FreeBSD's Linux binary compatibility is not emulation. Its not even really abstraction. Linux binaries don't have to go through any extra layers in order to execute on FreeBSD than they would in order to execute on Linux.
Linux binaries can run natively on FreeBSD because the FreeBSD kernel contains all the necessary ABIs, and the filesystem contains all the necessary libraries, and there is even now a linux procfs that runs alongside FreeBSD's procfs for the benefit of Linux binaries. FreeBSD simply executes a native FreeBSD binary one way, and native Linux binaries another way. There is no performance penalty for running a Linux binary on FreeBSD because nothing has to be emulated or translated from a Linux-style execution to a FreeBSD-style execution, it is simply executed as-is.
Maybe try to understand it this way: FreeBSD can run Linux binaries not because it emulates the way Linux executes binaries, but because it has implemented the way Linux executes binaries alongside its own implementation of binary execution.
So to ask "Why should I run linux programs on FreeBSD when I could just run them on their native OS?" is a question based on misunderstanding. The simple answer is that even when they are running on FreeBSD, the are running on their native OS.
Yes, nearly everything can be keyboard controlled in Konqueror-2.1. KDE in general now is pretty good about this now. You can even customize what key combinations trigger what actions. But, just like Netscape, IE, and Lynx, tabbing through an entire page of links to get to the one you want to follow is still a pain.
I've had no stability problems whatsoever with KDE2.1 (including Konqueror). Supposedly there was a threading issue with KDE2.1 and XFree86-4.0.x on FreeBSD, which has since been worked around in the latest KDE2.1 port.
But I don't understand why we need more than one Unix.
BSD was around before Linux. So using your logic, tell me again why Linux should exist. Why is choice among different operating systems a good thing according to you, yet choice among different Unix operating systems a bad thing? Because it splits development efforts you say. So I guess choices among licenses for developers is bad according to you also? So much for your "...choice is a good thing."
[hypothetical] I don't understand Ford. No part of it. I don't understand why we need more than one american motor company. Its not like the american motor industry doesn't have enemies. Yes choice is a good thing. People can already choose between Chevrolet and Toyota. We could make Chevrolet a stronger choice, or we could make the two american car companies weaker choices. [/hypothetical]
Understand now? Probably not. At least understand this: don't dicatate to me which Unix to use or develop for, no matter for whatever noble reason. For you to do so would be no better than Microsoft dictating to me which version of Windows to use. Its the very same mentality, and to see it expressed by open-source advocates saddens me.
Actually, NONE of the software you cited as being GPL'd uses the GPL. That's right, NONE.
Perl: Artistic License
Apache: The Apache Software License
PHP: The PHP License
All three of those licenses are _very_ BSD-like, pretty much letting you do whatever you like with them so long as credit is given where credit is due. NONE of those licenses require the distribution of modified source code along with any distribution of modified binaries.
Why do you, and most everyone else, just _assume_ that whatever piece of open-source software you happen to like is covered by the GPL. Do you bother to READ the licenses of the software you use? Apparently not.
Are you going to stop using Perl now? Or Apache? Or PHP? Didn't think so. Are they suddenly going to disappear off the face of the earth because they aren't covered by the GPL? Of course not.
I can't believe no one pointed this out to you sooner...
Google's interface for web searches is _useless_ for usenet.
Couldn't they keep the existing Deja functionality until they had something decent to offer? I can't believe how completely un-sympathetic to the needs of existing Deja users this sudden, and obviously not-at-all-thought-out, gutting of Deja is on the part of Google. I like Google, put they can't just shove Deja into their existing format and structure, leaving out 90% of the previous functionality, and expect everyone to just roll with it. And from what I could tell from the FAQ they have no real plans on making it any better anytime in the relatively near future. Quoting from the FAQ...
So much for browsing newsgroups: Google does not currently support browsing of the newsgroup hierarchy. Look for more complete browsing support as well as other new features to be available in future versions of our newsgroup search service.
So much for reading new messages as they come in: The archive is currently updated once per day. There may be a delay of up to 36 hours between articles being posted to a news server and those articles being searchable within Google's newsgroup search.
The least Google could have done is gotten their shit straight _before_ pulling this half-assed stunt.
