Right now, I'm going through a cynical phase where I feel like my university is more interested in being a factory for producing white-collar workers, than being a place of education and higher learning.
i agree wholeheartedly. i just graduated in january. sometime around sophomore/junior year i started exhibiting the cynicism you're talking about. i just didn't see the point anymore. my school felt like a job factory, and the classes weren't catching my interest. i majored in electrical and computer engineering, and, while i still conceptually love the field, i'm currently working (for the university, no less) in a software/systems design capacity. i took a bunch of CS classes for the minor, and those didn't really catch my interest either (though i will admit i enjoyed the graduate-level CS classes more than my ECE classes).
it took me an extra semester to graduate - i got depressed and stopped doing my work for a semester. so now i'm working for the university. i feel really lucky to even have a job, even if it's only a temporary position. i'm not really sure what i'm doing come july when my appointment ends. i'm not all too optimistic about finding anything anywhere else; i'm mainly holding onto the hope that i'll be offered a more long-term position here. we'll see...
Twenty seconds seems an eternity to wait. Especially for something as trivial as looking up a phone number, or someone's address.
I mean, you can do this in two to four seconds if you just had a regular address book. Every handheld designer knows this, and their power up times are minimal.
then use a tool appropriate for the job, such as a PDA. ignoring the bootup times, your desktop is ill-suited as a contacts manager/addressbook simply because it is non-portable (in the physical sense). regardless, i would doubt that all that many people carry around their laptops wherever they go.
this BIOS mis-feature is just a poor attempt to turn a desktop/laptop machine into something it's not. "general purpose" != "can do anything".
if you don't own a PDA and simply must use your desktop machine as an addressbook, just don't turn the thing off.
(personally, i use my cell phone as an address book. well, ok, it doesn't hold addresses, just phone numbers, but i find that i so rarely need actual addresses, my desktop, which is on all the time anyway, suffices.)
agreed on the "being a dick" point, but i think what he meant is that if there are improvements, say, in performance, gtk2 apps will get the bonus just by upgrading gtk. of course improvements that involve new APIs won't be usable without app updates as well.
granted, i haven't read the release announcement or any condensed changelogs, so i dunno if any such improvements have been made (to perf or otherwise).
not really. the empirical formula for a molecule doesn't necessarily reflect the bond structure imposed by the component atoms/molecules. for example, consider most hydrocarbons - CxHyOz. to use a similar example, alcohols contain a standard hydrocarbon chain bonded to an OH molecule. you'd still write that using the CxHyOz notation. (damn/. for not allowing the <sub> tag.)
hmm, that's odd that you didn't find it. perhaps it didn't come from there, though i'm relatively sure it comes from a book i read very recently, and that would be the only one i've touched in the past week or so. sorry i can't give you a specific page number.
Being a soldier means as much about loving war as being a firefighter does about loving fire.
-- Uggy, Slashdot
just an FYI: this quote actually comes from 'a painted house' by john grisham (which i just read last night). i don't know if its use in that novel was original or borrowed from somewhere else, but methinks that's a better atrribution than "uggy".
no, correlation does not imply causation, but i don't think that's the issue here.
fact: corporations with money have been able to unfairly influence the government in the past.
fact: the USPTO reexamines and rejects an insignificantly small number of patents.
fact: microsoft, in the past, has thrown its weight around when they've wanted things out of the gov't.
no, this isn't a ringing indictment, but i think it's suspicious enough to warrant some amount of skepticism that all this may not have been totally on the level.
I must disagree with your sentiments. Personally I feel the modern drive for more consistent grammar and spelling to be the result of a continuing cultural bias towards the superficial.
i'm not convinced that a drive to be consistent in _anything_ is superficial. and certainly not in a form of written communication. we have enough trouble in the form of language barriers - i see no need to add to that by perpetuating poor spelling and grammar.
Your right that proper presentation does help one in the world today. Yet I am far from convinced that the growing emphasis in society on presentation as opposed to content is a good thing.
i wouldn't call it "growing." i think it's been a relative constant - perhaps not several hundred years ago, but at least as long as items such as (consistent) dictionaries have been available to the vast majority of people.
i agree that, on the surface, content should always trump presentation. that seems to follow from oft-repeated sayings like "it's the thought that counts." but - for better or worse - we live in a world where we are not only judged by what we say and do, but how we say it and how we look while saying it. you seem to acknowledge this fact while disagreeing with it, so this obviously isn't new information for you.
