postfix's convoluted incoming vs outgoing filtering
It strikes me as more than a little bit ironic to call ANYTHING convoluted in comparison to sendmail.
And to turn that around, it strikes me as more than a little bit odd that postfix manages to be MORE complex than sendmail in this respect. I mean, it's not like rule-sets are a joy to behold, but postfix manages to make it even more of a pain, to the extent that I've heard seasoned postfix admins on mailing lists say, "well, that's not exactly doable out-of-the box, why don't you just set up an incoming and and outgoing server?..."
I like the *idea* of postfix, but it seems that the filtering system makes it rather a difficult migration away from sendmail.
This is exactly the problem with the OpenBSD, qmail (and the rest of DJB's software) and any other system that claims security through simplicity, but then refuses to either add features or accept code changes for the feature set that is needed in the real world. I respect this software, as I respect all functioning software that is contributed to the community (though qmail is contributed with some heavy provisos on what you are allowed to do in terms of modification and distribution).
However, you get the "unsupported majority" who run a modified/patched/extended version that might well have security flaws that no one knows about. Worse, when an exploit is found in one of those changes, the maintainer of the central package usually makes a point of saying, "look, see! My software was secure, it was just those icky add-ons that were broken!" (as OpenBSD did with apache).
Bottom line: if you run OpenBSD or qmail or any other like service, don't patch it, or add unsupported features.
If that's not a good enough feature-set for you, choose a platform that embraces the feature-set that you need.
Now, on to the myths of sendmail:
Recent sendmail holes have been found because careful security auditing by programmers who have no goal other than to find such problems is being PAID for on sendmail. Companies like Red Hat have found such bugs in the Linux kernel, sendmail, apache, samba, etc, etc because they are looking for them, fixing them, and patching their user-base proactively.
I'm not saying that this is a first. Many companies that can afford it perform such audits, and it's still not as helpful, IMHO, as the benefit of being open source in the first place. However, saying that software is "insecure" because paid auditors have discovered and fixed the problems is... questionable.
I like sendmail. It has its quirks and problems, but I've yet to see a replacement that doesn't insist on proving that it's "better than sendmail" by imposing some strange restriction on the users (e.g. exim's B&D approach to RFC-compliance; postfix's convoluted incoming vs outgoing filtering; qmail's B&D approach to software distribution).
I like these other packages too, but I don't see a role for them as-is in my environments. Perhaps someday someone will write a simple sendmail replacement that is feature-for-feature compatible, but simply has simpler code and a more straight-forward config syntax (the only two real failings of sendmail).
This is the only thing that works. I have NO pity for folks who also have hosting from these scumbags who are collateral damage. Find a new hosting or colo company or feel the wrath.
And it's exactly that attitude that devalues your blacklist with software like SpamAssassin (since it actually looks at the real-world impact to a user who *wants* to get their legit mail of each blacklist's performance history). Your blacklist's opinion will still be noted, but unless you carefully prune your list on the basis of minimizing collateral damage, you will be out-voted by other, more fastidious blacklists.
Postmaster errors, bounces, virus scanners... they all tell me that my email address is being used as the "from" address by dozens of helpless Windows users' virus-ridden computers. Sigh. I'm deleting about 20-50 messages a minute, and just barely keeping up.
The standard sysadmin reward-food is "anything flat". For programmers you can also add (for those from MIT) Suan La Chow Show. If you don't know what the Suan is, you have not yet lived. Go directly to Mary's, do not type go, do not clock 200MhZ.
I'm up for offering you a free Mary's run. Just drop me some mail.
One of the most controversial changes in GNOME 2.4 is the dumping of Galeon in favour of Epiphany as the default browser
For those of you who do not know, Epiphany is a new project from the creator of Galeon. It's aimed at staying small and light-weight while Galeon continues to be a sort of "Mozilla - XUL - mailer - chat + Gnome".
Epiphany is based on Mozilla, but is much more light and bloat free and features a much, much, much cleaner interface. I have not used Galeon very extensively, but Epiphany has already become my default browser. Startup is much faster than Mozilla, the interface is much more intuitive and clutter free and it merges nicely with the look and feel of the rest of the GNOME environment.
