"If someone connects to your web server, who authorised the access? In this case it's reasonable to assume that a web server that's open to the public network provides implicitly authorised access... or is it?"
Leaving a port open does not by itself constititute authorization.
If you publish a www.example.com record in the DNS, it would be entirely reasonable to assume that you've authorized people to stop by your website, yes. There are other conventions which apply, such as having a/robots.txt file, or a W3C privacy policy, or "Terms of use", etc.
Note that this does not mean that launching a set of requests to exploit, say, a PHP vulnerability, would be OK just because the machine is listed in the DNS. Legitimate users would have a very different click-trail than someone trying to break in.
Agreed, mostly. I've seen some portable music players make a point of noting their volume-limitation features in years past, so some vendors have obviously run into this issue before.
If an iPod plays louder than the local law permits, that's a clear and easily tested matter of fact, and even a clever lawyer is going to have a hard time spinning that on the defense.
On the other hand, if someone tells you about a problem, and you promptly release new firmware which makes your products compliant, well, doesn't that resolve the problem without risking that a judge would order fines or something more onerous like a product recall?
In any case, I'm happy to see this feature added. I locked my volume down to about 90% of the max, just to avoid accidentally blasting my own ears at max volume.
You're welcome...but note you asked for "better software written for the mass market", and that much I can point to [1]. Quantrix isn't widely deployed compared with Office, either, but anything written for both Windows and MacOS X seems to pass the goalposts originally set.
If you start adding terms like "must make it commercially" (according to whose definitions?) or "must interoperate with M$ Office tools", well, OK, but it's possible to shift the goalposts too far to be making an unbiased point....:-)
[1]: Note that you're welcome to hold the opinion that Excel 2003 is better than Improv, or that Quantrix is good at some things but lacks some other feature that the vast Office suite integration provides, but I feel that both are at least credible as contenders for being better than Excel in many areas.
You're on Slashdot, neighbor. Using base-2 for logarithms and exponents is generally taken as the default (ie, we assume that "KB" means 1024 or 2**10 bytes here, regardless of what the metric pendants say about KiB).
Question: why does testing all 256 combinations of this 8-bit IC not guarantee that a real system using it is going to work all of the time?
Let's suppose this IC was connected to a driver board which activated solinoids in a pinball machine, which operated the various pop bumpers, drop targets, and mechanical toys, and each bit activated a different physical device. Testing each bit pattern isn't enough to test all of the physical parameters of the real-world system, because the transitions from one bit pattern to another take time for the physical devices to move, and involve transient behavior and potential interactions, etc. So you would need to test combinations of bit patterns changing over an appropriate time interval to see what the physical system does in response, to be sure that it actually works the right way.
Second Question: why doesn't a memory tester like memtest86.org's try to write every possible bit pattern to RAM?
This has an obvious answer, but it's still worth thinking about. If you can't prove correctness without going through brute-force exponential testing which would require more time and resources then you have available, one should utilize problem decomposition and testing of smaller components which are manageable in size. And then use the principle of software composition where if your individual components work properly, and you've combined them properly into an architecture which solves the overall problem correctly, then you'll end up with a functioning system.
This leads to questions like:
"If I build the system using the architecture I've got, will it actually solve the problem it's supposed to?"
"Can the components I've got be assembled into the design? Do they implement the intra-component APIs that they are supposed to have?"
"Do the individual components work properly?"
Unit testing address the last of these three, and is often the easiest part of the business when dealing with large, multitiered systems.
Excel, as it happens, is the best software ever written for the mass market. Don't belive me? Well, give counterexamples.
I don't entirely disagree with you-- Excel is probably the best written part of the Office suite, and it is used so widely because it does provide very useful, well-implemented functionality, but I can still think of counterexamples:
If you need exponential time to test your code as it increases in size, you're probably going about the business in a less than ideally efficient way. Unit-testing can help. So will proper design.
I'm being too polite, what you've described sounds a lot like the testing equivalent of the bogosort algorithm, ie, sorting a deck of cards by shuffling them randomly and then checking to see whether you happened to shuffle them into sorted order. A bogosort takes exponential time, whereas an ideal sort is O(n * log(n)) worst-case.