I do not think that is what you are advocating but such sentiments are evident in other posts made suggesting that sub par grammar in a post is suitable grounds for out right rejection of any ideas so presented.
i think you'd be hard-pressed to find an example where poor spelling/grammar is publicly rejected in such a manner. of course, you'll see responders criticising the initial poster on his/her writing, but in general those are just people with too much time on their hands and probably some issues of their own. at the very least, a dose of perfectionism.
Perhaps I am biased due to the fact in my studies I was often required to read original historical source materials. In general the worst/. hack writes with more consistency than some of the greatest luminaries in history.
that may be the case, but i would bet that at the time, many of these people didn't have the benefit of consistent dictionaries, thesuaruses, or grammar guides. today, you can pick up a variety of any of these at a local bookstore.
I find the notion that 'correct' grammatical structure and consistent dictionary spellings have any impact on the substance of writting to be rather humorous. Presentation and content rarely have more than a passing aquaintence. While good grammar and spelling is gerenally a good marker of education it by no means conveys intelligence beyond a good memory and a penchant for mimicry.
here i have to disagree. there's the unrelated rebuttal: presentation and content are often tightly bound in an organisational sense depending on the subject matter.
i think "penchant for mimicry" is a bit strong. that's like saying "writing web pages that comply with standards is a sign that you have a penchant for mimicry." no, you don't. you simply want to make it easy for users of your product (web pages aka your writing) to be able to understand what's going on and view the information as you intend.
I could not agree more that ones prime courtesy to others in written discourse is to clearly and lucidly present your thoughts in a manner understandable to your audience. Can consistent grammar and spelling be benificial in that effort ? Most certainly. However there is a difference between grammatical and spelling inconsitencies which cause confusion and those which are undeniably superficial.
true, but this also depends upon the disposition of the reader. one of my pet peeves happens to be poor spelling and grammar. perhaps that is because, for
Had I released KillerApp under a BSD license, Bill or Darl could make a few changes to it and distribute the result under any terms they want - potntially making the new, improved KillerApp completely closed and proprietary. They might make megabucks by changing 10 lines of source code and selling the binary only. Is that fair to me, when I did most of the work originally? No - that's why I release under GPL.
that's fine - and that represents a _personal choice_ made by you. you are forcing the code to be perpetually free, but you are not giving as much freedom to others as you think. sure, they have the freedom to mess with it, modify it, re-release it, etc. but you aren't giving them the freedom to repackage it as they please if that manner is proprietary and source-code-less.
and on that, i agree. i personally don't want MS/SCO/(insert evil corporation of the week) taking my hard work, packaging it in a black box, and selling it for who knows how much. but this has nothing to do with 'freedom'. at least not freedom for _real people_.
information does not 'want' to be free. it just is. it has no feelings, no emotion, no desires. it doesn't 'care' if it's free. _people_ care, because the level of freedom determines how free they are to use/modify/redistribute the code.
The problem with the XFree86 1.1 license is that it imoses a requirement to give credit if someone chooses to distribute a binary build of their software.
how is that a problem? you're ok with requiring that your source code be distributed along with binaries of your software, but you think it's too much require that others give credit where it's due? that's reeks of hypocrisy.
If, in fact, they explicitly exclude client programs linked to the XFree86 libraries, then I don't see a problem.
they don't need to 'excplicitly exclude' the X libraries. they just need to not license them under their new license, which is what they've done. granted, i believe the exact wording was that they've "deferred" a decision to relicense the core X libs until later. regardless, i believe there's an implementation of libX11 maintained by freedesktop.org. if the xfree86 maintainers choose to relicense _their_ X11 libraries with the new license, binary distributors should be able to 'get around' the advertising clause by linking against the freedesktop xlibs, as long as they remain binary-compatible. (i presume they can do this, but IANAL.) personally, i'd like to see a shift toward something like XCB, but that's something that's a bit of a ways away i'm sure.
the fact of the matter is that the new xfree86 license is more free than most other licenses out there, including the GPL, when you consider giving _actual people_ the freedom to do what they wish with the code. personally, i'd only license my stuff under the GPL. i'm too much of a control freak to consider something like this new xfree86 license, or bsd, or ($DEITY forbid) public domain. but, unlike others, i'm not going to knock those that are more selfless than i am and will license their code under terms less restrictive than the GPL.
on a side note, given RMS's constant desire for attribution of work (example: "it's GNU/Linux, not Linux"), i'm utterly confused as to why the GPL doesn't contain something of an advertising clause.