Also for those who do not know, all of the above is true of Galeon as well. The choice to go with Epiphany is probably mostly based on a) it's newness and thus percieved coolness b) the ultra-lightweight nature of it lends to component-model desktop integration more easily. The Galeon or Epiphany bits are almost certainly not the major contributors to user-precieved performance, so I don't think you'll see as big a difference between Galeon and Epiphany as you did between, say, Mozilla and Galeon... there simply isn't that much in terms of the UI left to rip out when rendering a page compared to the overhead of parsing and rendering HTML/JavaScript/CSS/who-knows-what.
The thing that worries me is that Galeon used to be hands-down the best browser for a Gnome desktop, but now it has been ripped apart in terms of its code base and in terms of the developers. The Gnome 2.0 revision of Galeon was a major step backward in terms of user experience. Things changed like menus (e.g. "View Source" moving out of the "View" menu and into the "File" menu) becoming much less usable; major functionality present in the JavaScript configs no longer available in the preferences; and much more.
At the same time, the developers left to go work on a new toy while users of the now live version of the browser suffered.
I'm actually on the edge of going back to Mozilla, which while large and often cumbersome, still has some of the best maintentance-mode support going (we'll see if that lasts post-Netscape).
probability of their investigating the SCO fiasco is directly related to total amount of the the non-Geek news coverage
Please cite a major news outlet that has not already covered this. I just read a great piece in the Wall Street Journal (print) and have seen stories on every Web site that carries a Reuters feed (that includes CNN, MSNBC, etc).
If the SEC or FTC starts looking into it, that's when you'll see the prime-time coverage of the sort "Tonight on Local News 1, TiVo owners are being backed by the government as they investigate the claims of a Utah software company who says TiVo's "Linux" software is theirs, and users must pay them to use it! All that and the weather after these messages!"
You may think this is geek news, but it stopped being that around when SCO started threatening joe-average who bought a high-tech toy at Circut City. If you normalize your SCO-boneheaded-move meter so that only one big mistake shows up, that would be the one.
Just like all the executives at Enron and Worldcom went to jail
That was very, very different. First off, it's not clear that either of those companies is as much to blame for what happened as the banks that build them out of a constant stream of loans backed by their stock which the banks' analysts were propping up.
Second, that wasn't a case where the company was going directly to millions of consumers and saying "you owe us money for stolen property" (as they have done against TiVo owners) when a) there was no stolen property, just a dispute over licensing rights b) the dispute was far from settled, and wasn't likely to go their way c) the claims were repeatedly answered with a request for details which was denied. About the only part that's similar is the executives are using it as a chance to dump stock for profit.
In the first case, you have a company that's inflating stock using the market against itself to make money.
In the second you have a company that's inflating stock using fraudulent scare-tactics against millions of consumers to make money.
I think SCO is far, far, far more deserving of some major crack-downs from the regulatory bodies.
So, it's official: SCO's complaint is entirely about Sequent, and their contributions to UNIX and Linux.
For those who don't know, Sequent was a super-computer company that developed a lot of the software techniques that make today's multi-processor machines (especially the more-than-2 processor systems) work well. Sequent's code was its own, and IBM bought Sequent, so at first glance IBM has every right to contribute this code to Linux, even though it has also been sold to USL/SCO.
So why is SCO suing? Because they feel that this code was written "for UNIX" and thus cannot be contributed to Linux without carrying a "taint" of UNIX IP.
SCO has made many claims about "stolen UNIX code", but that's a sham as we now see. What SCO is really upset about is that code that they think is encumbered by them, but not actually theirs was inappropriately licensed without compensation to SCO. It's not UNIX code, it's allegedly-UNIX-encumbered code, but that doesn't sound as good in a press release demanding $32 from every TiVo user....
In fact it sounds suspiciously like a case where SCO will have to fight an up-hill battle over IP rights that aren't really clear to begin with, and even then they have to fight the GPL issue, which (given that they still offer a Linux kernel for download) will be difficult.
SCO is going to burn over this, and I really do think that the owners and all C-level executives of the company are going to end up in jail. It's just too brazen a lie and scam to go unchecked. The FTC and SEC will at least have to launch a token investigation to demonstrate that they're not 100% asleep at the wheel...