If you were writing a program which needed to convert between N different image formats, would you write something that converted between each combination (ie, N*N conversion routines), or would you be more clever and do what Jeff Pozanker did with PBM (ie, write a common intermediate format and only N * 2 conversion routines)...?
Why not? Windows, within the confines of computer applications, is a valid trade mark. No other product can be named windows within the operating system world. No BeOS Windows and no Mac OS Windows.
What about the "X11 Window System", commonly called "X Windows"?
I know that MIT and CMU were working on X10 and predecessors like the Andrew Window System as far back as 1985 according to http://en.wikipedia.org/wiki/Andrew_Project, and Apple was also around.
Anything wrong with that?
Ask Lindows what they think of the matter...?
Conversely, Microsoft Word, as a computer program has every right to have a trademark on it. No other computer program should be called Word.
If page rankings end up depending on how litigious the businesses they belong to are, then we can forget about having such a thing as a useful search engine altogether.
I'd actually find such a rating system useful, in the negative sense: I'd avoid the sites which are highly litigious.:)
1. attach magnets to these analysts 2. use spin they produce to generate some electricity 3. use electricity to power more game consoles 4. ????? 5. PROFIT!
You, sir, are a genius. (But, ha ha! I filed the patent application first!)
Um, I don't have the Python code from Civ4 immediately handy (I could check later tonight), but the language conventions in Python for triple-quoted strings and suchlike tend to result in code with good embedded documentation, much like what JavaDoc will give you.
Call help(foo) for pretty much anything in Python and you will get good module or class documenation...
You're welcome for the link to ClamAV. It started as a tool for people running Unix mailservers to have a useful tool for scanning for Windows malware like viruses and worms, and has since evolved to also do a good job of identifying phishing scams and other email frame-based exploits. ClamAV isn't perfect-- no anti-virus software is, apropos-- but it's worthwhile and certainly the price is right.
I wouldn't go so far to describe McAfee or Symantec as "crappy", but their track record has encountered far more than just one bad update. There have been other fairly significant problems with various anti-virus software over the years which seem to crop up every six months or so, you can hunt down some amusing articles on the Register for example, if you don't want to research the tech articles and knowledge base items in MSDN:
Search Results 1-10 of 1,153 containing Norton AntiVirus problem (0.14 seconds):
I don't want to blame McAfee or Symantec excessively, but you do realize that you are posting to a thread where an update to their products ended up breaking hundreds or thousands of machines.
As for freely available A/V software, try ClamAV at http://www.clamav.net/, or the associated ClamWin for Windows. The site has some studies and comparisons that people have done against other antivirus products. I think that ClamAV's scanner is somewhat slower than the big name AV products, but it seems to be more thorough about catching nested viruses (ie, a zip containing a rar containing a.exe or whatever).
A fair study on this topic would be interesting, no matter what the results are. However, it's really easy to set up a study that simply confirms the preconceptions you are looking for by biasing the questions you ask and by selecting a sample from the results that shows the conclusions you want to draw.
There is a big difference between "lets take a look at the effect of playing games on children", and "lets see whether we can find a link between playing games and violent behavior because that would be really useful for the people paying for the study to then be able to do something with that hoped-for result."
Lord British and Origin produced some of the finest computer games made, and in particular Ultima III through V, Wing Commander & Privateer, and Moebius. Games which actually had a plot, and significant interaction between player actions and consequences.
Would be nice to see something like this for all platforms.
Sure, no problem-- just connect a machine running the platform of your choice to the Internet and let it go.:-)
A well patched windows installation might be as good as a well patched OS X installation.
That certainly hasn't been true in the past, but sooner or later, OS X is likely going to have some high-profile incidents like a Slammer or Nimda equivalent. It's also true that Intel-based Macs are going to be easier targets for people used to writing x86-based malware than the older PowerPC based hardware.
In any event, Windows-based attacks on public IPs occur so frequently that an unpatched Windows machine, say a fresh install from a CD, is likely to be compromised in less than an hour, and it's entirely possible for a fresh install of Windows to be hacked before the admin can finish downloading the first run of security patches.