[/rant]
I guess its back to real usenet servers and clients for me. I feel sorry for those that don't have access to a real usenet server, until Google gets its act straight on this.
My own personal list of most influential games. Note that this is a personal list, i.e. games that influenced me (usually in bad ways like skipping class, not sleeping, etc.).
Tunnels of Doom (a game on the TI99/4a)
Moria (successor to Rogue)
Wing Commander
MechWarrior
Civilization
Castle Wolfenstein
Mortal Realms (a MUD)
Half-Life
Dummynet is part of FreeBSD. It does exactly what you are asking for:
dummynet is a system facility that permits the control of traffic going through the various network interfaces, by applying bandwidth and queue size limitations, and simulating delays and losses.
Noto Bene: This should in no way be considered an authoritative answer. As a matter of fact it probably contains less information than you already know.
From what I understand the biggest differences are along the lines of hardware support. BSD/OS has much better SMP support, and better drivers due to proprietary software/information provided by manufacturers which BSDi has signed NDA's for.
Server side, the CTM server runs a script which identifies changes, compresses them, puts a sequence number on them, and emails them out. Client side, ctm.rmail decodes and applies the changes.
CTM was developed for use with FreeBSD, but could be used for any source repository.
Of course they can come in without a search warrant. This isn't criminal law we are talking about, it is tort law - which is a completely different animal.
Want an example? How about MS shutting down the local government of Virginia Beach VA in order to look for license violations (link: http://www.theregister.co.uk/c ont ent/7/14496.html ).
I also am from Virginia. Yes, the UCITA legislation has passed. HoweverThere was also a resolution passed that the Joint Commission on Technology and Science would study the impact of the legislation and report back to the General Assembly by December 1, 2000. Furthermore, the UCITA legislation is not set to take effect until July 1, 2001.
It is not too late.
Write to your state senators and representatives (not your federal congressional representatives, but the ones that represent you in the state legislature - which you can find here - although it probably couldn't hurt to write to everyone in the state legislature if you are motivated to do so). Refer to the Uniform Computer Informations Transactions Act, or SB372 (from 2000 legistlative session) or CHAP996 (the actual additions to the Code of Virginia). Explain to them nicely, coherently, and concisely why UCITA is bad for Virginia, its citizens, and even its businesses. Point out how the other states have been unwilling to adopt UCITA and why.
UCITA can still be repealed in Virginia. Now is also the time to redouble our efforts towards that goal.
While I sympathize quite whole-heartedly with your sentiment, I cannot help but to point out the flaw in your solution. (I also provide my own solution at the end).
Absence of proof is not proof of absence. In this case, it would be impossible for an applicant to prove that an idea is their idea by simply failing to find proof to the contrary. The only possible "proof" to such a claim would be to explain how every idea ever in human history is not related to your idea.
This is why the system relies on challenges. It is far easier to disprove the originality of an idea by providing a counterexample.
Yes, these sound very similar to one another. The difference however, is that of proving a claim to be true versus proving something to be false.
What is broken is the means by which to perform the disproof. Currently, one must either challenge the application via the overworked and over-beaureucratic Patent Office, which simply does not have the time or the funding to care, or via the capricious and technologically-clueless court system with a civil suit, which takes a great deal of time by which point the patenet holder has strangled its competition into submission. In either case the patent continues to be enforced all during whatever time it may take to resolve the issue. It is due to this web of red tape that the applicant/holder has all the power rather than the people.
What needs to be instituted is a simpler way of challenging patents. My solution?
Remember: a single counterexample is sufficient to disprove a claim. The Patent Office needs to have the power to suspend or withdraw a patent instantly upon being furnished with a single counter-example of the originality of the idea, even after a patent has been granted. At that point, once a counter-example has been provided, it would be the burden of the applicant/holder to show that said counter-example is not related to their patent, either via the same overworked Patent Office or capricious court system. Yet in this way it is the applicant/holder which must struggle through the system, not the challenger. Thus the scales of power are balanced between the interests of business and the interests of humanity. And further, this would require no additional funding of the Patent Office.