The story was based on an article in a peer-reviewed journal. If you have a problem with the methods, you should be complaining to the journal, not the reporters, since the reporters, in theory, know much less about the science than those reviewing for the journal.
And it is important to not that Blaschko's Lines may or may not have any relation to Chimerism. They are generally caused by a birth defect that results from a very early cell-division. That is to say, only some percentage of the cells in the body get the defect, the rest are unaffected. Pigmentation striping (Blaschko's Lines) is an obvious result of this condition, which is called Mosaicism.
cd/usr/src/redhat/SRPMS # or whatever your %{_topdir}/SRPMS is wget ftp://ftp.sco.com/pub/updates/OpenLinux/3.1.1/Serv er/CSSA-2003-020.0/SRPMS/linux -2.4.13-21S.src.rpm rpm -i linux -2.4.13-21S.src.rpm cd/usr/src/redhat/SOURCES ls
You will see your new, shiny Linux kernel tar-ball as well as several SCO patches against it, which are of course distributed under the GPL as well (as they state in the spec file, I'm not assuming that on my own).
Yeah, I don't like that licensing scheme. I like their other one much better. You see, it's this new thing called the GPL. I know you're all skeptical, but it's really a cool license, and the price is much better. SCO may try to hold you to those terms and make you give out modifications that you make to anyone that you give binaries to, though so be careful!;-)
Conspiracy is a legal term, you're interpreting it in a casual sense.
In legal terms, a conspiracy is when two or more parties are involved in a "plan".
I think their angle here is that Red Hat is distributing SCO's code [their claim] and that code is being used by other developers in the community, and fed back to Red Hat to support their business. Thus, the whole community is conspiring to add value to Red Hat based on code that Red Hat should not have been using [again, SCO's claim, not mine].
If that's the angle they're using, then I think it's sound. But the problem is that it's based on the simple assumption that SCO can win their first and most basic point. That is, a company that has and continues to distribute code under the GPL can then sue another company for using it in compliance with the GPL. I can't imagine a case in which that's defensible.
Re:Check out Internet Mail 2000
on
Replacing SMTP?
·
· Score: 1
Trusted protocols abound. SMTP, for example, has a protocol within it called TLS. This allows key exchange, and once you've done key exchange, and are talking over a secure connection, the world is your oyster. You can give each identity a trust rating, or you can only alow mail from identities you've manually addded or you can just use a whitelist of valid identities.
It's not the trusted protocol that's hard, it's the infrastructure AROUND that protocol. You need a tool that a) allows (perhaps degraded) communication with non-conformant hosts b) everyone has an incentive to use and c) makes senders of valid mail more trusted over time.
There's a thing called "litigation risk". The theory is this (based on statistics, and as stated to me by a lawyer):
On a clear blue day, with all of the facts in your favor, and with no chance of losing the case, going to court means you have about a 20% chance of losing.
So, to turn that around, SCO probably has about a 20.01% chance of WINNING this case.
Of course, that won't happen, since the moment it gets close, IBM would sue them for infringing about 2000 patents that IBM holds in reserve just for such celebratory occasions; buy them cheap; and sell their HQ for parking space.
Even if they did win, though, they've lost in terms of Linux. A win against IBM is just that: a win against IBM. To press a case against a USER of Linux without having pressed a case against the distributors first would place them in such a messy situation, that the litigation risk might well be moot in the face of the counter-suits that they would be slapped with for using the legal system for purposes of implementing a protection racket!
Also, and all of the rest asside, go to their FTP server and download yourself a copy of the Linux source code, licensed under the GPL. There, you're done. If you're super-paranoid, don't use any version of Linux after 2.4.13 (the version they distribute) until this case is settled.
not trying to spread FUD.... but ill wait for a tighter distro before i promote *nix on the desktop
Tighter in what respect?
Try installing and managing a Red Hat 9 box. I've got several friends that I do admin for as well as working at a software firm that uses about 80-90% linux on the desktop. I find Linux to be a dream to manage compared to Windows, Mac even other free OSen which have great software, but often poorly supported ports of software that is developed under Linux (and complain as they may about developers who don't write prortable code, as an admin, I have to choose the desktop that works out of the box, and has decent support).
