, and frequented the BBS's I ran... he said he was writing a virus that would align all the lasers in your monitor to produce enough power to melt through the screen, and consequently, through your head.
Now, I know this is bullshit, but seeing this got me thinking... Foregoing that a virus was possible to do this, I'm just curious, would a concentrated mass of monitor lasers produce enough heat to even burn THROUGH the screen in the first place? Maybe some of you EE geeks can help me out on this, I'm very curious.
-Erik-
AFAIK Linus does not have to make any statements of what he feels is "Fair Use".
Those laws are already in place, the fact that he has a trademark is enough to contend any usage he feels is not kosher.
-Erik-
Big deal, it's understandable that Apple wants to protect their trademarks.
:)
Just slap a Kalista Apple on there instead
Erik