I received a CueCat in the mail. Apparently because I am a subscriber to Wired. I did not ask for the CueCat, did not order it, did not pay for it (yes I know its free anyway). Under US Postal Regulations, this item is now mine. It is not the property of D.C., they have not loaned, lent, nor licensed it to me. They can not ask for it back, they can not tell me what to do with it. It is mine, period. If they would like to claim differently, they can take the issue up with the Post Office, not me.
Interpretation: Speaking as a mindless zealot, I wish *BSD would just go away
BSD is arguably one of the most successful Unix, especially when considering its influence. Learn some history. Or even better, just grow up. The Linux vs BSD war is over, and everybody is winning. Even mindless zealots like you.
I'm a FreeBSD user, administrator, and advocate, BUT -
Mr. Stallman's website (www.stallman.org) is hosted by idiom.com, not his own personal web server. The choice of FreeBSD was made by Idiom Communications. A good choice if you ask me, but nothing to hold agains RMS.
First - In the U.S. alone there are 850,000 IT positions that are unfilled. No CFO wants to have to overinvest in the most critical resource--people.
And then later on - "Having somebody who can screw around with my operating system would make me very, very nervous," he says.
Fine then, hire people who don't have a clue how to manipulate the operating system. (Toungue-In-Cheek): I guess that explains the proliferation of MCSE's.
To borrow a commonly used metaphor: Port scanning is akin to looking at all the windows of a house to see which ones don't have their curtains drawn. While this behavior is certainly rude, it is not inherently evil.
Much more suspicious are probes of specific ports for daemons known to have vulnerabilities. Most crackers/kiddies don't run full scans against hosts. They choose a handful of ports and check those to determine if there is something listening there and more importantly what version of that daemon is listening there. This is the behavior that is akin to checking to see if the windows are locked.
Port scanning of the first type shouldn't get any seasoned admin's hackles raised - every host connected and available is going to get scanned eventually.
Port scanning of the second type shouldn't get any seasoned admin's hackles raised either - as long as they've taken proper security measures (Mr. Cracker/Kiddie's scanner will simply log the host as "not vulnerable" and move on). Furthermore, since such probes will either be stealthed or blend in with normal traffic, it is unlikely that they will even be noticed.
What does raise my hackles is when a host gets scanned over and over and over and over within a very short period of time from the same source. Such behavior, while not a DOS attack, can be resource-intensive on the target and is very rude. But there again, it is not suspicious per se because it is most likely indicative of a certain degree of cluelessness on the part of the scanner.
The bottom line to me is that port scanning happens but it is nothing to worry about as long as proper and normal security precautions have been taken anyway beforehand and continue to be taken as exploits emerge.
The admins that complain to the source network about port scans are worried about the wrong things, or worse want someone else to be responsible for their own security.
As for liability, who knows. Common sense would dictate that A) The target is responsible for their own security, and B) The source is responsible for their own actions. But since when has common sense born any resemblance to the law, especially in the context of a civil suit?
As for the dual-platform thing, I don't think it's possible (and certainly thing it *shouldn't* be possible) to call it open-source if you're discriminating against users of a particular platform.
There's two issues here. The first is that is is certainly possible, legal, and very common to license the very same product (code, prose, movies, etc) under two or more differing licenses. The licensee is bound by whatever terms they agreed to, and not to the terms someone else may have agreed to. If I license buddhafoo_1.1 to Tom under the GPL, and to Dick under the BSDL, then Tom's use of buddhafoo_1.1 is bound by the GPL, and Dick's use is bound by the BSDL. Absolutely no conflict there.
The second issue is basically aesthetic. Both the GPL and BSDL are considered open source licenses. But if I then license buddhafoo_1.1 to Harry under a proprietary license, do you still call buddhafoo an "open source" project? Tom and Dick still have open source licenses for buddhafoo_1.1, nothing has changed for them. Some would say "yes", others like you would say "no".
The lesson here is that any project can have both open source and closed source incarnations.
Architecture: The PC architecture just has way too much overhead and too many bottlenecks, since it is designed to handle many possible utilizations. A hardware router is overall designed to do what is supposed to do and nothing else, eliminating unecessary overhead and bottlenecks. A Cisco 2500 may only be a 68030 with 8 megs of ram, but it will beat the snot out of a PC; the PC will start dropping packets long before the Cisco.
