uh, sure they can. i'm using pop3-over-ssl to talk to one of my mail servers, and imap-over-ssl on another. i've never tried the smtp ssl support, but the option appears to be there.
regardless, why aren't you using your isp's mail server? it's there for a reason. unless there is something wrong with it, like it delays mail unreasonably or loses mail, you should be using it. if you really want a smtp log on your own machine, set up your mta to use a relay host. in any case, what does your isp blocking outgoing smtp have to do with what email client you use?
that's something that has annoyed me about the all-in-one mozilla suite. i have my middle click set to open links in a new tab. that's great for web browsing, but when i middle click on a link in an email, i get a new tab appearing in one of possibly several browser windows (each with several tabs), and i have to hunt down which one it stuck the new tab in. i'm so used to middle-clicking that i rarely remember to right click and select 'open in new window.' but i suppose with thunderbird and firebird separate, firebird could open tabs by default, and thunderbird could do the new window thing. nice.
my (unchecked) guess is that a goodly chunk of that is XPCOM and gecko, mozilla's component architecture and layout renderer, respectively. my hope is that eventually when that stuff stabilises enough we'll see a single download for the gecko runtimes, and then you can install thunderbird, firebird, etc. without having to download/install XPCOM/gecko along with the package.
then again, average-joe user generally doesn't care about download sizes and multiple copies of shared libraries - all they care about is how many clicks it takes from the website to having a usable app. perhaps they could just have a "if you don't know what you're doing, download this" package that has it all, with the smaller pieces for people that know what they want. i suppose the net-installers serve that purpose well enough, actually, as long as they can detect the presence/absence of the core libraries well enough...
i hate to be on microsoft's side, but i really can't see how you think anjuta is as good as VS. take the autocompletion, for example. anjuta will do it, but it's somewhat braindead. with VS, if you type the instance name of a c++ class, or a struct, and then type '.' or '->' (depending on if it's a pointer or not), the autocomplete box pops up with a list of all the public class members (or struct methods). anjuta just waits for you to type the first letter and then gives you a choice of any keyword you've typed so far that happens to start with that letter. that's useful to be sure, but sometimes, say when i'm doing initialisation, i want to see the list of members so i know i'm getting everything.
anjuta also seems to be hit-or-miss when displaying function prototypes after you hit the open parenthesis after typing a function name. sometimes it displays, sometimes not. and if the function prototype is being displayed, the autocomplete box won't pop up. plus, if you're in say PHP mode, and type "fopen(" you get the C prototype, which isn't too useful. granted, these are all just bugs, not design issues, and i presume they'll be fixed by the time the gtk2 version is considered stable.
and here i've only touched on the text editor itself - i don't think the other IDE features (project handling, building, debugging, configuring, etc.) are anywhere near as mature as VS's implementation is.
i like anjuta, i really do - i use it for all my linux dev work. but VS just seems to work better and feel more useful. i really do look forward to where the anjuta team is going with it, though. i'm quite confident in their ability to someday surpass VS. i just don't think it's quite there yet.
great, so what am i going to do with the B.S. in electrical engineering that i'm getting in january? my family and i have spent over $100k so i could get this degree and get a good job, and you're saying i missed the boat? looks like i'm going to end up with the wrong kind of B.S....
seriously, i'm a hybrid EE/CS guy, and it's what i love. but i'm not a supergenius. sure, i go to a damned good university, but it's not MIT, and i'm not exactly a stellar student. are you saying that while i've been sitting here getting an education, that for which i've been training has been slowly disappearing to the point that my hard work is going to be all for nothing?
while i agree that they are confusing and lame, you might want to check out this site before you attempt to "educate" others about your supposed knowledge. google is your friend.
oh c'mon, that's silly. you could also make the argument that because you sent a file to your friend using scp/sftp, the file wasn't recogniseable as copyrighted material while in transit and thus you aren't committing copyright infringement. saying that applying some reversible algorithm to data to "ensure that no single block has copyright content" will protect you from copyright infringement suits is ridiculous. it's merely a form of encryption - and the one you describe is fairly weak at that.