But don't get me wrong, anyone who runs a webserver nowadays can watch the computer equivalent of assault happening against PHP vulnerabilities, blogging software vulnerabilities, formmail.pl, and lots of other things.
Agreed, or better yet, pick up the sequel, Independence War 2. IW2 is a great game, right up there with Privateer or Starflight. IW1 was good but had some very difficult missions that involved hunting down some magic key that activate a special option that would be useful for that one mission alone, whereas the use of remote control in IW 2 was made more reasonable and more optional (ie, being clever would help you complete a mission more easily, but you could try to do it the hard way and even succeed in IW2, in IW1 you'd fail unless you followed the pre-scripted expected course).
FreeBSD can push 500K packets per second today, and some of the network developers are working on 10GB ethernet as well.
Andre Oppermann wrote: net.inet.ip.fastforwarding has precedence over net.isr.direct and enabling both at the same doesn't gain you anything. Fastforwarding is about 30% faster than all other methods available, including polling. On my test machine with two em(4) and an AMD Opteron 852 (2.6GHz) I can route 580'000 pps with zero packet loss on -CURRENT. An upcoming optimization that will go into -CURRENT in the next few days pushes that to 714'000 pps. Futher optimizations are underway to make a stock kernel do close to or above 1'000'000 pps on the same hardware.
[...and in another message with more detailed measurements... ]
As part my funded TCP/IP optimization work I'm doing lots of measurements and profiling with an Agilent N2X network tester and calibrated traffic generator.
The following data shall serve as baseline of the current performance we get out of FreeBSD 7-current. More to come tomorrow though.
OS: FreeBSD 7-current as of 20051222-1600 UTC KERNEL: Generic kernel, minus WITNESS and INVARIANTS, plus HWPMC, HZ=1000 HARDWARE: Dual Opteron 852 2.6Ghz, Tyan S2882 Mobo with AMD-8131 PCI-X tunnel HARDWARE: dual Broadcom Gigabit BMC5704C PCI-X-133 ("bge") HARDWARE: dual Intel Gigabit 82546EB PCI-X-133 ("em")
Uniprocessor kernel
bge:
normal forwarding bge0->bge1: @64/326kpps/166us/402kpps(30%Loss)/194us
normal forwarding bge0->bge1: @1500/81kpps/520us
normal forwarding bge0->disc0: @64/1205kpps
IP fastforwarding bge0->bge1: @64/565kpps/192us/575kpps(60%Loss)/1090us
IP fastforwarding bge0->bge1: @1500/81kpps/730us
IP fastforwarding bge0->disc0: @64/1160kpps
net.isr.direct=1 bge0->bge1: @64/476kpps/211us/487kpps(68%Loss)/1284us
net.isr.direct=1 bge0->bge1: @1500/81kpps/760us
net.isr.direct=1 bge0->disc0: @64/1250kpps
polling (*) bge0->bge1: @64/420kpps(9%Loss)/1385us/416kpps(72%Loss)/1600us
polling (*) bge0->bge1: @1500/71kpps(9%Loss)/850us
polling (*) bge0->disc0: @64/697kpps
Comments: Under full load the normal processing breaks completely down
while with IP fastforwarding it levels off but continues to forward.
Strangely with polling it has 9% loss at all loads (even at 1% wirespeed).
May be related to HZ=1000.
em:
normal forwarding em0->em1: @64/372kpps/112us/396kpps(11%Loss)/131us
normal forwarding em0->em1: @1500/81kpps/170us
normal forwarding em0->disc0: @64/1130kpps
IP fastforwarding em0->em1: @64/565kpps/45us/585kpps(4%Loss)/1600us
IP fastforwarding em0->em1: @1500/81kpps/135us
IP fastforwarding em0->disc0: @64/1116kpps
net.isr.direct=1 em0->em1: later
net.isr.direct=1 em0->disc0: later
polling (*) em0->em1: later
polling (*) em0->disc0: later
Cisco's routers are extremely reliable and can last for many years, but they aren't magic boxes that never fail, either. I've dealt with 2 out of about 10 Cisco 1800 or 2x00 routers dying because of the flash memory going wacky. And I just had an Cisco 871 go bad at a client site, too, taking out their VPN capabilities.