Hardware: A PC relies on too many moving parts, too many points of failure. Harddrives crash, and then you have to reinstall, restore from backups (you are going to make backups, right?). A hardware router has no moving parts except for a fan or two. Backing up a cisco's running-config is trivial, and easily transferred to a replacement router if there ever is a catastrophic hardware failure.
Data-Link Layer: Who's on the other end of the line, and what protocol do they want you to speak? What card are you going to find to drop in the PC that speaks X.25, or HDLC, or whatever? If you can find them, how much do they cost? Do you trust the manufacturer?
Network Layer: What routing protocol do you need to speak? BGP4, EGRP, IGRP, EIGRP? Forget using a PC. Even Zebra for Unix isn't mature enough, unfortunately.
Amdministration: Sure, Cisco's command-line is archaic. But it quite elegantly allows someone who knows what they are doing to do exactly what they need to do, without going through all the bullshit of a gui. There's also having to deal with all the administration responsibilities of configuring a full-blown PC hardware and OS, even for stuff in the OS you won't ever be using, as opposed to the administration of a hardware router which is, when it boils down to it, merely a matter of enabling the services you need, without ever having to touch anything you don't. Furthermore, as stated above, a cisco's entire configuration is contained in a single text file, and can be backed up and restored easily; the same is obviously not true of a PC based solution.
Upgrades: Do you want to have to administer Service Packs to your Windows2000 router? Upgrading a cisco's IOS is almost as trivial as restoring its running-config.
I hate to sound so one-sided. But the fact is that in a production environment it is professionally irresponsible to go with a solution other than that which is best suited for the task. A dedicated, hardware based router is much better suited for the task than a PC based router.
Slashdot Mirror
FreeBSD's Linux binary compatibility is not emulation. Its not even really abstraction. Linux binaries don't have to go through any extra layers in order to execute on FreeBSD than they would in order to execute on Linux.
Linux binaries can run natively on FreeBSD because the FreeBSD kernel contains all the necessary ABIs, and the filesystem contains all the necessary libraries, and there is even now a linux procfs that runs alongside FreeBSD's procfs for the benefit of Linux binaries. FreeBSD simply executes a native FreeBSD binary one way, and native Linux binaries another way. There is no performance penalty for running a Linux binary on FreeBSD because nothing has to be emulated or translated from a Linux-style execution to a FreeBSD-style execution, it is simply executed as-is.
Maybe try to understand it this way: FreeBSD can run Linux binaries not because it emulates the way Linux executes binaries, but because it has implemented the way Linux executes binaries alongside its own implementation of binary execution.
So to ask "Why should I run linux programs on FreeBSD when I could just run them on their native OS?" is a question based on misunderstanding. The simple answer is that even when they are running on FreeBSD, the are running on their native OS.
For a more technical explanation, read http://www.freebsd.org/handbook/x19213.html
Yes, nearly everything can be keyboard controlled in Konqueror-2.1. KDE in general now is pretty good about this now. You can even customize what key combinations trigger what actions. But, just like Netscape, IE, and Lynx, tabbing through an entire page of links to get to the one you want to follow is still a pain.
I've had no stability problems whatsoever with KDE2.1 (including Konqueror). Supposedly there was a threading issue with KDE2.1 and XFree86-4.0.x on FreeBSD, which has since been worked around in the latest KDE2.1 port.
"the only usable web browser available for any Unix
Try Konqueror from KDE 2.1. I think that you will be very pleasantly surprised.
But I don't understand why we need more than one Unix.
BSD was around before Linux. So using your logic, tell me again why Linux should exist. Why is choice among different operating systems a good thing according to you, yet choice among different Unix operating systems a bad thing? Because it splits development efforts you say. So I guess choices among licenses for developers is bad according to you also? So much for your "...choice is a good thing."
[hypothetical]
I don't understand Ford. No part of it. I don't understand why we need more than one american motor company. Its not like the american motor industry doesn't have enemies. Yes choice is a good thing. People can already choose between Chevrolet and Toyota. We could make Chevrolet a stronger choice, or we could make the two american car companies weaker choices.