now, if your point was to show that networks could be constructed such that it's incredibly difficult to determine what is being transmitted and from where, then ok. that's possible. we have freenet for that. but it sounds like you are leaning toward proposing that if you obfuscate the contents of a file enough and split it into enough pieces, all coming from different places, it somehow ceases being a copyrighted work. that's just plain rubbish.
in any case, you still need a method to _find_ the file you want in the first place, which is of course the hard part. if there is an indexing/search service, that's an easy vulnerability. if you have some type of hashed key structure (like freenet iirc), then you have the problem of disseminating the key. granted even if authorities could get ahold of a key that points to a copyrighted work, they still can't find out where it is located or who put it there. but, lacking some form of encryption for requests, they still have the opportunity to nab those that are receiving.
regardless, a quick look at the freenet
faq makes it quite clear that even freenet is not immune to all attacks that can reveal your identity.
i'm not particularly sure where i'm going with this (quite possibly nowhere), so i'll stop now.
simple. if quick and dirty is getting you "in hot water" after the fact, and you have to spend countless hours explaining why the q&d solution can't be the final one, you're wasting precious time that you could be using working on your proper and correct solution. try to find middle ground - find the happy medium between q&d and p&c. it's there, and most often won't be the same deal for different projects.
even if you're pressured to produce something - anything - that works in a short amount of time, at least have the foresight to put thought into it and plan for the need to do a partial redesign later. after your semi-q&d solution is released, begin working on turning it into as p&c as possible immediately. then when the phbs and marketroids come after you, you at least have something tangible to 'show' them.
To the average user netscape and IE are the only two browsers in existance.
i would venture to say that the average user is even unaware of netscape these days. i'm a university student - i find that an interesting perspective since most of the people here get a new computer to bring with them to college. back when i started 4 years ago i saw netscape all over. now most of the 1st- and 2nd-years i know have their fresh windows installs and never bother to download netscape (4.x or 7.x) because IE is fine for their purposes.
there is the random anomalous person i see using netscape 4.x (almost always not on windows). i don't know anyone using netscape 7.x. fortunately most of the more tech-savvy people i know are using mozilla (or i gently recommend it every now and then).
on a related note, why use netscape 7 over mozilla? perhaps the integration is a little better, but i don't see anything essential in netscape7 that i can't get out of mozilla. plus with mozilla i get much more frequent releases and the option to grab a nightly or build my own (granted these last two options don't interest that many people). as an added bonus you lose the "netscape activation" and all that bundled AOL crap.
so i read "The One Rack" and immediately thought you were talking about a pair of... erm... melons. judging from the other replies in this thread it appears i am not a true geek... ^_~
i think a better subject for your post might have been "word choice nazi," perhaps... whether or not a word is "nonstandard" - or incorrect in formal style - has nothing to do with whether or not it is grammatical. in the grandparent's usage, 'irregardless' is perfectly grammatical.
irregardless of that fact, i'd still disagree, as/. is hardly a place where formal style is required ^_~.
/me punches himself for caring enough to post this
If I leave my front door open and go to work and come home to an empty house, would the insurance company re-imburse me for the theft or would I be neglegent? Shouldn't the patterns be thrown in a locked trash bin?
i think you're drawing the wrong kind of analogy here. the fact of the matter is that the contents of your house were stolen, and, if the thieves are caught you have the option of suing them for the value of the items stolen (i presume).
however, an insurance company's reaction to the theft has nothing to do with the law: their policy may or may not include clauses for negligence as related to reimbursement eligibility.
as for a locked trash bin, theft is still theft, regardless of the measures that were taken to prevent it (or lack thereof).
with that said, question: are there any ways to access remote winxp directories using winxp other than SMB (and any other method available on linux that just so happens to be available on windows as well)?
answer: no. there are countless methods of accessing files on a linux machine. a good number of them are available in some shape or form for windows as well. google is your friend.
sure, there is another package or two that already does this, and maybe in a "better" way - better in the realm of being more general-purpose and perhaps more forward thinking. but from other posts here (yeah, i'm getting my 'facts' from/. posts; sunk to a new low ^_~), it appears that lufs lacks the caching that has been implemented in shfs, and so doesn't perform very well. a shell-only filesystem implementation would be more likely to be optimised for that purpose (after it is presumably worked on and polished), and would in theory perform better.
