maybe its just because i'm used to windows but i find that when linux works out of the box it does so butifully.
when your hardware is for whatever reason not properly supported out of the box you can't just grab a makers driver cd and install the driver you need like you can on windows. If there is a linux driver at all it will most likely require figuring out where to get headers for your kernel and then hoping that the module will build ok against your kernel version.
on the other hand linux does support a lot more hardware than windows out of the box. Installing windows on a system without the original driver cds for all the parts is a nightmare.
also there are specific classes of hardware with which linux really sucks. For example wireless network cards and usb adsl modems (i know you can use a home router with built in adsl modem but then you are stuck behind a cheap and nasty nat) you can get a few home routers that can terminate the pppoa whilst still providing the real ip to the box connected by ethernet but theese are rare and some of them also seem to not get on very well with the way linux handles dhcp.
Emm, here in the UK the digital cable services have been around for five years and they are over IP. do you have a reputable source for this? it sounds like bs to me. Why would they use something as complex as ip multicasting when what they are sending is broadcast tv?
sure the internet is ip that doesn't mean that everything that runs over the same cables has to be ip.
1: oracle support didn't know that the feature had been dropped 2: there was nothing that support could do about the issue anyway even once they admitted it.
the first one is definately an issue with the support department the second is the main issue with propietry software, if the vendor makes a descision you don't like you can be screwed.
every end site is supposed to get a/48 and some isps (for example xs4all) do actually provide them.
the trouble with a/64 is that it means you have the choice of either only one subnet or using a non-standard alocation setup (do operating systems even let you do this?) which will prevent use of normal stateless autoconfiguration setups.
ofc some isps may deliberately cripple accounts for home users.
btw 6to4 gives you a/48 for every public ipv4 ip by design
yes but hardware tends to come with usable windows drivers in the box.
very little hardware comes with linux drivers in the box and the whole way linux is designed discourages drivers that don't come with the kernel. (no binary compatibility between versions,you have to have just the right headers to compile a module for your existing kernel or start from scratch and build a new kernel etc)
sure linux comes with a lot of drivers but its not very friendly to third party driver distribution.
su type systems have always seemed vulnerable to me.
if you are running under a user account you can almost certainly find a way to trap that users next attempt to use a su like tool to run something with higher privilages
once you have done that you have the admin password and the run of the systems.
iirc the reason they overloaded it on the floating point registers was to allow it to work without support from the OS.
adding extra registers would have meant that the feature would have to be explicitly enabled by a supporting OS that new how to context switch the new registers and it would have been unusuable without such support.
the big thing thats difficult about the linux way is essentially you have several choices none particularlly nice.
1: use a stable distro and put up with the versions of software it ships with.
2: use a development version of a distro and put up with the continuous upgrades and possibility that something will break.
3: use a stable distro along with third party repositries that provide backported packages for it. The main problem here is limited choice of backported packages.
4: use a stable distro along with packages from the upstream for software you care about. This works fine if the do one for your version of your distro but can otherwise be hell.
5: use a stable distro and compile stuff you care from source manually. This is usually possible with most software but there may still be problems with library and compiler versions etc especially as your distro gets older.
in contrast on windows its usually just a case of grab installer and go even on windows versions as old as win95!!
lots of people regard debian woody as prehistoric and its often hard to make recent software run on it, yet it is newer than windows XP!
meanwhile in the windows world recent software will often work just fine on versions right down to windows 95.
see the issue yet? linux is a huge forced upgrade treadmill that shows no real sign of stopping. The problem is application developers develop on systems running development versions of distros (debian unstable gentoo etc) and so others are forced to stay fairly close to this edge or they can't use the apps.
it seems to me that both the original poster and the response were at 50% informative 50% offtopic.
i think the kind of people who are likely to mod down are the kind that browse at +2 or +3 so if someone mods an offtopic post up to that level (especially +3 and above) or it gets there automatically becase of karma bonus its likely to get modded down
if you are making offtopic posts then use the no karma bonus option to reduce the visibility of your posts to those likely to mod down.
