i think he's doing more of a rant on the fact that computers have been billed as being the devices that allow you to do anything. hell, i've even seen commercials that boast "You will be a better $whatever by using our computerized system."
i guess it's kind of like all the shoe manufacturers selling shoes by making it seem as though their shoe will make you jump higher/run faster/etc... and i can tell you from experience that none of those things happen;)
the firewall rules and other data structures required for the TCP/IP stack in the kernel all require various amounts of memory. with no swap, it would be possible to completely run out of memory given a sufficient amount of traffic going through the interfaces. that's when things get nasty.
i'm thinking though that it would take quite a bit of traffic for even a modest machine to have problems.
well, technically speaking you could remove the script that umount's the mounted filesystems. i'm not sure how the kernel would see the filesystems in the halted state, but it might be interesting to find out...
personable? Theo? Good God! he's like a living godzilla spewing white hot embers of death everywhere...
not that i don't love the OpenBSD project (i have several machines running it), but to say that Theo is personable is like saying everyone needs a porcupine to snuggle up with at night.
and on that same token it should be ok to say "here are all the parts to a bomb and here is it how it works". and thusly illegal to say "here is how to make a bomb; let's blow some people up".
it's not the information that poses the problem. it's the people who (mis)use the information.
one thing to mention though is the loss of space due to configuring RAID-0+1 vs RAID-5. let's say you have 8 drives (10GB each). with RAID-0+1 you get 40GB (4 drives for primary, 4 drives for mirror). with RAID-5, you would get somewhere around 70GB (all 8 drives configured for RAID-5, parity typically takes up about 1 drive).
so, it basically comes down to: is the loss of space worth the increased speed? and that is a question that should be asked on a case-by-case basis.
also, when you use hardware RAID controllers, the parity calculations are handled by the controller, so no CPU time is used. RAID-5 should still have slower access in theory... in practice, i don't think the difference is all *that* great.
it's funny how we have laws that provide for fair use and allow us to make back-up copies of software, CD's, DVD's, etc. but when it comes down to the courts actually upholding our right to do so, they typically fall towards the company that creates technology to prevent us from using these backups.
"yes, you have the right to make backups of Playstation games, but no, you do not have the right to play them since Sony has technology to prevent you from doing so."
seems like there really isn't any fair use anymore.
i'm wondering if i could purchase a Playstation CD and return it to the company that made it for a replacement in the event that it got scratched/damaged. has anyone done this? i mean, they say they don't want to let us make copies, but i should only have to pay for the game once. i'd be more than happy to pay for the replacement media and shipping. i'm just betting that they'd charge me $45 for a $50 game. now, what does that tell about how much the actual game is worth?
i just went to their website and noticed that they make some pretty hefty claims in their ad. i'm just wondering if the can really get away with it...
.... the first player that can double as a portable hard disk
erm... hasn't the archos jukebox been able to do this for quite awhile now? i bought the 20GB version and it's basically just a USB harddrive with an mp3 player stuck on top.
.... the first player with upgradeable firmware
erm... again, my archos can do this. you copy the firmware to the drive when it is off and then turn it back on to activate it.
i guess these claims work for the CF world, but that's not really clarified in their statement. and it's not that i'm knocking this product. i might have actually gotten one instead of the archos if they had been available when i bought my archos. i'm just wondering if they can really make these claims...
perhaps it would be better to spend time and money focusing on educating the people who run insecure software/hardware. i don't really know how you'd go about enforcing it except to hold people liable for any problems caused by cracked servers.
actually, i should probably look into that. my last fileserver had an Intel 10/100 Pro+ Management card, but FreeBSD choked on it every month or so. i haven't had any problems with Linux and the NetGear cards, but it never hurts to be cautious. thanks for the suggestion.
i just bought a new file server at home with the following specs:
1 x Tyan 2640 motherboard (Athlon MP)
2 x 1800+ MP (1.5GHz) Athlon CPU
2 x ThermalRight SK-6 heatsink
2 x Delta "Black Label" fan
2 x Tekram 390U3D Dual Ultra-160 SCSI HBA
6 x Seagate 15K RPM 36GB U160 SCSI Cheetah
1 x Seagate 10K RPM 4GB LVD SCSI Cheetah
1 x Plextor 40X SCSI CD-ROM
1 x Plextor 12/10/32X SCSI CD-Recordable
1 x NetGear FA310TX 10/100Mbps NIC
1 x Trident Video (4MB, PCI)
1 x CoolerMaster ATC-400 rackmount case
4 x Crucial 256MB ECC DDR-SDRAM
it's not much on the gaming front, but damn does it crank out the dnet keys (RC5: 10.5Mkeys/s)!