[/hypothetical]
Understand now? Probably not. At least understand this: don't dicatate to me which Unix to use or develop for, no matter for whatever noble reason. For you to do so would be no better than Microsoft dictating to me which version of Windows to use. Its the very same mentality, and to see it expressed by open-source advocates saddens me.
Actually, NONE of the software you cited as being GPL'd uses the GPL. That's right, NONE.
- Perl: Artistic License
- Apache: The Apache Software License
- PHP: The PHP License
All three of those licenses are _very_ BSD-like, pretty much letting you do whatever you like with them so long as credit is given where credit is due. NONE of those licenses require the distribution of modified source code along with any distribution of modified binaries.Why do you, and most everyone else, just _assume_ that whatever piece of open-source software you happen to like is covered by the GPL. Do you bother to READ the licenses of the software you use? Apparently not.
Are you going to stop using Perl now? Or Apache? Or PHP? Didn't think so. Are they suddenly going to disappear off the face of the earth because they aren't covered by the GPL? Of course not.
I can't believe no one pointed this out to you sooner...
Google's interface for web searches is _useless_ for usenet.
Couldn't they keep the existing Deja functionality until they had something decent to offer? I can't believe how completely un-sympathetic to the needs of existing Deja users this sudden, and obviously not-at-all-thought-out, gutting of Deja is on the part of Google. I like Google, put they can't just shove Deja into their existing format and structure, leaving out 90% of the previous functionality, and expect everyone to just roll with it. And from what I could tell from the FAQ they have no real plans on making it any better anytime in the relatively near future. Quoting from the FAQ...
The least Google could have done is gotten their shit straight _before_ pulling this half-assed stunt.
[/rant]
I guess its back to real usenet servers and clients for me. I feel sorry for those that don't have access to a real usenet server, until Google gets its act straight on this.
My own personal list of most influential games. Note that this is a personal list, i.e. games that influenced me (usually in bad ways like skipping class, not sleeping, etc.).
Tunnels of Doom (a game on the TI99/4a)
Moria (successor to Rogue)
Wing Commander
MechWarrior
Civilization
Castle Wolfenstein
Mortal Realms (a MUD)
Half-Life
Jim said:
I'm pretty sure BSD does not.
You would be wrong.
FreeBSD
NetBSD
BSD/OS
OpenBSD does not. But don't perpetuate the falsehood that all BSD does not.
Dummynet is part of FreeBSD. It does exactly what you are asking for:
& apropos=0&sektion=0&manpath=FreeBSD+4.2-RELEASE&fo rmat=html
dummynet is a system facility that permits the control of traffic going through the various network interfaces, by applying bandwidth and queue size limitations, and simulating delays and losses.
Check out the man page:
http://www.freebsd.org/cgi/man.cgi?query=dummynet
Noto Bene: This should in no way be considered an authoritative answer. As a matter of fact it probably contains less information than you already know.
From what I understand the biggest differences are along the lines of hardware support. BSD/OS has much better SMP support, and better drivers due to proprietary software/information provided by manufacturers which BSDi has signed NDA's for.
CTM does this.
Server side, the CTM server runs a script which identifies changes, compresses them, puts a sequence number on them, and emails them out. Client side, ctm.rmail decodes and applies the changes.
CTM was developed for use with FreeBSD, but could be used for any source repository.
Of course they can come in without a search warrant. This isn't criminal law we are talking about, it is tort law - which is a completely different animal. Want an example? How about MS shutting down the local government of Virginia Beach VA in order to look for license violations (link: http://www.theregister.co.uk/c ont ent/7/14496.html ).
I also am from Virginia. Yes, the UCITA legislation has passed. HoweverThere was also a resolution passed that the Joint Commission on Technology and Science would study the impact of the legislation and report back to the General Assembly by December 1, 2000. Furthermore, the UCITA legislation is not set to take effect until July 1, 2001.
It is not too late.
Write to your state senators and representatives (not your federal congressional representatives, but the ones that represent you in the state legislature - which you can find here - although it probably couldn't hurt to write to everyone in the state legislature if you are motivated to do so). Refer to the Uniform Computer Informations Transactions Act, or SB372 (from 2000 legistlative session) or CHAP996 (the actual additions to the Code of Virginia). Explain to them nicely, coherently, and concisely why UCITA is bad for Virginia, its citizens, and even its businesses. Point out how the other states have been unwilling to adopt UCITA and why.