Re:Zero-innovation: perhaps a good thing?
on
McVoy Strikes Back
·
· Score: 1
it should be noted that the original gnutella despite the name was NOT opensource in some ways i think this is what made things so vibrant you had several gnutella clones on the same network all competing on features etc
hehe whats wrong with saying "you can't slashdot my website"?
really/.ings are overrated and if you are prepared (either because you posted the link yoruself or you got wind of is ahead of time as a subscriber or a user of a certain mirror site;) ) you are unlikely to fall unless your hosting really sucks.
the reasons people fall to slashdottinlgs: 1: really poor hosting (modem dsl very limited shared hosting account etc) 2: large downloads (but if your prepared you can make them torrents or whatever) 3: dynamic websites that have a high server load cost per request (this honestly seems to make up the bulk of/.ings) 4: low traffic limits (to some extent this comes under the shitty hosting category)
one way would be to have some kind of rolling password system with the current code given either on a display (visible only inside the coffee shop) or on receipts
i'm not sure bandwidth is really the issue here anyway.
yes with client-server games you will generally be fine the issues come with games that use peer-peer techiniques and/or where its normal to host games on a whim (unlike most first person shooters where its normal to leave gameservers running and connect to them).
if you get 6to4 set up then you don't need a tunnel as such. 6to4 is totally stateless!
basically heres the way it works
your system has an ipv6 default gateway of 2002:c058:6301::
this is a 6to4 address so packets routed to it get wrapped in a 6to4 packet and sent to 192.88.99.1
this is an anycast address which takes the packet to a 6to4 relay router that will take it onto the ipv6 network (mostly theese seem to be run by research organisations right now but as 6to4 grows in popularity isps will find themselves under pressure to run thier own)
on the way back the packets find thier way to a 6to4 relay router through normal ipv6 routing and from there will be wrapped and sent to your ipv4 address.
well nat has the disadvantage over a more traditional firewall that you can't just say open port xyz to all systems so app abc can work you havw to give machines fixed lan ips and then make mappings to each one individually using different ports.
frankly i've always belived that firewalls are a kludge anyway the whole point of the internet is to be a network of computers accross the world. If your app can't live securely in that environment then imo it shouldn't be using ip in the first place.
there are tunnel brokers but finding a reliable one is hard
and there is 6to4 but that really needs the machine doing it to have a public ip or possiblly a router dmz setting (basically it needs to be able to get 6to4 packets which are neither tcp nor udp from the 6to4 relays serving the hosts you are communicating with).
thats bad math at least if the worm writer has a clue (i considered checking if that was what you had done but i'm too tiered to do the numbers right now)
all addresses on the ipv6 internet currently are in one of 3/16 blocks
2001::/16 production ipv6 internet 2002::/16 6to4 stateless ipv6 over ipv4 3FFE::/16 6bone experimental ipv6
threedegrees sets up teredo which is similar to 6to4 but works behind nat. It also sets up the system for 6to4 if its not behind nat but a large proportion of desktop systems nowadays have ended up behind nat.
unfortunately unlike with 6to4 teredo isn't yet an approved standard and does not have in place the relays needed to interact with the rest of the ipv6 internet so its only really usefull for connecting between systems on ipv4 right now even though it uses the ipv6 apis
for ipv6 to really take off imo we need to see ipv6 support including 6to4 in consumer routers 6to4 cannot be done by the computers behind theese routers as it can't be done behind nat.
the fact is a complete switch to ipv6 WILL to all practical perposes make tradidional net-scanning worms of this type an unworkable way of spreading. Some may view this as security though obscurity but then doesn't that apply to passwords etc as well after all a password is just an obscure cobination you use to gain access to a system.
i'm not sure how the gp got the figure of 10^-28% but the figure is still so small that a worm could hit random addresses for a very very long time before having a reasonable chance of hitting anything.
Slashdot Mirror
maybe its just because i'm used to windows but i find that when linux works out of the box it does so butifully.
when your hardware is for whatever reason not properly supported out of the box you can't just grab a makers driver cd and install the driver you need like you can on windows. If there is a linux driver at all it will most likely require figuring out where to get headers for your kernel and then hoping that the module will build ok against your kernel version.
on the other hand linux does support a lot more hardware than windows out of the box. Installing windows on a system without the original driver cds for all the parts is a nightmare.
also there are specific classes of hardware with which linux really sucks. For example wireless network cards and usb adsl modems (i know you can use a home router with built in adsl modem but then you are stuck behind a cheap and nasty nat) you can get a few home routers that can terminate the pppoa whilst still providing the real ip to the box connected by ethernet but theese are rare and some of them also seem to not get on very well with the way linux handles dhcp.
