Linux for System/390 can run native, without a hypervisor. Most of us mainframe geeks think there are serious advantages to running it under VM (the hypervisor), but if you happen to find a box by the side of the road and don't have the bucks for software, Linux will run on it.
You need to be up on very recent IBM stuff to recognize it, but when Red Hat says the entire "eServer" line, that means Netfinity, AS/400, RS/6000 and System/390. They all got renamed into one line of four families last week.
SuSE's already available (in beta), TurboLinux is commited, and there's a Debian port started. If yau want some Slack (Praise Bob!), ya gotta get it started!
And don't forget The Stone That Never Came Down either. First published in 1973, it discusses a designer drug that effectively eliminates ignorance and self-delusion in favor, and either saves or ruins humanity in the process (your call depending on your politics).
Brunner predicts an Internet-like network, which I think was quite an accomplishment, even if he didn't see the microcomputer revolution coming.
Huh? When Nicky first meets Kate, she's redoing her apartment, replacing the functional computer circuitry painted on the wall of her living room the previous year. How microcomputer revolution can you get?
The only difference is, instead of a PC on every desk, there is a terminal or phone connected to a mainframe somewhere, which is in turn linked to the worldwide data network.
The Pretender isn't so much a Brunner-ism as it is a riff on Roger Zelazny's My Name Is Legion. The lead character in MNIL is part of the team that builds the world wide data bank, and chooses to take his boss up on the offer to destroy his personal punched card deck (OK, so it's a little dated!) and become a non-person. After that, he gets lots of jobs that can best be done by an invisible man.
There's nothing wrong with making archival copies of information you can come by legitimately. Alexa isn't redistributing the information, so they're fine. The did, however, at one point promise to give a complete copy to Kahle's non-profit Internet Archive. My understanding at the time was that the Internet Archive is legally structured to look much more like a library than a business, and libraries have a special status under copyright law.
If you want an example of true Bazaar development, you need look no further than Unix itself. Unix, not Linux. "Return with us now, to those fabulous days of yesteryear!"...
The August 23rd Slashdot has a pointer to a wonderful Unix family tree in graphical form. As we all remember, Ken Thompson created Unix, and over time it leaked out from Bell Labs to assorted universities. It also flourished inside AT&T, producing quite a number of variants, each with it's own goals. Once UCBerkeley's Computer Science Research Group got going, we had the *ix equivalent of the Cambrian Explosion. Quite apart from the 2BSD vs. 4BSD vs. AT&T variants, every hardware manufacturer on the face of the planet started carving on 4.2BSD, all with the intent of "differentiating themselves from the rest of the market".
The end result? Just watch a GNU configure script run, or (shudder) install an older sendmail (circa v5). (Hey - remember when there were lots of sendmail variants running around too?) "Unix" doesn't mean a darn thing in terms of functionality, interfaces, etc.. There is, of course, a "Unix flavor", and that's why we all recognize Minix, Linux, et al. as Unix-oid systems. But it took repeated attempts like AT&T's SVID, IEEE's POSIX, and others to simply pare down the list to an almost-managable size, and even then the real factor was that the market couldn't support all those hardware variations, so their software variants vanished with them.
The Bazaar was bizzare for Unix, and it nearly killed it. The community sorted itself out after a time, and grudging annointed Scott McNealy and Solaris the winners, but still leaving a few others in play. In the mean time, a lot of effort was wasted and a lot of money was spent that could have fed the homeless or done something equally worthwhile.
Charles Connell is right, even if he does mistake the meanings of the Cathedral and Bazaar models for the shape of their org-charts. Fred Brooks was right too - if you haven't read TM3 lately, you owe it yourself and to your fellow human beings to let the Prof. lecture you once again.
Slightly off-topic, but the link above to Transparent Society triggered Opera's third-party-cookie detector, thus allowing me to defeat the web-bug placed in the article page.
Actually, yes. Give credit where credit is due - the IBMers in Boerblingen did an awful lot of work on their own. We know because there was a simultaneous open, user-driven attempt at getting Linux up on S/390 (the "i370" port, championed by Linas Vepstas) that some of us believe helped coax the S/390 port out of IBM's lab. Alan Cox and others have been extremely helpful to the crowd, and the kernel changes have certainly been integrated since 2.2.14, but if you read the LINUX-VM list archives it's clear that the IBM guys did a hell of a lot of work. There's no reason to belittle or discount their efforts.
From what I understand, IBM didn't even consider supporting TCP/IP until about ten years ago or so
Nope, not true. IBM S/390 systems have had TCP/IP available since the early 1980s when Dr. Larry Landweber's group out at U. Wisconsin wrote the precursor to IBM's current VM stack, "WISCNET" as the core for the CSNET network. To this day, IBM's "TCP/IP for VM/ESA" stack supports 14.0.0.1 as a loopback address because CSNET was net 14.