i'm also very pleased with the disk throughput. each individual 15K RPM Cheetah pulls ~55MB/s. all 6 done in RAID-5 maxes the PCI bus (133MB/s). theoretically it should cap around 266MB/s (due to the 64-bit PCI bus), but i'm not sure why it doesn't. still, 100+ MB/s is pretty damned fast.
i've just done a complete overhaul of my firewall (2.7 upgraded to 2.8) and in doing so, chose to use the new pf package in the 3.0 release. so far everything has been working great. the move to the pf package was extremely smooth. i only had to change a couple of ICMP rules in my ipfilter ruleset, and re-writing the NAT rules wasn't hard at all with a little help from the pfctl man page. from my testing it appears as though it is performing every bit as good as ipfilter.
the only downside to pf that i see is in the new packet logging scheme. i've written several scripts that used the old ipmon style of syslog logging to run stats for my firewall and the new pf tcpdump format won't work with them. however, it does log, so i can always re-write my scripts to use the new format.
but i think there is something to be said for intuitiveness. let's say that there isn't someone there to teach about either setup.exe or rpm/dpkg/apt/etc. i think you find that most people (yes, even the not-so-smart ones) could eventually figure out that you need to double-click on setup.exe and then hit "Next" a whole bunch of times. conversely, giving a user a.rpm,.deb, or.tgz package doesn't necessarily mean that they will ever be able to figure out what to do with it. in this case you'd be assuming that user would 1) find the binary responsible for package management, and 2) figure out how to use all of the command-line flags to install the package. quite a far cry from double-click... clicky, clicky, clicky.
however, i have to agree that if people are taught how to do things, this becomes a non-issue. i guess everyone just gets caught up in the "easy-as-possible" way of thinking and the teaching part of things goes to the wayside.
it seems as though there are two main types of metadata: OS dependent and user dependent. the OS dependent metadata is information important to the system (permissions, ownership, filesize, etc). everything else (file type, etc) would be user dependent.
i'm wondering if there would be a way to have the OS dependent data remain static (as it usually is), but have the user dependent data be variable. for instance, you could have a 'type' field, but you might also want an 'open', 'print', or 'edit' field to specify perhaps which applications would be responsible for those actions.
for example, i might view images in GQView, but i can't edit them in that program. so, i'd create a new metadata field called 'edit' that would maybe point to the Gimp as the editor for that file. as long as the OS and/or programs i used adhered to this, life would be simple.
since the user dependent metadata would be variable, not all files would need uneccessary metadata fields. but for those files where the end-user might want a broader range of actions associated with the file, many different possibilites could exist.
of course, file-type information would not have to be the only thing stored in metadata. you could have version information, author information, authoring program, etc... but the metadata would only include that which the user would want included.
why is it that everything has to be billed as opensource? like that's really a selling point? i mean sure, there are some people that are just looking into the opensource thing, but it makes it seem as though that's the only reason you might want to run FreeBSD. never mind the fact that it's stable, fast, and has many of the popular apps that linux does. plus it includes linux emulation so that you can run native linux apps under FreeBSD. why not have a title that shows this in addition to it's being opensource? "FreeBSD: The opensource OS for your PC that's fast, versatile, and dependable". Sounds much more catchy to me.
one thing i have noticed about Sun hardware as compared to Intel hardware is that the Sun stuff will take quite a load without crumbling under the pressure. my experience has been that whether you've got 1 user or 100 users on the system, the Sun stuff pretty much runs the same. with Intel hardware, your results will vary, especially among the different OS's, but overall, i don't think Intel hardware handles a heavy load as well as Sun equipment.
wouldn't it be possible to use lex/flex to generate a C scanner that could parse the data just as easily as perl? C by itself may not be the answer, but there are some tools that use C that can make life a lot easier.
one problem: you are using the assumption that l33t actually owes you an answer. he doesn't. by saying RTFM, he's probably just trying to brush you off. if you have spent any time in an IRC support channel, you'll see many people parroting the same question over and over. RTFM is a way of saying "yes, i heard you... there is some documentation out there... find it for yourself and leave us alone". this is especially true for common problems (like setting up PPP) where large volumes of information exist on the specific problem.
if you are new to X, the best way to find information about X would probably be to start with a search engine. then, if you truly cannot find any information, ask "anyone know where i can get some info on X?". and do not expect a reply. if you don't get an answer, just hop to another IRC server. and if you never get an answer... well, sorry. that's just the way it is. just because you logged into the IRC server and found the support channel for X doesn't mean you are actually entitled to support (unless you specifically paid for it).