Id also be thankfull not to be asked how to make packages work correctly between KDE, gnome, X, or whatever else joe moron decides to use
Allowing joe moron to decide what to use is your first mistake. Here's how you deploy Linux:
Grab 3-5 of your "power users" (preferably a spectrum from techie to non-techie tinkerer) and give them a second desktop with a full install (plus whatever external goodies you find most useful, check freshrpms for example) for a month. Let them go nuts, give them root and let them install/run whatever they like
After that trial period get them all in a room and get them to agree on what they found to be most useful, and define a standard install
Deploy that on their second desktops, and this time, don't allow them to change anything and don't give them root
Repeat the process a month later, and refine the install, this time install about 2-4 times the number of users
During the third trial, start developing your admin infrastructure in earnest. You'll need to decide how you're going to remotely update these boxes (I suggest apt for rpm out of a cron job, but up2date is fine too); what data if any users will be allowed to store locally; etc.
Begin your full-scale deployment.
It's hard work. Any conversion to a new OS is, but I think you'll find that not worrying about viruses alone makes it worth it.
In 1988, I could ask a question on a USENET newsgroup, and as long as the question was relatively mainstream among the particular crowd that I was asking (e.g. in the C programming group, if I asked about a pointer management problem), I would get several answers withing a day.
Today, when I ask very straight-forward questions on USENET groups that are focused on the specific issue that I'm asking about, I get nothing.
Why? Because no one READS USENET any more. Sure, there are a hand-full of die-hards, but compared to the number of people per group that were reading back in the late 80s and early 90s, USENET is essentially an un-monitored bit-bucket.
MS Research is not MS Software. They are wildly different, just as Bell Labs and AT&T were different, and AT&T never would have developed UNIX and given it to Berkeley. MS Research, for example, releases many pieces of software under open source licensing including the GPL.
The research arm of any large technology company is not really designed to "discover" new things (sometimes they do, but that's not their purpose). They are designed to take the ideas that others have had and find ways to make them more practical so that the engineering and production side of the company can eventually make a product out of it, and the sales side can then sell it. MS Research does this particularly well, and while they produce many things that aren't "product-ready", the number of technologies that they produce (e.g the language that was Java-ized into C#) that are later used in products is quite high.
Innovation is not a quality of using new parts to build something never before seen, but to find new ways to use the parts you have. The former is called invention, not innovation.
Slashdot Mirror
postfix's convoluted incoming vs outgoing filtering
It strikes me as more than a little bit ironic to call ANYTHING convoluted in comparison to sendmail.
And to turn that around, it strikes me as more than a little bit odd that postfix manages to be MORE complex than sendmail in this respect. I mean, it's not like rule-sets are a joy to behold, but postfix manages to make it even more of a pain, to the extent that I've heard seasoned postfix admins on mailing lists say, "well, that's not exactly doable out-of-the box, why don't you just set up an incoming and and outgoing server?..."
I like the *idea* of postfix, but it seems that the filtering system makes it rather a difficult migration away from sendmail.
This is exactly the problem with the OpenBSD, qmail (and the rest of DJB's software) and any other system that claims security through simplicity, but then refuses to either add features or accept code changes for the feature set that is needed in the real world. I respect this software, as I respect all functioning software that is contributed to the community (though qmail is contributed with some heavy provisos on what you are allowed to do in terms of modification and distribution).
However, you get the "unsupported majority" who run a modified/patched/extended version that might well have security flaws that no one knows about. Worse, when an exploit is found in one of those changes, the maintainer of the central package usually makes a point of saying, "look, see! My software was secure, it was just those icky add-ons that were broken!" (as OpenBSD did with apache).
Bottom line: if you run OpenBSD or qmail or any other like service, don't patch it, or add unsupported features.
If that's not a good enough feature-set for you, choose a platform that embraces the feature-set that you need.
Now, on to the myths of sendmail:
Recent sendmail holes have been found because careful security auditing by programmers who have no goal other than to find such problems is being PAID for on sendmail. Companies like Red Hat have found such bugs in the Linux kernel, sendmail, apache, samba, etc, etc because they are looking for them, fixing them, and patching their user-base proactively.
