As far as I can tell from the article, this only affects business students in the school. Judging from that, I'm guessing someone in the department was keeping a few spreadsheets or something of that nature around on a public windows share. This strikes me as far more of a careless employee problem than a truly insecure infrastructure problem. Thus, comments about CERT may be a bit premature.
A piece of suggested paper reading would be The Diamond Age by Neal Stephenson... Many of you have prolly read it already... The ideas presented there (which certainly aren't new ideas, but he puts them down nicely and wraps them up in a nifty plot while he's at it) give an interesting view of digital paper, etc. Although even within the bounds of the world created inside this work there are those who still choose to read things on real paper. It's a privledge for the rich, however.
QNX is actually a very nice operating system for any PC. The only real flaw it has is a lack of applications (and possibly a poorly designed crypt() function, however I'm not entirely sure if the crypt() supplied with the I-Opener comes in the actual QNX distribution). The nice thing about the system is it's Microkernel design. Similar to the concept behing Mach and the HURD, QNX has a very small kernel (although it seems to be a good deal larger than I'd expect these days) that handles IPC while smaller applications do the actual work of drivers. You can hot swap almost any driver in the system, with the possible exception (I've never been crazy enough to try) of the process manager (Proc32) and the general filesystem driver (Fsys). I'm amazed that they're giving this stuff away for free, I'd like to see what the free software (free as in beer) community can do with it once it's out. I do hope however, that they don't populate it with a bunch of bloated do-all kill-all applications and keep to the QNX (and actually, UNIX) mentality of do one thing, do it well.
Also, before you guys get to work programming, I ask that you please read the on-line manuals on the functions Send() Receive() and Reply() and that you actually use those functions (along with qnx_register_name() and qnx_name_locate()) together those functions form the basis of the best inter-process communication i've seen to date.
In your own comment, you point out that someone can still modify the server. This is really impossible to prevent even using JC's method, because one would still have to trust the server operator. Besides, the very point of opening up the Quake source is so that people can create useful modifications to it and release them to the public. JC's solution requires using a digest algorithm on the client itself. Obviously this is an entirely invalid strategy, because it prevents people from using their own slightly modified clients. The solution is (and always has been) to assume the server is trusted and the client is not. The majority of server's out there that anyone is willing to play on will be trustworthy anyway.
It's too bad this kind of thing would happen inside Corel. They were one of the biggest people backing Linux. If this shakes Corel up at all, other companies really aren't going to want to follow suit, which might inhibit increased Linux support. I was really pulling for Corel too, they've done some nice stuff in the past, and I thought they could produce some of the really clean and functional apps that IT managers need to see.
I sit here using my Linux box, and all seems uniform because all my window borders look the same. Then I look a little deeper. On my screen right now i've got Motif, Arena, GTK+, Qt, and whatever widget set Eterm uses. Sure they're ALMOST all alike, but they all have their unique traits. For example, if I click a menu in Motif, Qt, or GTK+, it stays down, but if I simply click it in the other two, it pops out for a second and disappears (similar to a Macintosh). Also, the menus look different, the whole app just feels different. This could very easily disorient users, and something really ought to be done about it. Obviously we can't ask people to rewrite all their applications. First of all, the programmers would never go for it, and second it would be too much work. My proposal is that we create a set of wrapper libraries. The ideal implementation would be to allow the user to choose the widget set they want, and suddenly all applications use it. Of course this may seem incredibly difficult at first, because it means creating seperate wrapper libraries for each widget set...or does it? I propose we create a very basic widget set, one that enables you to DO everything existing ones do, but can also be easily implemented on top of any other set. That way we simply wrap this around each set, and wrap all of the others around this, and we're done. I do of course see the two obvious problems with this: it's slow, and it's bloated. My idea was that it would be an interim(sic) solution, eventually we would need to wrap each widget set around each other one, but in the interest of fast development I would like to see my other idear implemented as well. I unfortunately don't have the time to do this, and because I'm posting so late I doubt anyone will see this, but I think this would go a lot way in bringing linux to the masses.