I don't see that failure rate as being much different from a Soekris 4501 running NetBSD via CF, or if you want to set up a bigger machine and perhaps run squid or openvpn on the router box, maybe use a diskful machine in RAID-1 rather than flash memory.
It is also missing the nice interface, and commercial applications.
What, you don't think KDE or Gnome are as nice a GUI as OS X? (Well, you'd be right. But they're tolerable.)
As for commercial applications, FreeBSD has at least a few. There's a list of approximately 500 FreeBSD ports which contain sufficient restrictions against redistribution that they cannot be redistributed on the ISO images that people use to burn CD's. Not all of these are "commercial applications", but things like Acrobat Reader or StarOffice or Oracle's database, Linux versions of commercial games running in FreeBSD's Linux-emulation mode, or perhaps the various virus scanners would count:
"""You know, that bit where it says "derivative code must also be GPL'd" """
Only, the GPL doesn't actually say any such thing. Third-party code which includes GPL'ed source code must be released under terms which are compatible with the GPL if the software is redistributed to others, but you can use a mix of GPL'ed code and proprietary code yourself if you don't redistribute (this is why you can mix a GPL'ed Linux kernel with proprietary drivers from ATI or nVidia, and not have a problem).
You can also release your third-party code together with GPL'ed code which forms a deriviative work under the terms of other permissive open source licenses, such as the BSD or MIT/X11 licenses. The entire list of "GPL-Compatible Free Software Licenses" is kept here:
Who made me the genius I am today The mathematician that others all quote Who's the professor that made me that way The greatest that ever got chalk on his coat
One man deserves the credit One man deserves the blame And Nicolai Ivanovich Lobachevsky is his name, hi! Nicolai Ivanovich Lobache-
I am never forget the day I first meet the great Lobachevsky. In one word he told me secret of success in mathematics. Plagiarize!
Plagiarize Let no one else's work evade your eyes Remember why the good Lord made your eyes So don't shade your eyes But plagiarize, plagiarize, plagiarize Only be sure always to call it please "research"
(Tom Lehrer, _Lobachevsky_. Imagine someone dancing the Mamuska while playing a banal Russian folk tune on a piano.:-)
The best rooms for sound are rectangular, where none of the dimensions of the space (length, width, height) are the same or are multiples of each other. For example, a room that is 16' x 24' x 8' will sound fairly bad and have resonance issues with the low-freqnency sound in particular.
A room that's 11' x 19' x 8' will sound much better.
It was not the connection between Slashdot and Newsforge I am wondering about. It's the relationship to industrybrains.com, pricegrabber.com, falkad.net, googlesyndication.com and whoever else that wasn't mentioned either. Do they do business with Ziff-Davis, too, perhaps?
[ You do acknowledge that there was approximately 25 ad links surrounding a single paragraph of content which could have been linked to directly, instead? ]
With regard to the other comment, yes, "Mr. Brockmeier mentioned the BSDs toward the end of the article -- and FreeDOS too". Is the difference between actually testing these alternatives and briefly mentioning them in passing unclear?
However, what bugs me about this supposed comparison of how well Linux and Windows perform on old hardware isn't that neither Ziff-Davis nor linux.com tried the BSDs, but the fact that nobody seemed to actually test whether the various Windows distributions worked on old hardware, either.
If you want compare A to B, and you only show A without ever discussing B, you haven't actually done the comparison you claimed you were doing, true...?
Slashdot Mirror
Perhaps, but it was a fairly good rant. :)
/robots.txt file, or a W3C privacy policy, or "Terms of use", etc.
"If someone connects to your web server, who authorised the access? In this case it's reasonable to assume that a web server that's open to the public network provides implicitly authorised access... or is it?"
Leaving a port open does not by itself constititute authorization.