In fact, you can probably run the linux VMs under MVS
Nope, MVS doesn't virtualize the underliying hardware. One of the other S/390-native OSes, VM/ESA ("VM" as in "Virtual Machine") does, and that's what many (most?) of us are running Linux under on S/390 hardware currently. It's just like what VMWare does for x86 hardware, in fact the precursor to VMWare was called "VM/386" when the precursor to VM/ESA was called "VM/370".
That fact isn't made clear in the announcement and AFAIK that is not a requirement to run Linux on a S390 server
Damn straight. Lots of us have been running Linux on S/390 for some time now, both on the metal and in virtual machines. Check out the archives of the LINUX-VM list for the war stories and glories. This is the group that was reported on here several months ago, the news is IBM's big-time commitment to the idea instead of just skunk-works'ing it.
The other problem is that Linux does not support any 'native' S390 FS so you have implement some extensions to the OS in order to overlay a Linux FS on top of the basic FS functions on the S390
Nope, Linux on S/390 uses EXT2 just like everybody else does. Two of the S/390 native OSes (OS/390 (aka MVS) and VM/ESA) already have NFS servers, so if you want access to the "native" filesystems, just mount them.
Back before Network Solutions took over the Internet, and Jon Postel et al. were in charge of numbers, the official registration for net 127.0.0 belonged to the University of Mars. Alas, today it's just the property of IANA.
Ah, but does Prodigy control what your From: header reads?
I don't know if it does, but it certainly should. There's a long-standing tradition in multi-user computing systems that the local system ensures mail originating from a user is tagged as having done so. RFC 822 explicitly states this (section 4.4.1), and provides a mechanism (the "Sender:" header) for cases when this isn't true.
Does your phone company control what caller ID reports for your phone number?
It sure does!
The From: header is set by the person making the post.
Perhaps, but it isn't supposed to be. Prodigy should (and I expect does - remember, it started as Sears and IBM, not a high school student in his parent's basement) ensure that the "From" header accurately reflects the author's userid. The case at hand doesn't imply any failure to do so. On the contrary, the suit (as quoted in the Wired article cited way south of here) apparently claims Prodigy is liable for their failure to adequately protect Lunney from imposture. In other words, Prodigy aided and abetted in the commission of a crime. While I'm glad the Supremes found no reason to review the case and therefore allowed Prodigy's common-carrier status to stand (at least in New York), I really think they dropped the ball in also allowing Prodigy to get away with their role in the imposture.
Yes, I suppose it is possible for the news/mail server to replace it with what their records say, but it would rely on some form of authenticating the user, and many mail clients don't support authenticating SMTP servers.
I haven't seen any report that suggests the e-mail in question was totally forged (i.e. that it was sent around Prodigy, not through it). As long as Prodigy was in the loop, it had an obligation to do exactly as you say.
What displays on caller ID is controlled by the phone company, with the exception that you can block it completly when making a call.
Not to mention the wonderful Caller Id Block Block! Now there's a great concept to carry over to the Internet. Imagine the ability to require your ISP to keep mail from unauthenticated senders out of your inbox - most spam would vanish in a puff of logic.
I'm not familiar with the case, but the article cited in the Slashdot headline makes this appear to be a loss in freedom-protection, not a win. The article briefly explains that someone posing as Alexander Lunney sent threatening e-mail from Prodigy, and that Prodigy (and by implication all ISPs) is not liable for the content of that e-mail. That's great, but it seems that Prodigy should have been liable to Lunney for their participation in the imposture.
The content-liability issue rests firmly on telecommuncations common-carrier law, but aren't the carriers still responsible for the accuracy of the telephone number that your caller-ID box displays? What if I start making threatening phone calls to the White House after convincing Bell Atlantic to give me a phone line that's listed in your name? When the Secret Service breaks down your door, will this ruling still look like a win?
The topic at hand is TCP/IP on VM/ESA, not on OS/390. Because VM's stack is entirely in userspace (i.e. running in it's own virtual machine), there has never been any limit on the number of stacks you could run on a single system, and they didn't need to have any interaction with each other. One cool use has been to run two stacks, one connected to the outside world, the other to the inside world, and to run the VM system spanning the firewall boundary. Unless you cross-connect the stacks, there's no network liability.
Ross Patterson Sterling Software (um, make that Computer Associates)
Sorry, but that's not the truth. VM had TCP/IP support back in the early 1980s, and at heart it's the same solid code running today. The implementation was originally done for use by CSNET by Dr. Larry Landweber's group out at U. Wisconsin, and was known in those days as "WISCNET". (Yes, *THAT* Landweber.) In fact, it still recognizes the CSNET loopback address (14.0.0.1)!