I'm from a small Catholic school (175 students) in Pittsburgh, PA. All of the stories in this article seem so distant to me, because in my school such discrimination would quickly be noticed, and would spread like wildfire. Although, as far as some students (and possibly faculty) are concerned, I could be R. Harris. I'm a `computer geek' who plays a lot of DooM and Quake. I wear a black trenchcoat. I know how to make explosives (oooh...pipe bomb, there's a tough one...take pipe, cap both ends, drill hole in one, full with explosive, insert detonator), but it's not a skill a learned on-line. It's something i figured out from Chemistry class, and because I have a mind larger than a small peanut. I could probably be put into the catagory of anti-social if you looked closely enough at me. Quite frankly however, I'm not a killer. The people who really know me can see that is obvious. But although I may not receive suspensions, or letters form the school Administration (yet). I still feel the stare of the principal on me every second I'm within her view. In a school as small as mine, there aren't many geeks to watch. In fact, I'll guarantee I'm the most computer literate person in the school (this is including teaching staff), simply because I've seen everyone else trying to use a computer. In all honesty, I'm more worried about the stupid people, or the jocks, or whatever group you'd like to choose for the blockhead masses. I actually heard once last year that someone said "I really wish I could kill Jon Olson. That kid just pisses me off with his `smarter than thou' attitude." Possibly the reason you think I have that attitude it because I am `smarter than thou.' Sure, I'll glady answer any question a teacher asks me. However, I can't shoot a basketball to save my life. Hell, I just got my learner's permit, and even the simple coordination required for driving is a challenge for me. It's just that I acquired different skills in my childhood. And the real kicker is, if you ever do something wrong academically, it's assumed that you're not really smart, and you're just faking it most of the time. Anyway, I'm rambling, and by the time this comment gets posted no one will be reading that far anyway.
Slashdot Mirror
Sure, we've got a structure that we lovingly refer to as the shaft, but at least we use protection.
As far as I can tell from the article, this only affects business students in the school. Judging from that, I'm guessing someone in the department was keeping a few spreadsheets or something of that nature around on a public windows share. This strikes me as far more of a careless employee problem than a truly insecure infrastructure problem. Thus, comments about CERT may be a bit premature.
A piece of suggested paper reading would be The Diamond Age by Neal Stephenson... Many of you have prolly read it already... The ideas presented there (which certainly aren't new ideas, but he puts them down nicely and wraps them up in a nifty plot while he's at it) give an interesting view of digital paper, etc. Although even within the bounds of the world created inside this work there are those who still choose to read things on real paper. It's a privledge for the rich, however.
Also, before you guys get to work programming, I ask that you please read the on-line manuals on the functions Send() Receive() and Reply() and that you actually use those functions (along with qnx_register_name() and qnx_name_locate()) together those functions form the basis of the best inter-process communication i've seen to date.
Anyway, I'm rambling now
-- Jon Olson
In your own comment, you point out that someone can still modify the server. This is really impossible to prevent even using JC's method, because one would still have to trust the server operator. Besides, the very point of opening up the Quake source is so that people can create useful modifications to it and release them to the public. JC's solution requires using a digest algorithm on the client itself. Obviously this is an entirely invalid strategy, because it prevents people from using their own slightly modified clients. The solution is (and always has been) to assume the server is trusted and the client is not. The majority of server's out there that anyone is willing to play on will be trustworthy anyway.
It's too bad this kind of thing would happen inside Corel. They were one of the biggest people backing Linux. If this shakes Corel up at all, other companies really aren't going to want to follow suit, which might inhibit increased Linux support. I was really pulling for Corel too, they've done some nice stuff in the past, and I thought they could produce some of the really clean and functional apps that IT managers need to see.
I sit here using my Linux box, and all seems uniform because all my window borders look the same. Then I look a little deeper. On my screen right now i've got Motif, Arena, GTK+, Qt, and whatever widget set Eterm uses. Sure they're ALMOST all alike, but they all have their unique traits. For example, if I click a menu in Motif, Qt, or GTK+, it stays down, but if I simply click it in the other two, it pops out for a second and disappears (similar to a Macintosh). Also, the menus look different, the whole app just feels different. This could very easily disorient users, and something really ought to be done about it. Obviously we can't ask people to rewrite all their applications. First of all, the programmers would never go for it, and second it would be too much work. My proposal is that we create a set of wrapper libraries. The ideal implementation would be to allow the user to choose the widget set they want, and suddenly all applications use it. Of course this may seem incredibly difficult at first, because it means creating seperate wrapper libraries for each widget set...or does it? I propose we create a very basic widget set, one that enables you to DO everything existing ones do, but can also be easily implemented on top of any other set. That way we simply wrap this around each set, and wrap all of the others around this, and we're done. I do of course see the two obvious problems with this: it's slow, and it's bloated. My idea was that it would be an interim(sic) solution, eventually we would need to wrap each widget set around each other one, but in the interest of fast development I would like to see my other idear implemented as well. I unfortunately don't have the time to do this, and because I'm posting so late I doubt anyone will see this, but I think this would go a lot way in bringing linux to the masses.
-- Jon Olson
morph@NOSPAM.jmss.com
I believe that those would fall into the catagory of GoodQuestions(TM) to ask at the forum. -- morph-
Geek power and all that jazz.