UCITA can still be repealed in Virginia. Now is also the time to redouble our efforts towards that goal.
From http://oss.software.i bm. com/developerworks/opensource/jfs/:
"JFS is licensed under the GNU General Public License."
While I sympathize quite whole-heartedly with your sentiment, I cannot help but to point out the flaw in your solution. (I also provide my own solution at the end).
Absence of proof is not proof of absence. In this case, it would be impossible for an applicant to prove that an idea is their idea by simply failing to find proof to the contrary. The only possible "proof" to such a claim would be to explain how every idea ever in human history is not related to your idea.
This is why the system relies on challenges. It is far easier to disprove the originality of an idea by providing a counterexample.
Yes, these sound very similar to one another. The difference however, is that of proving a claim to be true versus proving something to be false.
What is broken is the means by which to perform the disproof. Currently, one must either challenge the application via the overworked and over-beaureucratic Patent Office, which simply does not have the time or the funding to care, or via the capricious and technologically-clueless court system with a civil suit, which takes a great deal of time by which point the patenet holder has strangled its competition into submission. In either case the patent continues to be enforced all during whatever time it may take to resolve the issue. It is due to this web of red tape that the applicant/holder has all the power rather than the people.
What needs to be instituted is a simpler way of challenging patents. My solution?
Remember: a single counterexample is sufficient to disprove a claim. The Patent Office needs to have the power to suspend or withdraw a patent instantly upon being furnished with a single counter-example of the originality of the idea, even after a patent has been granted. At that point, once a counter-example has been provided, it would be the burden of the applicant/holder to show that said counter-example is not related to their patent, either via the same overworked Patent Office or capricious court system. Yet in this way it is the applicant/holder which must struggle through the system, not the challenger. Thus the scales of power are balanced between the interests of business and the interests of humanity. And further, this would require no additional funding of the Patent Office.
Sorry Mr. Tiemann, but I had working installations of both Slackware and FreeBSD before I had ever heard of Redhat.
I received a CueCat in the mail. Apparently because I am a subscriber to Wired. I did not ask for the CueCat, did not order it, did not pay for it (yes I know its free anyway). Under US Postal Regulations, this item is now mine. It is not the property of D.C., they have not loaned, lent, nor licensed it to me. They can not ask for it back, they can not tell me what to do with it. It is mine, period. If they would like to claim differently, they can take the issue up with the Post Office, not me.
In a word, no. That false sense of security is what gets a lot of machines compromised.
Security is a process, not a state.
Interpretation: Speaking as a mindless zealot, I wish *BSD would just go away
BSD is arguably one of the most successful Unix, especially when considering its influence. Learn some history. Or even better, just grow up. The Linux vs BSD war is over, and everybody is winning. Even mindless zealots like you.
I'm a FreeBSD user, administrator, and advocate, BUT -
Mr. Stallman's website (www.stallman.org) is hosted by idiom.com, not his own personal web server. The choice of FreeBSD was made by Idiom Communications. A good choice if you ask me, but nothing to hold agains RMS.
First - In the U.S. alone there are 850,000 IT positions that are unfilled. No CFO wants to have to overinvest in the most critical resource--people.
And then later on - "Having somebody who can screw around with my operating system would make me very, very nervous," he says.
Fine then, hire people who don't have a clue how to manipulate the operating system. (Toungue-In-Cheek): I guess that explains the proliferation of MCSE's.
To borrow a commonly used metaphor: Port scanning is akin to looking at all the windows of a house to see which ones don't have their curtains drawn. While this behavior is certainly rude, it is not inherently evil.
Much more suspicious are probes of specific ports for daemons known to have vulnerabilities. Most crackers/kiddies don't run full scans against hosts. They choose a handful of ports and check those to determine if there is something listening there and more importantly what version of that daemon is listening there. This is the behavior that is akin to checking to see if the windows are locked.
Port scanning of the first type shouldn't get any seasoned admin's hackles raised - every host connected and available is going to get scanned eventually.