have i tried either of them? no. do i intend to? yes. just saying "it's already been done elsewhere" is a lame reason for passing up on a new technology. if we all felt that way, we'd still all be windows users.
as for being "dangerous" - give me a break. the only harm i see this causing is to a user's own files. if you want to take that risk and have root on your box, go for it. any ssh server you connect to will have file permissions in place such that you can't muck up anything important, and, if not, using something like shfs on it is the least of their problems.
and as for why this merits an article on slashdot... well, i like trek as much as the next guy, but if a lego model of the enterprise warrants an article, shfs certainly does as well.
ok, i know i'm going to get flamed for this, but here goes.
first, let's ignore the legal implications - let's say that every file traded on current p2p networks is unencumbered by copyright or any other restriction. i say this because the article in quesiton seems to be doing the same - the article is not about college kids trading illegal stuff online, but it's about a certain kind of net traffic causing bottlenecks, simply because a large amount of it exists.
next, i will exempt special-purpose networks from the point i will soon make. corporations, etc., have a right to restrict their networks to company business use only. other organisations certainly have the right to do the same. colleges are in a grey area i think. while they own their networks and should be allowed to decide what constitutes acceptable use of network resources, i believe most colleges are more apt to be less restrictive.
so finally we are left with what i guess i'll call the "residential internet." all the residential broadband customers connected by whatever backbones.
so here we have this supposed "problem" that p2p networks are using up large amounts of the available bandwidth, and ISPs are hurting trying to keep up.
so finally, my point: who cares? what is the internet for? no, i'm serious. it started as a defense department research network, but it has grown far beyond that. i would venture to say that the internet is whatever its users want it for. if its users want it for p2p file sharing, so be it. if the broadband ISPs are providing "unlimited access" (excepting bandwidth caps), then they'd damn well better provide it. if they want to tighten bandwidth caps, and offer higher-price broadband packages to people that want to shell out the cash, that's their choice, but i don't see that when i look around - most ISPs (cable, anyway) seem to have at most two options - a "business" class and a normal, lower-bandwidth option (yes, i know, DSL does have multiple rate plans from a good number of providers, but i believe cable is far more prevalent).
but stop whining. broadband customers are paying for a service. broadband ISPs have a responsibility to provide that service, for whatever the customers want to use it for. instead of doing their jobs, they're trying to slap restrictions on everything.
unfortunately, my argument is weakened by the fact that the vast majority of data swapped over p2p networks appears to be copyrighted music, movies, and software. but, when you look at the ISPs' chief complaint here - bandwidth - i have no sympathy. provide what your customers want or get out of the business.
Just because MS windows has a monolithic kernel/graphics renderer it doesn't mean it isn't cobbled together behind it's surface.
um, windows doesn't have a monolithic kernel or graphics renderer. windows uses a microkernel, somewhat similar (even if superficially so) to mach/hurd. ever heard of shell32.dll? kernel32.dll? user32.dll? gdi32.dll? all user-level portions of what linux has bottled up in its kernel.
Contrary to popular belief (and, yes, contrary to my own usage patterns in college), Universities provide network resources for academic uses.
bullshit. if the university is going to provide the internet access just for "academic use," then there damned well better be an alternative for "nonacademic use." and no, i don't consider using dialup an alternative.
i am a cornell student (a senior now). i live off campus, so i'm on cable and fortunately don't have to deal with this crap. but the people on campus have no choice. two - only two - of the dorms, out of about 25-odd buildings, are wired for cable. and by cable i mean cable TV - i don't know if you could get cable internet on those (probably), but i wouldn't put it past cornell to disallow it if anyone tried and they found out.
and don't even think of suggesting that cornell allow people to get their room wired for cable, because i know that campus life (the portion of the administration that runs cornell housing) would expressly disallow any "modifications" to the buildings.