I'm not saying that this is a first. Many companies that can afford it perform such audits, and it's still not as helpful, IMHO, as the benefit of being open source in the first place. However, saying that software is "insecure" because paid auditors have discovered and fixed the problems is... questionable.
I like sendmail. It has its quirks and problems, but I've yet to see a replacement that doesn't insist on proving that it's "better than sendmail" by imposing some strange restriction on the users (e.g. exim's B&D approach to RFC-compliance; postfix's convoluted incoming vs outgoing filtering; qmail's B&D approach to software distribution).
I like these other packages too, but I don't see a role for them as-is in my environments. Perhaps someday someone will write a simple sendmail replacement that is feature-for-feature compatible, but simply has simpler code and a more straight-forward config syntax (the only two real failings of sendmail).
This is the only thing that works. I have NO pity for folks who also have hosting from these scumbags who are collateral damage. Find a new hosting or colo company or feel the wrath.
And it's exactly that attitude that devalues your blacklist with software like SpamAssassin (since it actually looks at the real-world impact to a user who *wants* to get their legit mail of each blacklist's performance history). Your blacklist's opinion will still be noted, but unless you carefully prune your list on the basis of minimizing collateral damage, you will be out-voted by other, more fastidious blacklists.
Join the club?! Heck, I liked it so much, I bought the company!
;-)
The crappiest INBOX you've ever had or your money back!
Postmaster errors, bounces, virus scanners... they all tell me that my email address is being used as the "from" address by dozens of helpless Windows users' virus-ridden computers. Sigh. I'm deleting about 20-50 messages a minute, and just barely keeping up.
The standard sysadmin reward-food is "anything flat". For programmers you can also add (for those from MIT) Suan La Chow Show. If you don't know what the Suan is, you have not yet lived. Go directly to Mary's, do not type go, do not clock 200MhZ.
I'm up for offering you a free Mary's run. Just drop me some mail.
Someone please mod the parent way the heck up. Good research is what Slashdot's mod system should encourage!
Yeah, I should help out. Problem is I have a full time job and two major hobbies... sigh.
One of the most controversial changes in GNOME 2.4 is the dumping of Galeon in favour of Epiphany as the default browser
For those of you who do not know, Epiphany is a new project from the creator of Galeon. It's aimed at staying small and light-weight while Galeon continues to be a sort of "Mozilla - XUL - mailer - chat + Gnome".
Epiphany is based on Mozilla, but is much more light and bloat free and features a much, much, much cleaner interface. I have not used Galeon very extensively, but Epiphany has already become my default browser. Startup is much faster than Mozilla, the interface is much more intuitive and clutter free and it merges nicely with the look and feel of the rest of the GNOME environment.
Also for those who do not know, all of the above is true of Galeon as well. The choice to go with Epiphany is probably mostly based on a) it's newness and thus percieved coolness b) the ultra-lightweight nature of it lends to component-model desktop integration more easily. The Galeon or Epiphany bits are almost certainly not the major contributors to user-precieved performance, so I don't think you'll see as big a difference between Galeon and Epiphany as you did between, say, Mozilla and Galeon... there simply isn't that much in terms of the UI left to rip out when rendering a page compared to the overhead of parsing and rendering HTML/JavaScript/CSS/who-knows-what.
The thing that worries me is that Galeon used to be hands-down the best browser for a Gnome desktop, but now it has been ripped apart in terms of its code base and in terms of the developers. The Gnome 2.0 revision of Galeon was a major step backward in terms of user experience. Things changed like menus (e.g. "View Source" moving out of the "View" menu and into the "File" menu) becoming much less usable; major functionality present in the JavaScript configs no longer available in the preferences; and much more.
At the same time, the developers left to go work on a new toy while users of the now live version of the browser suffered.
I'm actually on the edge of going back to Mozilla, which while large and often cumbersome, still has some of the best maintentance-mode support going (we'll see if that lasts post-Netscape).
probability of their investigating the SCO fiasco is directly related to total amount of the the non-Geek news coverage
Please cite a major news outlet that has not already covered this. I just read a great piece in the Wall Street Journal (print) and have seen stories on every Web site that carries a Reuters feed (that includes CNN, MSNBC, etc).