Port scanning of the second type shouldn't get any seasoned admin's hackles raised either - as long as they've taken proper security measures (Mr. Cracker/Kiddie's scanner will simply log the host as "not vulnerable" and move on). Furthermore, since such probes will either be stealthed or blend in with normal traffic, it is unlikely that they will even be noticed.
What does raise my hackles is when a host gets scanned over and over and over and over within a very short period of time from the same source. Such behavior, while not a DOS attack, can be resource-intensive on the target and is very rude. But there again, it is not suspicious per se because it is most likely indicative of a certain degree of cluelessness on the part of the scanner.
The bottom line to me is that port scanning happens but it is nothing to worry about as long as proper and normal security precautions have been taken anyway beforehand and continue to be taken as exploits emerge.
The admins that complain to the source network about port scans are worried about the wrong things, or worse want someone else to be responsible for their own security.
As for liability, who knows. Common sense would dictate that A) The target is responsible for their own security, and B) The source is responsible for their own actions. But since when has common sense born any resemblance to the law, especially in the context of a civil suit?
looking for a document that gives a complete list of all the hardware that is supported by FreeBSD, but can I find such a document? nah.
Try reading the release notes.
As for the dual-platform thing, I don't think it's possible (and certainly thing it *shouldn't* be possible) to call it open-source if you're discriminating against users of a particular platform.
There's two issues here. The first is that is is certainly possible, legal, and very common to license the very same product (code, prose, movies, etc) under two or more differing licenses. The licensee is bound by whatever terms they agreed to, and not to the terms someone else may have agreed to. If I license buddhafoo_1.1 to Tom under the GPL, and to Dick under the BSDL, then Tom's use of buddhafoo_1.1 is bound by the GPL, and Dick's use is bound by the BSDL. Absolutely no conflict there.
The second issue is basically aesthetic. Both the GPL and BSDL are considered open source licenses. But if I then license buddhafoo_1.1 to Harry under a proprietary license, do you still call buddhafoo an "open source" project? Tom and Dick still have open source licenses for buddhafoo_1.1, nothing has changed for them. Some would say "yes", others like you would say "no".
The lesson here is that any project can have both open source and closed source incarnations.
Architecture: The PC architecture just has way too much overhead and too many bottlenecks, since it is designed to handle many possible utilizations. A hardware router is overall designed to do what is supposed to do and nothing else, eliminating unecessary overhead and bottlenecks. A Cisco 2500 may only be a 68030 with 8 megs of ram, but it will beat the snot out of a PC; the PC will start dropping packets long before the Cisco.
Hardware: A PC relies on too many moving parts, too many points of failure. Harddrives crash, and then you have to reinstall, restore from backups (you are going to make backups, right?). A hardware router has no moving parts except for a fan or two. Backing up a cisco's running-config is trivial, and easily transferred to a replacement router if there ever is a catastrophic hardware failure.
Data-Link Layer: Who's on the other end of the line, and what protocol do they want you to speak? What card are you going to find to drop in the PC that speaks X.25, or HDLC, or whatever? If you can find them, how much do they cost? Do you trust the manufacturer?
Network Layer: What routing protocol do you need to speak? BGP4, EGRP, IGRP, EIGRP? Forget using a PC. Even Zebra for Unix isn't mature enough, unfortunately.
Amdministration: Sure, Cisco's command-line is archaic. But it quite elegantly allows someone who knows what they are doing to do exactly what they need to do, without going through all the bullshit of a gui. There's also having to deal with all the administration responsibilities of configuring a full-blown PC hardware and OS, even for stuff in the OS you won't ever be using, as opposed to the administration of a hardware router which is, when it boils down to it, merely a matter of enabling the services you need, without ever having to touch anything you don't. Furthermore, as stated above, a cisco's entire configuration is contained in a single text file, and can be backed up and restored easily; the same is obviously not true of a PC based solution.
Upgrades: Do you want to have to administer Service Packs to your Windows2000 router? Upgrading a cisco's IOS is almost as trivial as restoring its running-config.
I hate to sound so one-sided. But the fact is that in a production environment it is professionally irresponsible to go with a solution other than that which is best suited for the task. A dedicated, hardware based router is much better suited for the task than a PC based router.