And people expect Open Source Software to be adopted by the unwashed masses?
i don't. or rather, i'm not particularly concerned with whether it is or isn't. all of these "linux for the desktop", "linux to overthrow microsoft" people really annoy me. on the surface, i really don't care if linux kills microsoft or not. as long as it is there for me to use and hack on, i'll be happy. indirectly, i _do_ care, because the continuing existence and growth of microsoft is somewhat of a threat to linux.
What about the OSS people who, upon reading about bugs in various Microsoft programs, go "damned Microsoft, we're not supposed to beta test for them! Don't they test things before releasing them?" and then turn around and say "oh but Open Source Software means that whoever uses it is a beta tester!" It smacks of hypocracy.
it's not hypocrisy. the closed source, proprietary business model _requires_ that the company/developers thoroughly test the code themselves and/or use a controlled group of beta testers. the end users not only do not want to beta test the software (that, in many cases, they have paid money for) but they are incapable of doing the kind of testing that you can do on OSS. by contrast, OSS not only lends itself well to the user-as-tester/developer model, but frankly professes that to be How Things Are Done.
it's hypocritical to accuse someone of using the same practices in which you engage. it's not hypocritical to have a model, and accuse others of attempting to use that same model in a manner that is incompatible with the rest of their system (when it works well in your system).
The method comprises receiving an indication from a user to preview the multiple virtual desktops...
granted, i'd bet there's prior art here too, but does this line actually exclude the standard desktop pager that sits on a panel or whatnot? the fact that it's always visible means that there is no "indication from a user" that he/she wants to preview the desktops. it just does it all the time.
at the very least, i believe metacity has done this from its inception - pressing ctrl+alt+arrows displays a row of desktops in the middle of the screen while allowing you to navigate. when did the first release of metacity come out? MS applied for the patent in april of 2002, so that might be close... regardless, i would be very surprised if there wasn't an older WM that did this as well.
Re:I've got one reason to choose Linux over UNIX-S
on
SCOoby Snacks
·
· Score: 2, Informative
Now, correct me if I am wrong, but an ABI definition tells you what to put in which registers and how to make the system call.
i wouldn't call that an ABI - that's really just a calling convention.
the ABIs include such things as the integers that correspond to each syscall, the actual integers behind error codes that we think of as ENOMEM, EAGAIN, etc., the numbers used for each signal (SIGBUS, SIGTERM, etc.). also you'd include in the ABI e.g. the format of the ioctl() call - what arguments it takes, in what order, and what each argument means and does.
ever seen a run-time linker error when you try to run something compiled against an incompatible glibc? that's a user-space ABI incompatibility at work.
think about the words that make up the acronym 'ABI'. it's an 'interface' (function calls, defined constants) that 'applications' use. the 'binary' portion just means that differing ABIs don't impact how the code is written (i.e., you can have correctly-compiling code that is ABI-incompatible with a particular library or whatever). assuming that all else was the same (binary format, linker format, calling conventions, etc.), you could theoretically compile a binary on BSD or SCO UNIX that would run on linux. unfortunately, it wouldn't actually do anything - it would probably crash almost immediately - because, for example, when linux sends a signal to the app with number 14, linux is sending SIGALRM. but the app, compiled against a different ABI, may believe that 14 means something else, say, SIGIO (contrived example).
when we're talking about user-space, ABI incompatibilities usually manifest themselves as link time errors (compile-time or run-time). when we're talking about the user/kernel interface, ABI incompatibilities manifest themselves as crashes, because the application erroneously believes it is doing one thing, but the kernel believes it's trying to do something else (possibly in violation of what it's allowed to do).