what about setting up a system of multiple sets of root servers that are able to handle requests for different DNS networks? the same kind of idea can be seen with IRC (Undernet vs. Efnet vs. Dalnet vs. etc... -- all are IRC networks, but each is seperate). Perhaps you could modify the way that resolvers work to allow for a new DNS network convention. URL examples:
using a system such as this would allow for users to select which DNS network they want to use for resolving name collisions.
of course, something like this would require modifications to the current name resolution system, but i think that with a little time and effort, implementing a system like this could work.
doh... express.com has Disney listed as the copyright holder (which makes sense), but yep, if you look at the enlarged picture of the back, there is the miramax logo.
my friends told me it was a Disney release and got me all worked up over nothing. now they must be punished >0
i've talked with a few people who have seen the japanese version (import), but have not seen the US version (it didn't come to any theaters around my area). they've talked about the fact that since this is a Disney release (not Miramax), then it might be edited for content. my friends mention one scene that involves a sword, blood, and a disembodied head.
so, to anyone who has seen both versions, is the US release the same? or watered down for the kiddies (well, as watered down as PG-13 gets anyways)?
yes, and they also use them for green (and yellow possibly??). there is an intersection next to my apartment complex that has one of these newer light systems installed. they are BRIGHT (especially the green for some reason)! there is definately no mistaking what color the light is. it even hurts my eyes at night sometimes!
Network apps do not query DNS servers directly (typically, though there are exceptions). Most simply make calls to the OS's resolver, which then forwards requests to the primary DNS server, which then queries a root DNS server.
So, adding code into Mozilla to use alternate root servers would simply be a waste of time and space.
Hacking BIND would be the way to go. You could have bind check to see which kind of TLD is being requested (official or alternate) and then have it query whichever root server. However, the same problems that are associated with having an alternate DNS system are still present (collisions, etc...).
this is a topic that i covered in my sociology class in college. basically what was said is that we (as a society) today are trading in our authority to allow the community to regulate the activities of our children rather than take a pro-active stance and do it ourself. You can see many examples of this with after-school programs and community programs that are there to 'keep your kids out of trouble'. It is quite interesting to see this in action.
Slashdot Mirror
i think he's doing more of a rant on the fact that computers have been billed as being the devices that allow you to do anything. hell, i've even seen commercials that boast "You will be a better $whatever by using our computerized system."
;)
i guess it's kind of like all the shoe manufacturers selling shoes by making it seem as though their shoe will make you jump higher/run faster/etc... and i can tell you from experience that none of those things happen
ahh.. ok.
...
... and i didn't realize that AIX's kernel would page out. interesting
the firewall rules and other data structures required for the TCP/IP stack in the kernel all require various amounts of memory. with no swap, it would be possible to completely run out of memory given a sufficient amount of traffic going through the interfaces. that's when things get nasty.
i'm thinking though that it would take quite a bit of traffic for even a modest machine to have problems.
well, technically speaking you could remove the script that umount's the mounted filesystems. i'm not sure how the kernel would see the filesystems in the halted state, but it might be interesting to find out...
personable? Theo? Good God! he's like a living godzilla spewing white hot embers of death everywhere...
not that i don't love the OpenBSD project (i have several machines running it), but to say that Theo is personable is like saying everyone needs a porcupine to snuggle up with at night.
and on that same token it should be ok to say "here are all the parts to a bomb and here is it how it works". and thusly illegal to say "here is how to make a bomb; let's blow some people up".
it's not the information that poses the problem. it's the people who (mis)use the information.
one thing to mention though is the loss of space due to configuring RAID-0+1 vs RAID-5. let's say you have 8 drives (10GB each). with RAID-0+1 you get 40GB (4 drives for primary, 4 drives for mirror). with RAID-5, you would get somewhere around 70GB (all 8 drives configured for RAID-5, parity typically takes up about 1 drive).
so, it basically comes down to: is the loss of space worth the increased speed? and that is a question that should be asked on a case-by-case basis.
also, when you use hardware RAID controllers, the parity calculations are handled by the controller, so no CPU time is used. RAID-5 should still have slower access in theory... in practice, i don't think the difference is all *that* great.
it's funny how we have laws that provide for fair use and allow us to make back-up copies of software, CD's, DVD's, etc. but when it comes down to the courts actually upholding our right to do so, they typically fall towards the company that creates technology to prevent us from using these backups.