If you publish a www.example.com record in the DNS, it would be entirely reasonable to assume that you've authorized people to stop by your website, yes. There are other conventions which apply, such as having a
Note that this does not mean that launching a set of requests to exploit, say, a PHP vulnerability, would be OK just because the machine is listed in the DNS. Legitimate users would have a very different click-trail than someone trying to break in.
Agreed, mostly. I've seen some portable music players make a point of noting their volume-limitation features in years past, so some vendors have obviously run into this issue before.
If an iPod plays louder than the local law permits, that's a clear and easily tested matter of fact, and even a clever lawyer is going to have a hard time spinning that on the defense.
On the other hand, if someone tells you about a problem, and you promptly release new firmware which makes your products compliant, well, doesn't that resolve the problem without risking that a judge would order fines or something more onerous like a product recall?
In any case, I'm happy to see this feature added. I locked my volume down to about 90% of the max, just to avoid accidentally blasting my own ears at max volume.
You're welcome...but note you asked for "better software written for the mass market", and that much I can point to [1]. Quantrix isn't widely deployed compared with Office, either, but anything written for both Windows and MacOS X seems to pass the goalposts originally set.
:-)
If you start adding terms like "must make it commercially" (according to whose definitions?) or "must interoperate with M$ Office tools", well, OK, but it's possible to shift the goalposts too far to be making an unbiased point....
[1]: Note that you're welcome to hold the opinion that Excel 2003 is better than Improv, or that Quantrix is good at some things but lacks some other feature that the vast Office suite integration provides, but I feel that both are at least credible as contenders for being better than Excel in many areas.
You're on Slashdot, neighbor. Using base-2 for logarithms and exponents is generally taken as the default (ie, we assume that "KB" means 1024 or 2**10 bytes here, regardless of what the metric pendants say about KiB).
Question: why does testing all 256 combinations of this 8-bit IC not guarantee that a real system using it is going to work all of the time?
Let's suppose this IC was connected to a driver board which activated solinoids in a pinball machine, which operated the various pop bumpers, drop targets, and mechanical toys, and each bit activated a different physical device. Testing each bit pattern isn't enough to test all of the physical parameters of the real-world system, because the transitions from one bit pattern to another take time for the physical devices to move, and involve transient behavior and potential interactions, etc. So you would need to test combinations of bit patterns changing over an appropriate time interval to see what the physical system does in response, to be sure that it actually works the right way.
Second Question: why doesn't a memory tester like memtest86.org's try to write every possible bit pattern to RAM?
This has an obvious answer, but it's still worth thinking about. If you can't prove correctness without going through brute-force exponential testing which would require more time and resources then you have available, one should utilize problem decomposition and testing of smaller components which are manageable in size. And then use the principle of software composition where if your individual components work properly, and you've combined them properly into an architecture which solves the overall problem correctly, then you'll end up with a functioning system.
This leads to questions like:
"If I build the system using the architecture I've got, will it actually solve the problem it's supposed to?"
"Can the components I've got be assembled into the design? Do they implement the intra-component APIs that they are supposed to have?"
"Do the individual components work properly?"
Unit testing address the last of these three, and is often the easiest part of the business when dealing with large, multitiered systems.
I don't entirely disagree with you-- Excel is probably the best written part of the Office suite, and it is used so widely because it does provide very useful, well-implemented functionality, but I can still think of counterexamples:
Lotus Improv
Quantrix
If you need exponential time to test your code as it increases in size, you're probably going about the business in a less than ideally efficient way. Unit-testing can help. So will proper design.
I'm being too polite, what you've described sounds a lot like the testing equivalent of the bogosort algorithm, ie, sorting a deck of cards by shuffling them randomly and then checking to see whether you happened to shuffle them into sorted order. A bogosort takes exponential time, whereas an ideal sort is O(n * log(n)) worst-case.
If you were writing a program which needed to convert between N different image formats, would you write something that converted between each combination (ie, N*N conversion routines), or would you be more clever and do what Jeff Pozanker did with PBM (ie, write a common intermediate format and only N * 2 conversion routines)...?
What about the "X11 Window System", commonly called "X Windows"?