but the point that you made is still correct - much of linux's ABI does differ from SCO UNIX (and BSD), so i don't see where SCO is going with this...
i would agree with this, except for one caveat. (this may not be an issue, as i haven't RTFAed, but...)
kerberos is designed with the concept of a single authoritative authentication directory in mind. that seems to be pretty much at odds with jabber's goals here.
now, it _is_ possible to form "trust relationships" between disparate kerberos realms, but that isn't really an oft-used feature, and it seems to me to be something that was almost tacked on last minute without being truly designed into the system.
now, what i'd really like to see is a fundamental redesign of kerberos (version 6, anyone?) that takes into account some of the open, decentralised concepts that jabber seems to be pushing here.
of course, the final issue is application support. despite being around forever, kerberos still has little to no support in web browsers. certainly none of the major browsers (moz, IE, safari) support kerberos auth. other kerberised apps are hard to come by - sure, the krb5 distribution comes with kerberised telnet, ftp, rsh, etc., but GUI clients are hard to come by. and they want to add _yet another_ authentication protocol? who's going to support it?
i think something like this is a great idea, but unfortunately i believe that you'd need some major corporate backing to get this into current applications.
Do you really think that more than 1% of the traffic on Kazaa comes from legitimate sharing?
i didn't think so myself, but i recall reading something here on/. the other day that linked to a report (sorry, don't have it handy) stating that 10% of p2p traffic (which granted could be different from kazaa traffic) was legitimate.
The network operators could easily log the names of files that are being downloaded. They don't, because that information could be subpeonaed...
uh-uh. you don't appear to quite understand how this works. there is no big super-server that keeps track of all the files on the network. there are instead a series of many many not-quite-as-super servers that run all over the world. they're called 'supernodes', and they're users running the kazaa client just like any other user. the supernodes act as a sort of semi-central hubs that help with the routing of searches (among other things). in fact, if you were to connect to the network at two (geographically) different locations and simultaneously initiate the same search, the results you get will almost certainly be different, because connection #1's nearest supernode may or may not have access to the same resources that connection #2 has.
so they _could_ try to log what goes on on the network, but their picture of it would be far from the whole picture. and even then, with the current infrastructure they could only find out what's on the network, not what people are actually downloading. they could get a sample of what people are downloading by hosting a multitude of content, and then collecting statistics on what people get from them, but otherwise that's about it. at any rate, any such stats-gathering mission is severely limited based on the scope of the material they offer.
now, they can figure out what's being shared simply by doing a bunch of searches. but, as i said, the lists they get back will be far from complete, and will change very rapidly.
that's correct, but i think you're missing the point. this article is about uncapping your modem so it is capable of transferring data at speeds _above_ what the cable company advertises. if you're not even getting what you're _supposed_ to get, uncapping the modem isn't going to change anything.
certainly, if you're promised 2/256, that's what you should get. but that's a separate issue.
sure, you can do that, and they have every right to refuse you service when you do, since most of these blanket contracts usually have a clause in them that basically say they can terminate your service whenever they damn well please.
no, online games really do have low bandwidth requirements. the sticking point is the latency of the connection. if all the building is doing is playing halo with the outside world all day, a 2-3mbit cable line should be sufficient. if people are saturating the connection with downloads and uploads, however, you'll run into problems. wireless is also a bit more latent than wired, so on a loaded network there might be some issues as well.
contrary to somewhat-popular belief, a house that you live in, especially when it is mortgaged, is a liability, not an asset.
furthermore, if he's going to be putting down something in the realm of $10k on a house, he's going to have a huge mortgage. the monthly payments (depending upon where he lives) will likely exceed his current monthly rent, and a large portion of those mortgage payments are going to be on interest (more money wasted).
if he's into real estate, i'd suggest putting that money into real estate that's rented out itself. nice bit of supplemental income if you do it right and are willing to accept the risk and take the time and research to do it properly.
however, this guy didn't ask _what_ to do with his money. he already has a plan. everything need not be about saving and planning. sometimes people just want to buy stuff, and there's nothing wrong with that. besides, i'd bet he can get them a respectable setup for under a grand, and that's $6k he can invest or put away or whatever.