Ground0 would seem to be mistaking OS/390 for VM. See jms's excellant summary post for the key differences - it's like saying Linux and Windows are the same 'cause they both run on Intel hardware.
Ross Patterson Sterling Software (um, make that Computer Associates)
Slashdot Mirror
Linux for System/390 can run native, without a hypervisor. Most of us mainframe geeks think there are serious advantages to running it under VM (the hypervisor), but if you happen to find a box by the side of the road and don't have the bucks for software, Linux will run on it.
You need to be up on very recent IBM stuff to recognize it, but when Red Hat says the entire "eServer" line, that means Netfinity, AS/400, RS/6000 and System/390. They all got renamed into one line of four families last week.
There's an s390 Debian port under way by the user community (of course, it's Debian ferchrissake!). It just isn't coming from IBM.
SuSE's already available (in beta), TurboLinux is commited, and there's a Debian port started. If yau want some Slack (Praise Bob!), ya gotta get it started!
SuSE has already delivered beta code for IBM's System/390 mainframes, and TurboLinux is commited to doing so as well.
And don't forget The Stone That Never Came Down either. First published in 1973, it discusses a designer drug that effectively eliminates ignorance and self-delusion in favor, and either saves or ruins humanity in the process (your call depending on your politics).
Brunner predicts an Internet-like network, which I think was quite an accomplishment, even if he didn't see the microcomputer revolution coming.
Huh? When Nicky first meets Kate, she's redoing her apartment, replacing the functional computer circuitry painted on the wall of her living room the previous year. How microcomputer revolution can you get?
The only difference is, instead of a PC on every desk, there is a terminal or phone connected to a mainframe somewhere, which is in turn linked to the worldwide data network.
Ah, so they're WAP phones!
The Pretender isn't so much a Brunner-ism as it is a riff on Roger Zelazny's My Name Is Legion. The lead character in MNIL is part of the team that builds the world wide data bank, and chooses to take his boss up on the offer to destroy his personal punched card deck (OK, so it's a little dated!) and become a non-person. After that, he gets lots of jobs that can best be done by an invisible man.
There's nothing wrong with making archival copies of information you can come by legitimately. Alexa isn't redistributing the information, so they're fine. The did, however, at one point promise to give a complete copy to Kahle's non-profit Internet Archive. My understanding at the time was that the Internet Archive is legally structured to look much more like a library than a business, and libraries have a special status under copyright law.
If you want an example of true Bazaar development, you need look no further than Unix itself. Unix, not Linux. "Return with us now, to those fabulous days of yesteryear!" ...
The August 23rd Slashdot has a pointer to a wonderful Unix family tree in graphical form. As we all remember, Ken Thompson created Unix, and over time it leaked out from Bell Labs to assorted universities. It also flourished inside AT&T, producing quite a number of variants, each with it's own goals. Once UCBerkeley's Computer Science Research Group got going, we had the *ix equivalent of the Cambrian Explosion. Quite apart from the 2BSD vs. 4BSD vs. AT&T variants, every hardware manufacturer on the face of the planet started carving on 4.2BSD, all with the intent of "differentiating themselves from the rest of the market".
The end result? Just watch a GNU configure script run, or (shudder) install an older sendmail (circa v5). (Hey - remember when there were lots of sendmail variants running around too?) "Unix" doesn't mean a darn thing in terms of functionality, interfaces, etc.. There is, of course, a "Unix flavor", and that's why we all recognize Minix, Linux, et al. as Unix-oid systems. But it took repeated attempts like AT&T's SVID, IEEE's POSIX, and others to simply pare down the list to an almost-managable size, and even then the real factor was that the market couldn't support all those hardware variations, so their software variants vanished with them.
The Bazaar was bizzare for Unix, and it nearly killed it. The community sorted itself out after a time, and grudging annointed Scott McNealy and Solaris the winners, but still leaving a few others in play. In the mean time, a lot of effort was wasted and a lot of money was spent that could have fed the homeless or done something equally worthwhile.
Charles Connell is right, even if he does mistake the meanings of the Cathedral and Bazaar models for the shape of their org-charts. Fred Brooks was right too - if you haven't read TM3 lately, you owe it yourself and to your fellow human beings to let the Prof. lecture you once again.
Slightly off-topic, but the link above to Transparent Society triggered Opera's third-party-cookie detector, thus allowing me to defeat the web-bug placed in the article page.
