I've often wonders about p2p text messaging, not for anything noble but just because it seams that certain communities would be great candidates for it (schools, universities, festivals, etc). There are many problems including: 1)Battery life, almost all phones switch of wifi/bluetooth asap to save battery life, which on smart phones is apparently fairly short anyway. To constantly have wifi/bluetooth active would reduce the idle life of most phones so much most people would disable this feature. 2)Carriers wouldn't like it, sending text messages across a room via bluetooth seams fairly trivial, however AFAIK (my current phone is a cheap POS) its much easier to send MMS than SMS because few phones support receiving text, without 3rd party apps. Either this is down to bickering over standards or more likely the carriers putting pressure on phone makers not to implement/standardize such features. 3)Range and complexity, even in ideal situations (high density, high diffusion), the range of wifi/bluetooth is still fairly limited. And as neither are particularly great at handling walls, you would end up with large subgroups of completely disconnected users, and its not that useful to only be able to communicated with people in the room. The way around the separation problem is to buffer and send messages when other users are in range, however this increase complexity & starts introducing possible security problems and DOS attacks. 4)Chicken and egg, this being useful relies on it being installed on most phones, but being installed on most phones relies on it being useful (even ignoring carrier interference)
While 1,3 can be solved technically and 2 will become a moot point as carriers allow smart phones, 4 is the real problem, especially as there is little demand for yet another way of sending text between phones. While fellow/.ers may appreciate the beauty of open-p2p phone networks, most normals wont give a damn.
p.s encryption would be fairly trivial to implement as you could exchange keys when physically near each-other and/or due to the mesh/p2p nature making consistent MITM attacks tricky(especially as if one pear slows down to re-encrypt messages a faster route will probably be found) sign all outgoing messages and encrypt replies (ofc you wouldn't want to send anything sensitive until you trust the keys, probably achieved by standing next to each-other and sending a message)
suddenly imposing economic sanctions on families with >1 child doesn't seam so harsh (in china anyway), it should also be noted that population growth is much less of an issue inn developed countries and curing diseases would help lots of developing countries out.
switched from debian lenny to fedora and the biggest PITA was defenitly pulseaudio, and thats coming from somebody who cant use yum/rpm or selinux for shit.
while agree that PulseAudio is not ready for the primetime: 1) the three APIs are not that incompatible as they always include legacy modes for old apps. 2) I never figured out how to stop audio playing if a 2nd user logged in with alsa, but it happened by default with PA.
IMH(umble)O it would have been better for PA's features to be implemented as scripts around ALSA, but those doing the work thought differently and as im too stupid and/or lazy to do it myself, I have to live with a slightly broken PA until its finished!
yeah it had nothing to do with BMPx no longer being maintained! as for xmms i suspect that gtk1.x is no longer being maintained. Canonical can't be arsed to maintain a few old winamp style players when there are many other well maintained ones about:O
Yeah either that or only just fucking heard about this DOS and its gone on the bottom of a fairly long todo list. And while its a fairly lame attack its down to a minor design flaw, so it will require more work than GP suggests. I also get the impression that the problem is that Apache is trying to do too much, prevent the server demanding too much from the OS by limiting processes, when the OS should prevent Apache from doing that itself, but hey I'm the sort of noob that still thinks DDOS/DOS is a firewall problem not a server problem.
While i agree that its a by choice that IIS (rather than fluke) is immune to this, i disagree that the worker-pool threading model is superior. It has advantages in terms of performance (somewhat limited on a modern os) but is both more complex and less secure (theoretically anyway, although implementation (see complexity) counts more)).
even copyrighted tv-shows don't (same video was taken down on youtube lol). why i pirate: *it makes me a pirate and pirates are cool (i wouldn't be half as happy about smoking weed if it was legal) *I don't feel sorry for the "victims" (record label execs, big rockstars loosing a few cents) *It's easier, enter credit card details vs click link *i don't have the money for an 20gb music collection (if it wasn't 20gb then most of the artists wouldn't of had me go see them) *too many good artists are ruined by money when they make it big (i still pay for small bands stuff, but im probably doing a favour as you can't right lyrics about having a hard life if your fucking loaded!) *copyright law is broken (0 is closer to what it should be 14-25 than life+50)
As somebody who has smashed up laptops, I think tis harder than you'd imagine to damage your HDD in a drop, i've abused the crap out of 2.5 acers and apart from some problems if i move the laptop too suddenly when its on, I've been fine*!