"yes, you have the right to make backups of Playstation games, but no, you do not have the right to play them since Sony has technology to prevent you from doing so."
seems like there really isn't any fair use anymore.
i'm wondering if i could purchase a Playstation CD and return it to the company that made it for a replacement in the event that it got scratched/damaged. has anyone done this? i mean, they say they don't want to let us make copies, but i should only have to pay for the game once. i'd be more than happy to pay for the replacement media and shipping. i'm just betting that they'd charge me $45 for a $50 game. now, what does that tell about how much the actual game is worth?
i just went to their website and noticed that they make some pretty hefty claims in their ad. i'm just wondering if the can really get away with it...
.... the first player that can double as a portable hard disk
.... the first player with upgradeable firmware
erm... hasn't the archos jukebox been able to do this for quite awhile now? i bought the 20GB version and it's basically just a USB harddrive with an mp3 player stuck on top.
erm... again, my archos can do this. you copy the firmware to the drive when it is off and then turn it back on to activate it.
i guess these claims work for the CF world, but that's not really clarified in their statement. and it's not that i'm knocking this product. i might have actually gotten one instead of the archos if they had been available when i bought my archos. i'm just wondering if they can really make these claims...
perhaps it would be better to spend time and money focusing on educating the people who run insecure software/hardware. i don't really know how you'd go about enforcing it except to hold people liable for any problems caused by cracked servers.
actually, i should probably look into that. my last fileserver had an Intel 10/100 Pro+ Management card, but FreeBSD choked on it every month or so. i haven't had any problems with Linux and the NetGear cards, but it never hurts to be cautious. thanks for the suggestion.
i just bought a new file server at home with the following specs:
1 x Tyan 2640 motherboard (Athlon MP)
2 x 1800+ MP (1.5GHz) Athlon CPU
2 x ThermalRight SK-6 heatsink
2 x Delta "Black Label" fan
2 x Tekram 390U3D Dual Ultra-160 SCSI HBA
6 x Seagate 15K RPM 36GB U160 SCSI Cheetah
1 x Seagate 10K RPM 4GB LVD SCSI Cheetah
1 x Plextor 40X SCSI CD-ROM
1 x Plextor 12/10/32X SCSI CD-Recordable
1 x NetGear FA310TX 10/100Mbps NIC
1 x Trident Video (4MB, PCI)
1 x CoolerMaster ATC-400 rackmount case
4 x Crucial 256MB ECC DDR-SDRAM
it's not much on the gaming front, but damn does it crank out the dnet keys (RC5: 10.5Mkeys/s)!
i'm also very pleased with the disk throughput. each individual 15K RPM Cheetah pulls ~55MB/s. all 6 done in RAID-5 maxes the PCI bus (133MB/s). theoretically it should cap around 266MB/s (due to the 64-bit PCI bus), but i'm not sure why it doesn't. still, 100+ MB/s is pretty damned fast.
i've just done a complete overhaul of my firewall (2.7 upgraded to 2.8) and in doing so, chose to use the new pf package in the 3.0 release. so far everything has been working great. the move to the pf package was extremely smooth. i only had to change a couple of ICMP rules in my ipfilter ruleset, and re-writing the NAT rules wasn't hard at all with a little help from the pfctl man page. from my testing it appears as though it is performing every bit as good as ipfilter.
the only downside to pf that i see is in the new packet logging scheme. i've written several scripts that used the old ipmon style of syslog logging to run stats for my firewall and the new pf tcpdump format won't work with them. however, it does log, so i can always re-write my scripts to use the new format.
but i think there is something to be said for intuitiveness. let's say that there isn't someone there to teach about either setup.exe or rpm/dpkg/apt/etc. i think you find that most people (yes, even the not-so-smart ones) could eventually figure out that you need to double-click on setup.exe and then hit "Next" a whole bunch of times. conversely, giving a user a .rpm, .deb, or .tgz package doesn't necessarily mean that they will ever be able to figure out what to do with it. in this case you'd be assuming that user would 1) find the binary responsible for package management, and 2) figure out how to use all of the command-line flags to install the package. quite a far cry from double-click... clicky, clicky, clicky.
however, i have to agree that if people are taught how to do things, this becomes a non-issue. i guess everyone just gets caught up in the "easy-as-possible" way of thinking and the teaching part of things goes to the wayside.
it seems as though there are two main types of metadata: OS dependent and user dependent. the OS dependent metadata is information important to the system (permissions, ownership, filesize, etc). everything else (file type, etc) would be user dependent.
i'm wondering if there would be a way to have the OS dependent data remain static (as it usually is), but have the user dependent data be variable. for instance, you could have a 'type' field, but you might also want an 'open', 'print', or 'edit' field to specify perhaps which applications would be responsible for those actions.