I know that MIT and CMU were working on X10 and predecessors like the Andrew Window System as far back as 1985 according to http://en.wikipedia.org/wiki/Andrew_Project, and Apple was also around.
Anything wrong with that?
Ask Lindows what they think of the matter...?
Conversely, Microsoft Word, as a computer program has every right to have a trademark on it. No other computer program should be called Word.
Hmph. Tell that to AbiWord.
I'd actually find such a rating system useful, in the negative sense: I'd avoid the sites which are highly litigious. :)
1. attach magnets to these analysts
2. use spin they produce to generate some electricity
3. use electricity to power more game consoles
4. ?????
5. PROFIT!
You, sir, are a genius.
(But, ha ha! I filed the patent application first!)
Um, I don't have the Python code from Civ4 immediately handy (I could check later tonight), but the language conventions in Python for triple-quoted strings and suchlike tend to result in code with good embedded documentation, much like what JavaDoc will give you.
Call help(foo) for pretty much anything in Python and you will get good module or class documenation...
You're welcome for the link to ClamAV. It started as a tool for people running Unix mailservers to have a useful tool for scanning for Windows malware like viruses and worms, and has since evolved to also do a good job of identifying phishing scams and other email frame-based exploits. ClamAV isn't perfect-- no anti-virus software is, apropos-- but it's worthwhile and certainly the price is right.
% 3Ben-us%3B295824; en-us;265824; en-us;276504 ...etc, etc.
I wouldn't go so far to describe McAfee or Symantec as "crappy", but their track record has encountered far more than just one bad update. There have been other fairly significant problems with various anti-virus software over the years which seem to crop up every six months or so, you can hunt down some amusing articles on the Register for example, if you don't want to research the tech articles and knowledge base items in MSDN:
Search Results 1-10 of 1,153 containing Norton AntiVirus problem (0.14 seconds):
http://support.microsoft.com/default.aspx?scid=kb
http://support.microsoft.com/default.aspx?scid=kb
http://support.microsoft.com/default.aspx?scid=kb
I don't want to blame McAfee or Symantec excessively, but you do realize that you are posting to a thread where an update to their products ended up breaking hundreds or thousands of machines.
.exe or whatever).
As for freely available A/V software, try ClamAV at http://www.clamav.net/, or the associated ClamWin for Windows. The site has some studies and comparisons that people have done against other antivirus products. I think that ClamAV's scanner is somewhat slower than the big name AV products, but it seems to be more thorough about catching nested viruses (ie, a zip containing a rar containing a
What you've asked is the operative question.
A fair study on this topic would be interesting, no matter what the results are. However, it's really easy to set up a study that simply confirms the preconceptions you are looking for by biasing the questions you ask and by selecting a sample from the results that shows the conclusions you want to draw.
There is a big difference between "lets take a look at the effect of playing games on children", and "lets see whether we can find a link between playing games and violent behavior because that would be really useful for the people paying for the study to then be able to do something with that hoped-for result."
Lord British and Origin produced some of the finest computer games made, and in particular Ultima III through V, Wing Commander & Privateer, and Moebius. Games which actually had a plot, and significant interaction between player actions and consequences.
...let's just say it's been a while since I've played a game which made me think as much about such things.
http://en.wikipedia.org/wiki/Virtues_of_Ultima
Sure, no problem-- just connect a machine running the platform of your choice to the Internet and let it go. :-)
A well patched windows installation might be as good as a well patched OS X installation.
That certainly hasn't been true in the past, but sooner or later, OS X is likely going to have some high-profile incidents like a Slammer or Nimda equivalent. It's also true that Intel-based Macs are going to be easier targets for people used to writing x86-based malware than the older PowerPC based hardware.
In any event, Windows-based attacks on public IPs occur so frequently that an unpatched Windows machine, say a fresh install from a CD, is likely to be compromised in less than an hour, and it's entirely possible for a fresh install of Windows to be hacked before the admin can finish downloading the first run of security patches.
But don't get me wrong, anyone who runs a webserver nowadays can watch the computer equivalent of assault happening against PHP vulnerabilities, blogging software vulnerabilities, formmail.pl, and lots of other things.