i'm sure part of the problem is the speed at which this is occurring. when i was a freshman, network fees were around $60 per SEMESTER. the next year i believe it went up to $80, and the last year there was another price hike (i think slightly over $100). but still it was a semester price. i believe that this year they're using a per-month fee, not sure about that. figure you're in the dorm for about 8 months out of the year (i know the article says 9; that's incorrect). even at $200, that's $25 a month. damned great deal. and yes, i do agree that that's amazingly cheap for the speed - at non-peak times i'd see 800kB/s. on a rare occassion i'd see 1MB/s, 4/5 of the 10baset limit (like anyone could really get 1.25MB/s). but now we're going from $200/yr, to $240/yr. not a big jump, but i would bet that the bandwidth cap will piss off way more than the 10% of users that the article suggests.
then again, i'm sitting here paying $40/month for inferior cable - 250kB/s down on a rare good day, with an upload cap of 45kB/s. 9 months, since i have to pay while i'm not here over winter break (otherwise i get to pay extra for them to reconnect it). that's what, $360? so i probably have more of a right to complain than the dorm-dwellers ^_~.
ok, so i digressed a bit from what i was originally replying to. sue me ^_~
Last time I checked, most web servers could reject persistent connections.
Speaks pretty poorly of the server (or network architecture) if your only recourse is to say "it's the client's fault!"
this isn't the same deal. based on the TCP specs, here is what a server (or client, for that matter) is supposed to do when it wants to close the connection:
1) send FIN
2) wait for ACK
3) wait for FIN
4) send an ACK
if the server never receives the FIN in step 3, it assumes that the client wants to keep the connection open for some reason. this is _correct behaviour_ with regards to the TCP spec. if this article is correct, MS is merely exploiting the TCP spec to its advantage. yes, it's dirty and wastes resources, but it works.
the thing that bothers me tho, is this is what should be happening on the server end (a non-IIS server, that is):
1) send FIN
2) wait for ACK
3) ok, got ACK, now wait for FIN
4) (after timeout) hmm, no FIN, must have been lost, so we'll resend our FIN
5) client ACKs that FIN, but doesn't send its FIN
6) server thinks the response FIN is lost again, so probably resends its FIN
now the server will have a max amount of retries before it gives up and finally drops the connection (which is what it was trying for in the first place anyway). this should be a relatively low number, and the timeouts between each retransmission shouldn't be that long either. so unless IE comes back and requests another page fairly quickly, the server _should_ go ahead and drop the connection, so i fail to see how this is a problem.
the only thing i can think of is that the client keeps responding with an ACK to the server's FINs (despite not sending its own FIN), so maybe the server won't drop the connection for that reason (since the client is obviously still alive, just not responding as expected). i don't remember the TCP spec all that clearly with regards to connection teardown, so that may be where IE is able to keep the connection open.
then again, i could be totally wrong here, but i don't think so...
Slashdot Mirror
uh, sure they can. i'm using pop3-over-ssl to talk to one of my mail servers, and imap-over-ssl on another. i've never tried the smtp ssl support, but the option appears to be there.
regardless, why aren't you using your isp's mail server? it's there for a reason. unless there is something wrong with it, like it delays mail unreasonably or loses mail, you should be using it. if you really want a smtp log on your own machine, set up your mta to use a relay host. in any case, what does your isp blocking outgoing smtp have to do with what email client you use?
that's something that has annoyed me about the all-in-one mozilla suite. i have my middle click set to open links in a new tab. that's great for web browsing, but when i middle click on a link in an email, i get a new tab appearing in one of possibly several browser windows (each with several tabs), and i have to hunt down which one it stuck the new tab in. i'm so used to middle-clicking that i rarely remember to right click and select 'open in new window.' but i suppose with thunderbird and firebird separate, firebird could open tabs by default, and thunderbird could do the new window thing. nice.
my (unchecked) guess is that a goodly chunk of that is XPCOM and gecko, mozilla's component architecture and layout renderer, respectively. my hope is that eventually when that stuff stabilises enough we'll see a single download for the gecko runtimes, and then you can install thunderbird, firebird, etc. without having to download/install XPCOM/gecko along with the package.
then again, average-joe user generally doesn't care about download sizes and multiple copies of shared libraries - all they care about is how many clicks it takes from the website to having a usable app. perhaps they could just have a "if you don't know what you're doing, download this" package that has it all, with the smaller pieces for people that know what they want. i suppose the net-installers serve that purpose well enough, actually, as long as they can detect the presence/absence of the core libraries well enough...