for example, i might view images in GQView, but i can't edit them in that program. so, i'd create a new metadata field called 'edit' that would maybe point to the Gimp as the editor for that file. as long as the OS and/or programs i used adhered to this, life would be simple.
since the user dependent metadata would be variable, not all files would need uneccessary metadata fields. but for those files where the end-user might want a broader range of actions associated with the file, many different possibilites could exist.
of course, file-type information would not have to be the only thing stored in metadata. you could have version information, author information, authoring program, etc... but the metadata would only include that which the user would want included.
just an idea...
why is it that everything has to be billed as opensource? like that's really a selling point? i mean sure, there are some people that are just looking into the opensource thing, but it makes it seem as though that's the only reason you might want to run FreeBSD. never mind the fact that it's stable, fast, and has many of the popular apps that linux does. plus it includes linux emulation so that you can run native linux apps under FreeBSD. why not have a title that shows this in addition to it's being opensource? "FreeBSD: The opensource OS for your PC that's fast, versatile, and dependable". Sounds much more catchy to me.
one thing i have noticed about Sun hardware as compared to Intel hardware is that the Sun stuff will take quite a load without crumbling under the pressure. my experience has been that whether you've got 1 user or 100 users on the system, the Sun stuff pretty much runs the same. with Intel hardware, your results will vary, especially among the different OS's, but overall, i don't think Intel hardware handles a heavy load as well as Sun equipment.
wouldn't it be possible to use lex/flex to generate a C scanner that could parse the data just as easily as perl? C by itself may not be the answer, but there are some tools that use C that can make life a lot easier.
one problem: you are using the assumption that l33t actually owes you an answer. he doesn't. by saying RTFM, he's probably just trying to brush you off. if you have spent any time in an IRC support channel, you'll see many people parroting the same question over and over. RTFM is a way of saying "yes, i heard you... there is some documentation out there... find it for yourself and leave us alone". this is especially true for common problems (like setting up PPP) where large volumes of information exist on the specific problem.
if you are new to X, the best way to find information about X would probably be to start with a search engine. then, if you truly cannot find any information, ask "anyone know where i can get some info on X?". and do not expect a reply. if you don't get an answer, just hop to another IRC server. and if you never get an answer... well, sorry. that's just the way it is. just because you logged into the IRC server and found the support channel for X doesn't mean you are actually entitled to support (unless you specifically paid for it).
what about setting up a system of multiple sets of root servers that are able to handle requests for different DNS networks? the same kind of idea can be seen with IRC (Undernet vs. Efnet vs. Dalnet vs. etc... -- all are IRC networks, but each is seperate). Perhaps you could modify the way that resolvers work to allow for a new DNS network convention. URL examples:
http://user:pass@ICANN!slashdot.org:80/
-or-
http://user:pass@NEWDNS!hostname.tld:8080/
using a system such as this would allow for users to select which DNS network they want to use for resolving name collisions.
of course, something like this would require modifications to the current name resolution system, but i think that with a little time and effort, implementing a system like this could work.
just an idea...
doh... express.com has Disney listed as the copyright holder (which makes sense), but yep, if you look at the enlarged picture of the back, there is the miramax logo.
my friends told me it was a Disney release and got me all worked up over nothing. now they must be punished >0
i've talked with a few people who have seen the japanese version (import), but have not seen the US version (it didn't come to any theaters around my area). they've talked about the fact that since this is a Disney release (not Miramax), then it might be edited for content. my friends mention one scene that involves a sword, blood, and a disembodied head.
so, to anyone who has seen both versions, is the US release the same? or watered down for the kiddies (well, as watered down as PG-13 gets anyways)?
yes, and they also use them for green (and yellow possibly??). there is an intersection next to my apartment complex that has one of these newer light systems installed. they are BRIGHT (especially the green for some reason)! there is definately no mistaking what color the light is. it even hurts my eyes at night sometimes!
A worthy idea, but the implementation is wrong.
Network apps do not query DNS servers directly (typically, though there are exceptions). Most simply make calls to the OS's resolver, which then forwards requests to the primary DNS server, which then queries a root DNS server.
So, adding code into Mozilla to use alternate root servers would simply be a waste of time and space.
Hacking BIND would be the way to go. You could have bind check to see which kind of TLD is being requested (official or alternate) and then have it query whichever root server. However, the same problems that are associated with having an alternate DNS system are still present (collisions, etc...).
this is a topic that i covered in my sociology class in college. basically what was said is that we (as a society) today are trading in our authority to allow the community to regulate the activities of our children rather than take a pro-active stance and do it ourself. You can see many examples of this with after-school programs and community programs that are there to 'keep your kids out of trouble'. It is quite interesting to see this in action.