Emm, here in the UK the digital cable services have been around for five years and they are over IP.
do you have a reputable source for this? it sounds like bs to me. Why would they use something as complex as ip multicasting when what they are sending is broadcast tv?
sure the internet is ip that doesn't mean that everything that runs over the same cables has to be ip.
well it seems he had two main issues
1: oracle support didn't know that the feature had been dropped
2: there was nothing that support could do about the issue anyway even once they admitted it.
the first one is definately an issue with the support department the second is the main issue with propietry software, if the vendor makes a descision you don't like you can be screwed.
every end site is supposed to get a /48 and some isps (for example xs4all) do actually provide them.
/64 is that it means you have the choice of either only one subnet or using a non-standard alocation setup (do operating systems even let you do this?) which will prevent use of normal stateless autoconfiguration setups.
/48 for every public ipv4 ip by design
the trouble with a
ofc some isps may deliberately cripple accounts for home users.
btw 6to4 gives you a
yes but hardware tends to come with usable windows drivers in the box.
very little hardware comes with linux drivers in the box and the whole way linux is designed discourages drivers that don't come with the kernel. (no binary compatibility between versions,you have to have just the right headers to compile a module for your existing kernel or start from scratch and build a new kernel etc)
sure linux comes with a lot of drivers but its not very friendly to third party driver distribution.
su type systems have always seemed vulnerable to me.
if you are running under a user account you can almost certainly find a way to trap that users next attempt to use a su like tool to run something with higher privilages
once you have done that you have the admin password and the run of the systems.
this isn't correct either
they must either distribute the source and the binary together OR include a written offer with the binaries to give the source to anyone.
iirc the reason they overloaded it on the floating point registers was to allow it to work without support from the OS.
adding extra registers would have meant that the feature would have to be explicitly enabled by a supporting OS that new how to context switch the new registers and it would have been unusuable without such support.
http://www.networkmirror.com/cPx3D_npNfe2mF7p/www. sevendaysvt.com/features/2005/working_his_connecti ons/index.html
the big thing thats difficult about the linux way is essentially you have several choices none particularlly nice.
1: use a stable distro and put up with the versions of software it ships with.
2: use a development version of a distro and put up with the continuous upgrades and possibility that something will break.
3: use a stable distro along with third party repositries that provide backported packages for it. The main problem here is limited choice of backported packages.
4: use a stable distro along with packages from the upstream for software you care about. This works fine if the do one for your version of your distro but can otherwise be hell.
5: use a stable distro and compile stuff you care from source manually. This is usually possible with most software but there may still be problems with library and compiler versions etc especially as your distro gets older.
in contrast on windows its usually just a case of grab installer and go even on windows versions as old as win95!!
lots of people regard debian woody as prehistoric and its often hard to make recent software run on it, yet it is newer than windows XP!
meanwhile in the windows world recent software will often work just fine on versions right down to windows 95.
see the issue yet? linux is a huge forced upgrade treadmill that shows no real sign of stopping. The problem is application developers develop on systems running development versions of distros (debian unstable gentoo etc) and so others are forced to stay fairly close to this edge or they can't use the apps.
i'm pretty sure you can set the runlevel by passing an option on the kernel comand line in your bootloader.
it seems to me that both the original poster and the response were at 50% informative 50% offtopic.
i think the kind of people who are likely to mod down are the kind that browse at +2 or +3 so if someone mods an offtopic post up to that level (especially +3 and above) or it gets there automatically becase of karma bonus its likely to get modded down
if you are making offtopic posts then use the no karma bonus option to reduce the visibility of your posts to those likely to mod down.
it should be noted that the original gnutella despite the name was NOT opensource in some ways i think this is what made things so vibrant you had several gnutella clones on the same network all competing on features etc
hehe whats wrong with saying "you can't slashdot my website"?