Agreed, or better yet, pick up the sequel, Independence War 2. IW2 is a great game, right up there with Privateer or Starflight. IW1 was good but had some very difficult missions that involved hunting down some magic key that activate a special option that would be useful for that one mission alone, whereas the use of remote control in IW 2 was made more reasonable and more optional (ie, being clever would help you complete a mission more easily, but you could try to do it the hard way and even succeed in IW2, in IW1 you'd fail unless you followed the pre-scripted expected course).
FreeBSD can push 500K packets per second today, and some of the network developers are working on 10GB ethernet as well.
...and in another message with more detailed measurements... ]
s
Andre Oppermann wrote:
net.inet.ip.fastforwarding has precedence over net.isr.direct and
enabling both at the same doesn't gain you anything. Fastforwarding
is about 30% faster than all other methods available, including
polling. On my test machine with two em(4) and an AMD Opteron 852
(2.6GHz) I can route 580'000 pps with zero packet loss on -CURRENT.
An upcoming optimization that will go into -CURRENT in the next
few days pushes that to 714'000 pps. Futher optimizations are
underway to make a stock kernel do close to or above 1'000'000 pps
on the same hardware.
[
As part my funded TCP/IP optimization work I'm doing lots of measurements
and profiling with an Agilent N2X network tester and calibrated traffic
generator.
The following data shall serve as baseline of the current performance we
get out of FreeBSD 7-current. More to come tomorrow though.
OS: FreeBSD 7-current as of 20051222-1600 UTC
KERNEL: Generic kernel, minus WITNESS and INVARIANTS, plus HWPMC, HZ=1000
HARDWARE: Dual Opteron 852 2.6Ghz, Tyan S2882 Mobo with AMD-8131 PCI-X tunnel
HARDWARE: dual Broadcom Gigabit BMC5704C PCI-X-133 ("bge")
HARDWARE: dual Intel Gigabit 82546EB PCI-X-133 ("em")
Uniprocessor kernel
bge:
normal forwarding bge0->bge1: @64/326kpps/166us/402kpps(30%Loss)/194us
normal forwarding bge0->bge1: @1500/81kpps/520us
normal forwarding bge0->disc0: @64/1205kpps
IP fastforwarding bge0->bge1: @64/565kpps/192us/575kpps(60%Loss)/1090us
IP fastforwarding bge0->bge1: @1500/81kpps/730us
IP fastforwarding bge0->disc0: @64/1160kpps
net.isr.direct=1 bge0->bge1: @64/476kpps/211us/487kpps(68%Loss)/1284us
net.isr.direct=1 bge0->bge1: @1500/81kpps/760us
net.isr.direct=1 bge0->disc0: @64/1250kpps
polling (*) bge0->bge1:
@64/420kpps(9%Loss)/1385us/416kpps(72%Loss)/1600u
polling (*) bge0->bge1: @1500/71kpps(9%Loss)/850us
polling (*) bge0->disc0: @64/697kpps
Comments: Under full load the normal processing breaks completely down
while with IP fastforwarding it levels off but continues to forward.
Strangely with polling it has 9% loss at all loads (even at 1% wirespeed).
May be related to HZ=1000.
em:
normal forwarding em0->em1: @64/372kpps/112us/396kpps(11%Loss)/131us
normal forwarding em0->em1: @1500/81kpps/170us
normal forwarding em0->disc0: @64/1130kpps
IP fastforwarding em0->em1: @64/565kpps/45us/585kpps(4%Loss)/1600us
IP fastforwarding em0->em1: @1500/81kpps/135us
IP fastforwarding em0->disc0: @64/1116kpps
net.isr.direct=1 em0->em1: later
net.isr.direct=1 em0->disc0: later
polling (*) em0->em1: later
polling (*) em0->disc0: later
(*) max_burst=1000, user_frac=0, each_burst=30
Sponsored by: TCP/IP Optimization Fundraise 2005
Cisco's routers are extremely reliable and can last for many years, but they aren't magic boxes that never fail, either. I've dealt with 2 out of about 10 Cisco 1800 or 2x00 routers dying because of the flash memory going wacky. And I just had an Cisco 871 go bad at a client site, too, taking out their VPN capabilities.