well there was the issue that the CD drive fell out once when i droped it, but the store full of muppets i bought it from where probably to blame for that and the dodge mobo i had.
since when has ext been the best choice for anything, ext has always been about balance i doubt its the best choice for SSD, Id put my money on a log filesystem, e.g you couldn't be more wrong and GP is correct because NILFS2 will write to used blocks much less often than conventional systems. OFC ext will be better than FAT because file-allocation table block is going to be a problem and it turns out ext4 with COW will also be good (but not as good as a log system and the journal itself will be a problem)
preload and readahead speedup firefox startup times significantly, 3.5 is pretty quick here but i have taken care with the extensions i use (no tmp,fasterfox,etc)
yes, as with 3 the memory footprint is significantly reduced, its still a bit on the high side here 128mb (the most any one program uses) but i have plenty of ram (~1.5Gb) so much of that may be features instead of leaks.
BT can charge an arm and a leg for somethings though it was ~£100 to reconnect my line when i moved into a flat and that was even if the physical line was still usable, so i went with virgin.
2MB/s to EVERYBODY is a big step, a networks value depends on the number of people connected to it, 50p/month in exchange for a much more valuble internet is a great deal, those looking for faster connections should pay for it themselves as once everybody has access to "broadband content" iplayer/etc the incremental advantage of giving them 10mb or more is very low
but the problem with that, is that you then have to embed the encryption key into the client executable (or one of its libraries), which a hacker could then extract?
that's why drm is fundamentally flawed, however the same can be said for closed source programs. While I'm not an expert on licensing and GPL3 may forbid keeping keys secret there are many licenses that would allow you to hide your keys and require memory dumps and the same tricks that are used for getting keys from closed products, while still having a fundamentally open engine.
The signed stack is just so that you only have to trust 1 executable, instead of trusting that openGL isn't hacked to not render wall, I think most programs simply trust that your openGL/driecctX/audio/network/etc will do what they are told.
Weather a game is open/closed is secondary to the security measures they take, IMO closed games have been getting away (sometimes well, othertimes not so much) with a very poor security model by hiding it entirely. And because no game can ever be perfect (they have to trust they are talking the correct executable), it will always be down to server admins to protect gamers from cheats anyway.
It's tricky but i fail to see how getting major distros to sign their openGL/audio stacks, then requiring the clients to verify they are using a valid stack, is impossible? And you can also verify the client is valid by using a small 3rd party executable, like closed source games do. yeah its not perfect as you either need to close some of the source (the verifying software) or just accept that no protection is perfect and hopefully raise the bar above that of HL/whatever game cheats love playing at the moment.
. Additionally many cheats can be caught out, and subsequently banned by using secure servers, with dumb clients (speedhack,aimbot,etc are easy to see server side) and a lot of other stuff is virtually undetectable (ESP,wallhacks,etc) by anything other than a human admin, with a feel for the game and experience in catching cheats.
Not if you do it right, require a signed stack and other DRM-like tricks and your at least as safe as a closed source game. Open source means you can't hide your server side deficiencies (trusting the client) behind client side code, but at the end of the day between decompiers are reverse engineers the bulk of your code will be out there anyway. TBH whatever you do your going to get cheats and hackers (although speed hacks are down to poor server-side code), I'm an FPS guy and whether your paying HL(w/ VAC and a) or assaultcube the only thing that protects you from cheats are good server admins and id wager a large amount that the same can be said of MMOs
There are Repositories that are pretty fresh, fedora/arch/etc however there is a damn good reason that fewer people run fedora/arch/etc instead of debian/ubuntu/etc. (typing this from fedora 11). If novel want to show of OSS they will have to patch/stabilise/maintain those apps otherwise you end up with a store full of crappy unstable apps that don't work. Now novell could do something clever like allow developers to upload thier latest versions but have the default install stick to stable stabilized apps, however this is a PITA to maintain and would require much more work than either a crappy app store (ala apple) or an outdated repository (ala debian)
Slashdot Mirror
I've often wonders about p2p text messaging, not for anything noble but just because it seams that certain communities would be great candidates for it (schools, universities, festivals, etc). There are many problems including:
1)Battery life, almost all phones switch of wifi/bluetooth asap to save battery life, which on smart phones is apparently fairly short anyway. To constantly have wifi/bluetooth active would reduce the idle life of most phones so much most people would disable this feature.