Actually, yes. Give credit where credit is due - the IBMers in Boerblingen did an awful lot of work on their own. We know because there was a simultaneous open, user-driven attempt at getting Linux up on S/390 (the "i370" port, championed by Linas Vepstas) that some of us believe helped coax the S/390 port out of IBM's lab. Alan Cox and others have been extremely helpful to the crowd, and the kernel changes have certainly been integrated since 2.2.14, but if you read the LINUX-VM list archives it's clear that the IBM guys did a hell of a lot of work. There's no reason to belittle or discount their efforts.
IRON PENGUIN FOREVER!!!!
http://linux.s390.org/
Rock on! Check out http://www.linuxvm.org too.
Nope, not true. IBM S/390 systems have had TCP/IP available since the early 1980s when Dr. Larry Landweber's group out at U. Wisconsin wrote the precursor to IBM's current VM stack, "WISCNET" as the core for the CSNET network. To this day, IBM's "TCP/IP for VM/ESA" stack supports 14.0.0.1 as a loopback address because CSNET was net 14.
Nope, MVS doesn't virtualize the underliying hardware. One of the other S/390-native OSes, VM/ESA ("VM" as in "Virtual Machine") does, and that's what many (most?) of us are running Linux under on S/390 hardware currently. It's just like what VMWare does for x86 hardware, in fact the precursor to VMWare was called "VM/386" when the precursor to VM/ESA was called "VM/370".
Damn straight. Lots of us have been running Linux on S/390 for some time now, both on the metal and in virtual machines. Check out the archives of the LINUX-VM list for the war stories and glories. This is the group that was reported on here several months ago, the news is IBM's big-time commitment to the idea instead of just skunk-works'ing it.
The other problem is that Linux does not support any 'native' S390 FS so you have implement some extensions to the OS in order to overlay a Linux FS on top of the basic FS functions on the S390
Nope, Linux on S/390 uses EXT2 just like everybody else does. Two of the S/390 native OSes (OS/390 (aka MVS) and VM/ESA) already have NFS servers, so if you want access to the "native" filesystems, just mount them.
Back before Network Solutions took over the Internet, and Jon Postel et al. were in charge of numbers, the official registration for net 127.0.0 belonged to the University of Mars. Alas, today it's just the property of IANA.
I don't know if it does, but it certainly should. There's a long-standing tradition in multi-user computing systems that the local system ensures mail originating from a user is tagged as having done so. RFC 822 explicitly states this (section 4.4.1), and provides a mechanism (the "Sender:" header) for cases when this isn't true.
Does your phone company control what caller ID reports for your phone number?
It sure does!
The From: header is set by the person making the post.
Perhaps, but it isn't supposed to be. Prodigy should (and I expect does - remember, it started as Sears and IBM, not a high school student in his parent's basement) ensure that the "From" header accurately reflects the author's userid. The case at hand doesn't imply any failure to do so. On the contrary, the suit (as quoted in the Wired article cited way south of here) apparently claims Prodigy is liable for their failure to adequately protect Lunney from imposture. In other words, Prodigy aided and abetted in the commission of a crime. While I'm glad the Supremes found no reason to review the case and therefore allowed Prodigy's common-carrier status to stand (at least in New York), I really think they dropped the ball in also allowing Prodigy to get away with their role in the imposture.
Yes, I suppose it is possible for the news/mail server to replace it with what their records say, but it would rely on some form of authenticating the user, and many mail clients don't support authenticating SMTP servers.
I haven't seen any report that suggests the e-mail in question was totally forged (i.e. that it was sent around Prodigy, not through it). As long as Prodigy was in the loop, it had an obligation to do exactly as you say.
What displays on caller ID is controlled by the phone company, with the exception that you can block it completly when making a call.
Not to mention the wonderful Caller Id Block Block! Now there's a great concept to carry over to the Internet. Imagine the ability to require your ISP to keep mail from unauthenticated senders out of your inbox - most spam would vanish in a puff of logic.
The content-liability issue rests firmly on telecommuncations common-carrier law, but aren't the carriers still responsible for the accuracy of the telephone number that your caller-ID box displays? What if I start making threatening phone calls to the White House after convincing Bell Atlantic to give me a phone line that's listed in your name? When the Secret Service breaks down your door, will this ruling still look like a win?
Ross Patterson
Sterling Software (um, make that Computer Associates)
Sorry, but that's not the truth. VM had TCP/IP support back in the early 1980s, and at heart it's the same solid code running today. The implementation was originally done for use by CSNET by Dr. Larry Landweber's group out at U. Wisconsin, and was known in those days as "WISCNET". (Yes, *THAT* Landweber.) In fact, it still recognizes the CSNET loopback address (14.0.0.1)!
Ground0 would seem to be mistaking OS/390 for VM. See jms's excellant summary post for the key differences - it's like saying Linux and Windows are the same 'cause they both run on Intel hardware.
Ross Patterson
Sterling Software (um, make that Computer Associates)