i hate to be on microsoft's side, but i really can't see how you think anjuta is as good as VS. take the autocompletion, for example. anjuta will do it, but it's somewhat braindead. with VS, if you type the instance name of a c++ class, or a struct, and then type '.' or '->' (depending on if it's a pointer or not), the autocomplete box pops up with a list of all the public class members (or struct methods). anjuta just waits for you to type the first letter and then gives you a choice of any keyword you've typed so far that happens to start with that letter. that's useful to be sure, but sometimes, say when i'm doing initialisation, i want to see the list of members so i know i'm getting everything.
anjuta also seems to be hit-or-miss when displaying function prototypes after you hit the open parenthesis after typing a function name. sometimes it displays, sometimes not. and if the function prototype is being displayed, the autocomplete box won't pop up. plus, if you're in say PHP mode, and type "fopen(" you get the C prototype, which isn't too useful. granted, these are all just bugs, not design issues, and i presume they'll be fixed by the time the gtk2 version is considered stable.
and here i've only touched on the text editor itself - i don't think the other IDE features (project handling, building, debugging, configuring, etc.) are anywhere near as mature as VS's implementation is.
i like anjuta, i really do - i use it for all my linux dev work. but VS just seems to work better and feel more useful. i really do look forward to where the anjuta team is going with it, though. i'm quite confident in their ability to someday surpass VS. i just don't think it's quite there yet.
great, so what am i going to do with the B.S. in electrical engineering that i'm getting in january? my family and i have spent over $100k so i could get this degree and get a good job, and you're saying i missed the boat? looks like i'm going to end up with the wrong kind of B.S....
seriously, i'm a hybrid EE/CS guy, and it's what i love. but i'm not a supergenius. sure, i go to a damned good university, but it's not MIT, and i'm not exactly a stellar student. are you saying that while i've been sitting here getting an education, that for which i've been training has been slowly disappearing to the point that my hard work is going to be all for nothing?
is it just me, or did anyone else read the headline and wonder why a bunch of guys smoking joints on their segways was newsworthy?
then again, i'm on medication now so i may not be thinking clearly...
while i agree that they are confusing and lame, you might want to check out this site before you attempt to "educate" others about your supposed knowledge. google is your friend.
do you have a source for this info, or should we just take your word for it?
yeah, i know, i could look around myself, but i'm lazy.
oh c'mon, that's silly. you could also make the argument that because you sent a file to your friend using scp/sftp, the file wasn't recogniseable as copyrighted material while in transit and thus you aren't committing copyright infringement. saying that applying some reversible algorithm to data to "ensure that no single block has copyright content" will protect you from copyright infringement suits is ridiculous. it's merely a form of encryption - and the one you describe is fairly weak at that.
now, if your point was to show that networks could be constructed such that it's incredibly difficult to determine what is being transmitted and from where, then ok. that's possible. we have freenet for that. but it sounds like you are leaning toward proposing that if you obfuscate the contents of a file enough and split it into enough pieces, all coming from different places, it somehow ceases being a copyrighted work. that's just plain rubbish.
in any case, you still need a method to _find_ the file you want in the first place, which is of course the hard part. if there is an indexing/search service, that's an easy vulnerability. if you have some type of hashed key structure (like freenet iirc), then you have the problem of disseminating the key. granted even if authorities could get ahold of a key that points to a copyrighted work, they still can't find out where it is located or who put it there. but, lacking some form of encryption for requests, they still have the opportunity to nab those that are receiving.
regardless, a quick look at the freenet faq makes it quite clear that even freenet is not immune to all attacks that can reveal your identity.
i'm not particularly sure where i'm going with this (quite possibly nowhere), so i'll stop now.
simple. if quick and dirty is getting you "in hot water" after the fact, and you have to spend countless hours explaining why the q&d solution can't be the final one, you're wasting precious time that you could be using working on your proper and correct solution. try to find middle ground - find the happy medium between q&d and p&c. it's there, and most often won't be the same deal for different projects.