2)Carriers wouldn't like it, sending text messages across a room via bluetooth seams fairly trivial, however AFAIK (my current phone is a cheap POS) its much easier to send MMS than SMS because few phones support receiving text, without 3rd party apps. Either this is down to bickering over standards or more likely the carriers putting pressure on phone makers not to implement/standardize such features.
3)Range and complexity, even in ideal situations (high density, high diffusion), the range of wifi/bluetooth is still fairly limited. And as neither are particularly great at handling walls, you would end up with large subgroups of completely disconnected users, and its not that useful to only be able to communicated with people in the room. The way around the separation problem is to buffer and send messages when other users are in range, however this increase complexity & starts introducing possible security problems and DOS attacks.
4)Chicken and egg, this being useful relies on it being installed on most phones, but being installed on most phones relies on it being useful (even ignoring carrier interference)
While 1,3 can be solved technically and 2 will become a moot point as carriers allow smart phones, 4 is the real problem, especially as there is little demand for yet another way of sending text between phones. While fellow /.ers may appreciate the beauty of open-p2p phone networks, most normals wont give a damn.
p.s encryption would be fairly trivial to implement as you could exchange keys when physically near each-other and/or due to the mesh/p2p nature making consistent MITM attacks tricky(especially as if one pear slows down to re-encrypt messages a faster route will probably be found) sign all outgoing messages and encrypt replies (ofc you wouldn't want to send anything sensitive until you trust the keys, probably achieved by standing next to each-other and sending a message)
suddenly imposing economic sanctions on families with >1 child doesn't seam so harsh (in china anyway), it should also be noted that population growth is much less of an issue inn developed countries and curing diseases would help lots of developing countries out.
switched from debian lenny to fedora and the biggest PITA was defenitly pulseaudio, and thats coming from somebody who cant use yum/rpm or selinux for shit.
while agree that PulseAudio is not ready for the primetime:
1) the three APIs are not that incompatible as they always include legacy modes for old apps.
2) I never figured out how to stop audio playing if a 2nd user logged in with alsa, but it happened by default with PA.
IMH(umble)O it would have been better for PA's features to be implemented as scripts around ALSA, but those doing the work thought differently and as im too stupid and/or lazy to do it myself, I have to live with a slightly broken PA until its finished!
yeah it had nothing to do with BMPx no longer being maintained! as for xmms i suspect that gtk1.x is no longer being maintained. Canonical can't be arsed to maintain a few old winamp style players when there are many other well maintained ones about :O
Yeah either that or only just fucking heard about this DOS and its gone on the bottom of a fairly long todo list. And while its a fairly lame attack its down to a minor design flaw, so it will require more work than GP suggests. I also get the impression that the problem is that Apache is trying to do too much, prevent the server demanding too much from the OS by limiting processes, when the OS should prevent Apache from doing that itself, but hey I'm the sort of noob that still thinks DDOS/DOS is a firewall problem not a server problem.
While i agree that its a by choice that IIS (rather than fluke) is immune to this, i disagree that the worker-pool threading model is superior. It has advantages in terms of performance (somewhat limited on a modern os) but is both more complex and less secure (theoretically anyway, although implementation (see complexity) counts more)).
Their a university, they could just start teaching cobal again!
why not just run a course on the old language then employ one of your graduates to maintain it?
even copyrighted tv-shows don't (same video was taken down on youtube lol).
why i pirate:
*it makes me a pirate and pirates are cool (i wouldn't be half as happy about smoking weed if it was legal)
*I don't feel sorry for the "victims" (record label execs, big rockstars loosing a few cents)
*It's easier, enter credit card details vs click link
*i don't have the money for an 20gb music collection (if it wasn't 20gb then most of the artists wouldn't of had me go see them)
*too many good artists are ruined by money when they make it big (i still pay for small bands stuff, but im probably doing a favour as you can't right lyrics about having a hard life if your fucking loaded!)
*copyright law is broken (0 is closer to what it should be 14-25 than life+50)
As somebody who has smashed up laptops, I think tis harder than you'd imagine to damage your HDD in a drop, i've abused the crap out of 2.5 acers and apart from some problems if i move the laptop too suddenly when its on, I've been fine*!
well there was the issue that the CD drive fell out once when i droped it, but the store full of muppets i bought it from where probably to blame for that and the dodge mobo i had.