I don't see that failure rate as being much different from a Soekris 4501 running NetBSD via CF, or if you want to set up a bigger machine and perhaps run squid or openvpn on the router box, maybe use a diskful machine in RAID-1 rather than flash memory.
What, you don't think KDE or Gnome are as nice a GUI as OS X?
(Well, you'd be right. But they're tolerable.)
As for commercial applications, FreeBSD has at least a few. There's a list of approximately 500 FreeBSD ports which contain sufficient restrictions against redistribution that they cannot be redistributed on the ISO images that people use to burn CD's. Not all of these are "commercial applications", but things like Acrobat Reader or StarOffice or Oracle's database, Linux versions of commercial games running in FreeBSD's Linux-emulation mode, or perhaps the various virus scanners would count:
http://www.freebsd.org/cgi/cvsweb.cgi/ports/LEGAL
(click on download)
And there is no DRM in OS X except that used on the ITMS music and possibly that used to make it only run on Macs.
Well, just how much more DRM do you really want?
To me, "no DRM" means "no DRM" and not "some DRM".
"""You know, that bit where it says "derivative code must also be GPL'd" """
Only, the GPL doesn't actually say any such thing. Third-party code which includes GPL'ed source code must be released under terms which are compatible with the GPL if the software is redistributed to others, but you can use a mix of GPL'ed code and proprietary code yourself if you don't redistribute (this is why you can mix a GPL'ed Linux kernel with proprietary drivers from ATI or nVidia, and not have a problem).
You can also release your third-party code together with GPL'ed code which forms a deriviative work under the terms of other permissive open source licenses, such as the BSD or MIT/X11 licenses. The entire list of "GPL-Compatible Free Software Licenses" is kept here:
http://www.fsf.org/licensing/licenses/index_html
Thanks. (At least nobody modded it "offtopic" or "flamebait"...)
Who made me the genius I am today
:-)
The mathematician that others all quote
Who's the professor that made me that way
The greatest that ever got chalk on his coat
One man deserves the credit
One man deserves the blame
And Nicolai Ivanovich Lobachevsky is his name, hi!
Nicolai Ivanovich Lobache-
I am never forget the day I first meet the great Lobachevsky.
In one word he told me secret of success in mathematics. Plagiarize!
Plagiarize
Let no one else's work evade your eyes
Remember why the good Lord made your eyes
So don't shade your eyes
But plagiarize, plagiarize, plagiarize
Only be sure always to call it please "research"
(Tom Lehrer, _Lobachevsky_. Imagine someone dancing the Mamuska while playing a banal Russian folk tune on a piano.
The best rooms for sound are rectangular, where none of the dimensions of the space (length, width, height) are the same or are multiples of each other. For example, a room that is 16' x 24' x 8' will sound fairly bad and have resonance issues with the low-freqnency sound in particular.
A room that's 11' x 19' x 8' will sound much better.
Try more like 18 cents per KW/h in NY City, three times as much...
It was not the connection between Slashdot and Newsforge I am wondering about. It's the relationship to industrybrains.com, pricegrabber.com, falkad.net, googlesyndication.com and whoever else that wasn't mentioned either. Do they do business with Ziff-Davis, too, perhaps?
[ You do acknowledge that there was approximately 25 ad links surrounding a single paragraph of content which could have been linked to directly, instead? ]
With regard to the other comment, yes, "Mr. Brockmeier mentioned the BSDs toward the end of the article -- and FreeDOS too". Is the difference between actually testing these alternatives and briefly mentioning them in passing unclear?
However, what bugs me about this supposed comparison of how well Linux and Windows perform on old hardware isn't that neither Ziff-Davis nor linux.com tried the BSDs, but the fact that nobody seemed to actually test whether the various Windows distributions worked on old hardware, either.
If you want compare A to B, and you only show A without ever discussing B, you haven't actually done the comparison you claimed you were doing, true...?