even if you're pressured to produce something - anything - that works in a short amount of time, at least have the foresight to put thought into it and plan for the need to do a partial redesign later. after your semi-q&d solution is released, begin working on turning it into as p&c as possible immediately. then when the phbs and marketroids come after you, you at least have something tangible to 'show' them.
there is the random anomalous person i see using netscape 4.x (almost always not on windows). i don't know anyone using netscape 7.x. fortunately most of the more tech-savvy people i know are using mozilla (or i gently recommend it every now and then).
on a related note, why use netscape 7 over mozilla? perhaps the integration is a little better, but i don't see anything essential in netscape7 that i can't get out of mozilla. plus with mozilla i get much more frequent releases and the option to grab a nightly or build my own (granted these last two options don't interest that many people). as an added bonus you lose the "netscape activation" and all that bundled AOL crap.
i thought IANA took care of the well-known ports list. or are they a subset of/controlled by ICANN?
so i read "The One Rack" and immediately thought you were talking about a pair of... erm... melons. judging from the other replies in this thread it appears i am not a true geek... ^_~
i think a better subject for your post might have been "word choice nazi," perhaps... whether or not a word is "nonstandard" - or incorrect in formal style - has nothing to do with whether or not it is grammatical. in the grandparent's usage, 'irregardless' is perfectly grammatical.
/. is hardly a place where formal style is required ^_~.
/me punches himself for caring enough to post this
irregardless of that fact, i'd still disagree, as
however, an insurance company's reaction to the theft has nothing to do with the law: their policy may or may not include clauses for negligence as related to reimbursement eligibility.
as for a locked trash bin, theft is still theft, regardless of the measures that were taken to prevent it (or lack thereof).
i know, i know, don't feed the troll...
with that said, question: are there any ways to access remote winxp directories using winxp other than SMB (and any other method available on linux that just so happens to be available on windows as well)?
answer: no. there are countless methods of accessing files on a linux machine. a good number of them are available in some shape or form for windows as well. google is your friend.
i don't quite agree here.
/. posts; sunk to a new low ^_~), it appears that lufs lacks the caching that has been implemented in shfs, and so doesn't perform very well. a shell-only filesystem implementation would be more likely to be optimised for that purpose (after it is presumably worked on and polished), and would in theory perform better.
sure, there is another package or two that already does this, and maybe in a "better" way - better in the realm of being more general-purpose and perhaps more forward thinking. but from other posts here (yeah, i'm getting my 'facts' from
have i tried either of them? no. do i intend to? yes. just saying "it's already been done elsewhere" is a lame reason for passing up on a new technology. if we all felt that way, we'd still all be windows users.
as for being "dangerous" - give me a break. the only harm i see this causing is to a user's own files. if you want to take that risk and have root on your box, go for it. any ssh server you connect to will have file permissions in place such that you can't muck up anything important, and, if not, using something like shfs on it is the least of their problems.
and as for why this merits an article on slashdot... well, i like trek as much as the next guy, but if a lego model of the enterprise warrants an article, shfs certainly does as well.
ok, i know i'm going to get flamed for this, but here goes.
first, let's ignore the legal implications - let's say that every file traded on current p2p networks is unencumbered by copyright or any other restriction. i say this because the article in quesiton seems to be doing the same - the article is not about college kids trading illegal stuff online, but it's about a certain kind of net traffic causing bottlenecks, simply because a large amount of it exists.
next, i will exempt special-purpose networks from the point i will soon make. corporations, etc., have a right to restrict their networks to company business use only. other organisations certainly have the right to do the same. colleges are in a grey area i think. while they own their networks and should be allowed to decide what constitutes acceptable use of network resources, i believe most colleges are more apt to be less restrictive.
so finally we are left with what i guess i'll call the "residential internet." all the residential broadband customers connected by whatever backbones.
so here we have this supposed "problem" that p2p networks are using up large amounts of the available bandwidth, and ISPs are hurting trying to keep up.