/.ings are overrated and if you are prepared (either because you posted the link yoruself or you got wind of is ahead of time as a subscriber or a user of a certain mirror site ;) ) you are unlikely to fall unless your hosting really sucks.
/.ings)
really
the reasons people fall to slashdottinlgs:
1: really poor hosting (modem dsl very limited shared hosting account etc)
2: large downloads (but if your prepared you can make them torrents or whatever)
3: dynamic websites that have a high server load cost per request (this honestly seems to make up the bulk of
4: low traffic limits (to some extent this comes under the shitty hosting category)
one way would be to have some kind of rolling password system with the current code given either on a display (visible only inside the coffee shop) or on receipts
i'm not sure bandwidth is really the issue here anyway.
yes with client-server games you will generally be fine the issues come with games that use peer-peer techiniques and/or where its normal to host games on a whim (unlike most first person shooters where its normal to leave gameservers running and connect to them).
if you get 6to4 set up then you don't need a tunnel as such. 6to4 is totally stateless!
basically heres the way it works
your system has an ipv6 default gateway of 2002:c058:6301::
this is a 6to4 address so packets routed to it get wrapped in a 6to4 packet and sent to 192.88.99.1
this is an anycast address which takes the packet to a 6to4 relay router that will take it onto the ipv6 network (mostly theese seem to be run by research organisations right now but as 6to4 grows in popularity isps will find themselves under pressure to run thier own)
on the way back the packets find thier way to a 6to4 relay router through normal ipv6 routing and from there will be wrapped and sent to your ipv4 address.
well nat has the disadvantage over a more traditional firewall that you can't just say open port xyz to all systems so app abc can work you havw to give machines fixed lan ips and then make mappings to each one individually using different ports.
frankly i've always belived that firewalls are a kludge anyway the whole point of the internet is to be a network of computers accross the world. If your app can't live securely in that environment then imo it shouldn't be using ip in the first place.
by make support a nightmare i mean supporing any app that wan'ts to use a tcp listen port
you have to teach the user what nat and port forwards are and try to help them use a router control interface that you have never seen!
then you find you have to fix ips on your lan and listen ports in apps and create a whole load of port forward rules to make anything else
yeah sure they work out of the box but only for people who thing the internet is the interweb and they make support a nightmare
there are methods such as stun for getting around some of theese issues but they are pretty nasty for app coders and therefore not widely adopted.
there are tunnel brokers but finding a reliable one is hard
and there is 6to4 but that really needs the machine doing it to have a public ip or possiblly a router dmz setting (basically it needs to be able to get 6to4 packets which are neither tcp nor udp from the 6to4 relays serving the hosts you are communicating with).
thats bad math at least if the worm writer has a clue (i considered checking if that was what you had done but i'm too tiered to do the numbers right now)
/16 blocks
all addresses on the ipv6 internet currently are in one of 3
2001::/16 production ipv6 internet
2002::/16 6to4 stateless ipv6 over ipv4
3FFE::/16 6bone experimental ipv6
threedegrees sets up teredo which is similar to 6to4 but works behind nat. It also sets up the system for 6to4 if its not behind nat but a large proportion of desktop systems nowadays have ended up behind nat.
unfortunately unlike with 6to4 teredo isn't yet an approved standard and does not have in place the relays needed to interact with the rest of the ipv6 internet so its only really usefull for connecting between systems on ipv4 right now even though it uses the ipv6 apis
for ipv6 to really take off imo we need to see ipv6 support including 6to4 in consumer routers 6to4 cannot be done by the computers behind theese routers as it can't be done behind nat.
the fact is a complete switch to ipv6 WILL to all practical perposes make tradidional net-scanning worms of this type an unworkable way of spreading. Some may view this as security though obscurity but then doesn't that apply to passwords etc as well after all a password is just an obscure cobination you use to gain access to a system.
i'm not sure how the gp got the figure of 10^-28% but the figure is still so small that a worm could hit random addresses for a very very long time before having a reasonable chance of hitting anything.
do you have a real ip on the mac or are you behind nat?
6to4 needs to be done by a machine with a real ipv4 ip (you can have other machines behind the 6to4 box though)