If the SEC or FTC starts looking into it, that's when you'll see the prime-time coverage of the sort "Tonight on Local News 1, TiVo owners are being backed by the government as they investigate the claims of a Utah software company who says TiVo's "Linux" software is theirs, and users must pay them to use it! All that and the weather after these messages!"
You may think this is geek news, but it stopped being that around when SCO started threatening joe-average who bought a high-tech toy at Circut City. If you normalize your SCO-boneheaded-move meter so that only one big mistake shows up, that would be the one.
Just like all the executives at Enron and Worldcom went to jail
That was very, very different. First off, it's not clear that either of those companies is as much to blame for what happened as the banks that build them out of a constant stream of loans backed by their stock which the banks' analysts were propping up.
Second, that wasn't a case where the company was going directly to millions of consumers and saying "you owe us money for stolen property" (as they have done against TiVo owners) when a) there was no stolen property, just a dispute over licensing rights b) the dispute was far from settled, and wasn't likely to go their way c) the claims were repeatedly answered with a request for details which was denied. About the only part that's similar is the executives are using it as a chance to dump stock for profit.
In the first case, you have a company that's inflating stock using the market against itself to make money.
In the second you have a company that's inflating stock using fraudulent scare-tactics against millions of consumers to make money.
I think SCO is far, far, far more deserving of some major crack-downs from the regulatory bodies.
So, it's official: SCO's complaint is entirely about Sequent, and their contributions to UNIX and Linux.
For those who don't know, Sequent was a super-computer company that developed a lot of the software techniques that make today's multi-processor machines (especially the more-than-2 processor systems) work well. Sequent's code was its own, and IBM bought Sequent, so at first glance IBM has every right to contribute this code to Linux, even though it has also been sold to USL/SCO.
So why is SCO suing? Because they feel that this code was written "for UNIX" and thus cannot be contributed to Linux without carrying a "taint" of UNIX IP.
SCO has made many claims about "stolen UNIX code", but that's a sham as we now see. What SCO is really upset about is that code that they think is encumbered by them, but not actually theirs was inappropriately licensed without compensation to SCO. It's not UNIX code, it's allegedly-UNIX-encumbered code, but that doesn't sound as good in a press release demanding $32 from every TiVo user....
In fact it sounds suspiciously like a case where SCO will have to fight an up-hill battle over IP rights that aren't really clear to begin with, and even then they have to fight the GPL issue, which (given that they still offer a Linux kernel for download) will be difficult.
SCO is going to burn over this, and I really do think that the owners and all C-level executives of the company are going to end up in jail. It's just too brazen a lie and scam to go unchecked. The FTC and SEC will at least have to launch a token investigation to demonstrate that they're not 100% asleep at the wheel...
The story was based on an article in a peer-reviewed journal. If you have a problem with the methods, you should be complaining to the journal, not the reporters, since the reporters, in theory, know much less about the science than those reviewing for the journal.
Hmm, that would be good. A whole new set of negative mods: "Evil", "Unenlightened", "Blasphemy", "Offensive", "Obscene", "VMS-related".
;-)
Could be a big hit
That discussion has moved to K5:
7
http://www.kuro5hin.org/story/2003/8/6/32819/5182
Enjoy.
And it is important to not that Blaschko's Lines may or may not have any relation to Chimerism. They are generally caused by a birth defect that results from a very early cell-division. That is to say, only some percentage of the cells in the body get the defect, the rest are unaffected. Pigmentation striping (Blaschko's Lines) is an obvious result of this condition, which is called Mosaicism.
Use and enjoy!
with 1 CPU $699
;-)
Yeah, I don't like that licensing scheme. I like their other one much better. You see, it's this new thing called the GPL. I know you're all skeptical, but it's really a cool license, and the price is much better. SCO may try to hold you to those terms and make you give out modifications that you make to anyone that you give binaries to, though so be careful!
Conspiracy is a legal term, you're interpreting it in a casual sense.
In legal terms, a conspiracy is when two or more parties are involved in a "plan".