How do file-systems play into this? if you use a log file system (NILFS etc) would that make random access time irrelevant?
since when has ext been the best choice for anything, ext has always been about balance i doubt its the best choice for SSD, Id put my money on a log filesystem, e.g you couldn't be more wrong and GP is correct because NILFS2 will write to used blocks much less often than conventional systems. OFC ext will be better than FAT because file-allocation table block is going to be a problem and it turns out ext4 with COW will also be good (but not as good as a log system and the journal itself will be a problem)
preload and readahead speedup firefox startup times significantly, 3.5 is pretty quick here but i have taken care with the extensions i use (no tmp,fasterfox,etc)
yes, as with 3 the memory footprint is significantly reduced, its still a bit on the high side here 128mb (the most any one program uses) but i have plenty of ram (~1.5Gb) so much of that may be features instead of leaks.
if your so happy about dillo and lynx why did you give enough of a shit about an alternative post?
open a bug report?
dunno about xmarks, but I've been using http://www.andyhalford.com/syncplaces/index.html with minefield after fedora11betas had issues with firefox3.5.
BT can charge an arm and a leg for somethings though it was ~£100 to reconnect my line when i moved into a flat and that was even if the physical line was still usable, so i went with virgin.
2MB/s to EVERYBODY is a big step, a networks value depends on the number of people connected to it, 50p/month in exchange for a much more valuble internet is a great deal, those looking for faster connections should pay for it themselves as once everybody has access to "broadband content" iplayer/etc the incremental advantage of giving them 10mb or more is very low
but the problem with that, is that you then have to embed the encryption key into the client executable (or one of its libraries), which a hacker could then extract?
that's why drm is fundamentally flawed, however the same can be said for closed source programs. While I'm not an expert on licensing and GPL3 may forbid keeping keys secret there are many licenses that would allow you to hide your keys and require memory dumps and the same tricks that are used for getting keys from closed products, while still having a fundamentally open engine.
The signed stack is just so that you only have to trust 1 executable, instead of trusting that openGL isn't hacked to not render wall, I think most programs simply trust that your openGL/driecctX/audio/network/etc will do what they are told.
Weather a game is open/closed is secondary to the security measures they take, IMO closed games have been getting away (sometimes well, othertimes not so much) with a very poor security model by hiding it entirely. And because no game can ever be perfect (they have to trust they are talking the correct executable), it will always be down to server admins to protect gamers from cheats anyway.
It's tricky but i fail to see how getting major distros to sign their openGL/audio stacks, then requiring the clients to verify they are using a valid stack, is impossible? And you can also verify the client is valid by using a small 3rd party executable, like closed source games do. yeah its not perfect as you either need to close some of the source (the verifying software) or just accept that no protection is perfect and hopefully raise the bar above that of HL/whatever game cheats love playing at the moment.
.
Additionally many cheats can be caught out, and subsequently banned by using secure servers, with dumb clients (speedhack,aimbot,etc are easy to see server side) and a lot of other stuff is virtually undetectable (ESP,wallhacks,etc) by anything other than a human admin, with a feel for the game and experience in catching cheats.
I thought they just burned all their fans using copyright law, aren't we supposed to hate these guys ATM?
Not if you do it right, require a signed stack and other DRM-like tricks and your at least as safe as a closed source game. Open source means you can't hide your server side deficiencies (trusting the client) behind client side code, but at the end of the day between decompiers are reverse engineers the bulk of your code will be out there anyway. TBH whatever you do your going to get cheats and hackers (although speed hacks are down to poor server-side code), I'm an FPS guy and whether your paying HL(w/ VAC and a) or assaultcube the only thing that protects you from cheats are good server admins and id wager a large amount that the same can be said of MMOs
There are Repositories that are pretty fresh, fedora/arch/etc however there is a damn good reason that fewer people run fedora/arch/etc instead of debian/ubuntu/etc. (typing this from fedora 11). If novel want to show of OSS they will have to patch/stabilise/maintain those apps otherwise you end up with a store full of crappy unstable apps that don't work. Now novell could do something clever like allow developers to upload thier latest versions but have the default install stick to stable stabilized apps, however this is a PITA to maintain and would require much more work than either a crappy app store (ala apple) or an outdated repository (ala debian)