so finally, my point: who cares? what is the internet for? no, i'm serious. it started as a defense department research network, but it has grown far beyond that. i would venture to say that the internet is whatever its users want it for. if its users want it for p2p file sharing, so be it. if the broadband ISPs are providing "unlimited access" (excepting bandwidth caps), then they'd damn well better provide it. if they want to tighten bandwidth caps, and offer higher-price broadband packages to people that want to shell out the cash, that's their choice, but i don't see that when i look around - most ISPs (cable, anyway) seem to have at most two options - a "business" class and a normal, lower-bandwidth option (yes, i know, DSL does have multiple rate plans from a good number of providers, but i believe cable is far more prevalent).
but stop whining. broadband customers are paying for a service. broadband ISPs have a responsibility to provide that service, for whatever the customers want to use it for. instead of doing their jobs, they're trying to slap restrictions on everything.
unfortunately, my argument is weakened by the fact that the vast majority of data swapped over p2p networks appears to be copyrighted music, movies, and software. but, when you look at the ISPs' chief complaint here - bandwidth - i have no sympathy. provide what your customers want or get out of the business.
bullshit. if the university is going to provide the internet access just for "academic use," then there damned well better be an alternative for "nonacademic use." and no, i don't consider using dialup an alternative.
i am a cornell student (a senior now). i live off campus, so i'm on cable and fortunately don't have to deal with this crap. but the people on campus have no choice. two - only two - of the dorms, out of about 25-odd buildings, are wired for cable. and by cable i mean cable TV - i don't know if you could get cable internet on those (probably), but i wouldn't put it past cornell to disallow it if anyone tried and they found out.
and don't even think of suggesting that cornell allow people to get their room wired for cable, because i know that campus life (the portion of the administration that runs cornell housing) would expressly disallow any "modifications" to the buildings.
i'm sure part of the problem is the speed at which this is occurring. when i was a freshman, network fees were around $60 per SEMESTER. the next year i believe it went up to $80, and the last year there was another price hike (i think slightly over $100). but still it was a semester price. i believe that this year they're using a per-month fee, not sure about that. figure you're in the dorm for about 8 months out of the year (i know the article says 9; that's incorrect). even at $200, that's $25 a month. damned great deal. and yes, i do agree that that's amazingly cheap for the speed - at non-peak times i'd see 800kB/s. on a rare occassion i'd see 1MB/s, 4/5 of the 10baset limit (like anyone could really get 1.25MB/s). but now we're going from $200/yr, to $240/yr. not a big jump, but i would bet that the bandwidth cap will piss off way more than the 10% of users that the article suggests.
then again, i'm sitting here paying $40/month for inferior cable - 250kB/s down on a rare good day, with an upload cap of 45kB/s. 9 months, since i have to pay while i'm not here over winter break (otherwise i get to pay extra for them to reconnect it). that's what, $360? so i probably have more of a right to complain than the dorm-dwellers ^_~.
ok, so i digressed a bit from what i was originally replying to. sue me ^_~
1) send FIN
2) wait for ACK
3) wait for FIN
4) send an ACK
if the server never receives the FIN in step 3, it assumes that the client wants to keep the connection open for some reason. this is _correct behaviour_ with regards to the TCP spec. if this article is correct, MS is merely exploiting the TCP spec to its advantage. yes, it's dirty and wastes resources, but it works.
the thing that bothers me tho, is this is what should be happening on the server end (a non-IIS server, that is):
1) send FIN
2) wait for ACK
3) ok, got ACK, now wait for FIN
4) (after timeout) hmm, no FIN, must have been lost, so we'll resend our FIN
5) client ACKs that FIN, but doesn't send its FIN
6) server thinks the response FIN is lost again, so probably resends its FIN
now the server will have a max amount of retries before it gives up and finally drops the connection (which is what it was trying for in the first place anyway). this should be a relatively low number, and the timeouts between each retransmission shouldn't be that long either. so unless IE comes back and requests another page fairly quickly, the server _should_ go ahead and drop the connection, so i fail to see how this is a problem.
the only thing i can think of is that the client keeps responding with an ACK to the server's FINs (despite not sending its own FIN), so maybe the server won't drop the connection for that reason (since the client is obviously still alive, just not responding as expected). i don't remember the TCP spec all that clearly with regards to connection teardown, so that may be where IE is able to keep the connection open.
then again, i could be totally wrong here, but i don't think so...