I think their angle here is that Red Hat is distributing SCO's code [their claim] and that code is being used by other developers in the community, and fed back to Red Hat to support their business. Thus, the whole community is conspiring to add value to Red Hat based on code that Red Hat should not have been using [again, SCO's claim, not mine].
If that's the angle they're using, then I think it's sound. But the problem is that it's based on the simple assumption that SCO can win their first and most basic point. That is, a company that has and continues to distribute code under the GPL can then sue another company for using it in compliance with the GPL. I can't imagine a case in which that's defensible.
Trusted protocols abound. SMTP, for example, has a protocol within it called TLS. This allows key exchange, and once you've done key exchange, and are talking over a secure connection, the world is your oyster. You can give each identity a trust rating, or you can only alow mail from identities you've manually addded or you can just use a whitelist of valid identities.
It's not the trusted protocol that's hard, it's the infrastructure AROUND that protocol. You need a tool that a) allows (perhaps degraded) communication with non-conformant hosts b) everyone has an incentive to use and c) makes senders of valid mail more trusted over time.
I seem to have an oddly missing link there. That was supposed to be "go to their FTP server..."
There's a thing called "litigation risk". The theory is this (based on statistics, and as stated to me by a lawyer):
On a clear blue day, with all of the facts in your favor, and with no chance of losing the case, going to court means you have about a 20% chance of losing.
So, to turn that around, SCO probably has about a 20.01% chance of WINNING this case.
Of course, that won't happen, since the moment it gets close, IBM would sue them for infringing about 2000 patents that IBM holds in reserve just for such celebratory occasions; buy them cheap; and sell their HQ for parking space.
Even if they did win, though, they've lost in terms of Linux. A win against IBM is just that: a win against IBM. To press a case against a USER of Linux without having pressed a case against the distributors first would place them in such a messy situation, that the litigation risk might well be moot in the face of the counter-suits that they would be slapped with for using the legal system for purposes of implementing a protection racket!
Also, and all of the rest asside, go to their FTP server and download yourself a copy of the Linux source code, licensed under the GPL. There, you're done. If you're super-paranoid, don't use any version of Linux after 2.4.13 (the version they distribute) until this case is settled.
Tighter in what respect?
Try installing and managing a Red Hat 9 box. I've got several friends that I do admin for as well as working at a software firm that uses about 80-90% linux on the desktop. I find Linux to be a dream to manage compared to Windows, Mac even other free OSen which have great software, but often poorly supported ports of software that is developed under Linux (and complain as they may about developers who don't write prortable code, as an admin, I have to choose the desktop that works out of the box, and has decent support).
Id also be thankfull not to be asked how to make packages work correctly between KDE, gnome, X, or whatever else joe moron decides to use
Allowing joe moron to decide what to use is your first mistake. Here's how you deploy Linux:
- Grab 3-5 of your "power users" (preferably a spectrum from techie to non-techie tinkerer) and give them a second desktop with a full install (plus whatever external goodies you find most useful, check freshrpms for example) for a month. Let them go nuts, give them root and let them install/run whatever they like
- After that trial period get them all in a room and get them to agree on what they found to be most useful, and define a standard install
- Deploy that on their second desktops, and this time, don't allow them to change anything and don't give them root
- Repeat the process a month later, and refine the install, this time install about 2-4 times the number of users
- During the third trial, start developing your admin infrastructure in earnest. You'll need to decide how you're going to remotely update these boxes (I suggest apt for rpm out of a cron job, but up2date is fine too); what data if any users will be allowed to store locally; etc.
- Begin your full-scale deployment.
It's hard work. Any conversion to a new OS is, but I think you'll find that not worrying about viruses alone makes it worth it.In 1988, I could ask a question on a USENET newsgroup, and as long as the question was relatively mainstream among the particular crowd that I was asking (e.g. in the C programming group, if I asked about a pointer management problem), I would get several answers withing a day.
Today, when I ask very straight-forward questions on USENET groups that are focused on the specific issue that I'm asking about, I get nothing.
Why? Because no one READS USENET any more. Sure, there are a hand-full of die-hards, but compared to the number of people per group that were reading back in the late 80s and early 90s, USENET is essentially an un-monitored bit-bucket.