as for the actual question of network setup, that unfortunately depends quite a bit on the house itself. if the walls/ceilings/floors aren't too thick, you might get away with two access points and a little bit of wiring to connect the APs to the wired network. worst case, you put two APs on each floor, one on either side. depending on what standard you go with (802.11{a,b,g}), that will run you anywhere between $1000 and $1500. if you luck out and the two-AP approach works (or even a one-per-floor approach), you'll finish with a lot of money to spare.
i can't really recommend what brands are the best, because my experience is limited. but i'd suggest that after you pick an AP model, get two of them, and see what kind of coverage you can get. every building is different, and experimentation is the only way you'll learn what your particular building needs. if you need more APs, go out and buy them, but after experimenting you should have a pretty fair idea of how many you'll need in total. for a 4-story house i'd say two is the bare minimum if you're lucky, but you may want more because the quality of the connection at the fringes of range may not be acceptable to you.
the point has nothing to do with the government interfering with business. the point is that when the government contracts a corporation to do work for them, if that corporation is offshoring the work, then the government is essentially paying the wages of non-US citizens. it's like importing IT from overseas. i'm not saying there's anything wrong with imports, but when you have a perfectly willing workforce available domestically, and you're a government that pledges to act in the best interests of your constitutents, you choose the domestic labor, not the foreign.
just my opinion, of course, tho it also happens to be the right one ^_~.
and the amusing fact is, although there are some GPL zealots on the lkml, linus torvalds himself is not one of them. i've seem him quoted to the effect that the GPL worked for him at the time to further his goals as an OS developer, and that's why he used it - not because he felt an overwhelming desire to preach about how software should be free. hell, he even uses bitkeeper, a proprietary SCM, to manage the kernel. why? because it does what he needs it to, simple as that.
Slashdot Mirror
it took me an extra semester to graduate - i got depressed and stopped doing my work for a semester. so now i'm working for the university. i feel really lucky to even have a job, even if it's only a temporary position. i'm not really sure what i'm doing come july when my appointment ends. i'm not all too optimistic about finding anything anywhere else; i'm mainly holding onto the hope that i'll be offered a more long-term position here. we'll see...
this BIOS mis-feature is just a poor attempt to turn a desktop/laptop machine into something it's not. "general purpose" != "can do anything".
if you don't own a PDA and simply must use your desktop machine as an addressbook, just don't turn the thing off.
(personally, i use my cell phone as an address book. well, ok, it doesn't hold addresses, just phone numbers, but i find that i so rarely need actual addresses, my desktop, which is on all the time anyway, suffices.)
agreed on the "being a dick" point, but i think what he meant is that if there are improvements, say, in performance, gtk2 apps will get the bonus just by upgrading gtk. of course improvements that involve new APIs won't be usable without app updates as well.
granted, i haven't read the release announcement or any condensed changelogs, so i dunno if any such improvements have been made (to perf or otherwise).
not really. the empirical formula for a molecule doesn't necessarily reflect the bond structure imposed by the component atoms/molecules. for example, consider most hydrocarbons - CxHyOz. to use a similar example, alcohols contain a standard hydrocarbon chain bonded to an OH molecule. you'd still write that using the CxHyOz notation. (damn /. for not allowing the <sub> tag.)
hmm, that's odd that you didn't find it. perhaps it didn't come from there, though i'm relatively sure it comes from a book i read very recently, and that would be the only one i've touched in the past week or so. sorry i can't give you a specific page number.
no, correlation does not imply causation, but i don't think that's the issue here.
fact: corporations with money have been able to unfairly influence the government in the past.
fact: the USPTO reexamines and rejects an insignificantly small number of patents.
fact: microsoft, in the past, has thrown its weight around when they've wanted things out of the gov't.
no, this isn't a ringing indictment, but i think it's suspicious enough to warrant some amount of skepticism that all this may not have been totally on the level.
i'm not convinced that a drive to be consistent in _anything_ is superficial. and certainly not in a form of written communication. we have enough trouble in the form of language barriers - i see no need to add to that by perpetuating poor spelling and grammar.
i wouldn't call it "growing." i think it's been a relative constant - perhaps not several hundred years ago, but at least as long as items such as (consistent) dictionaries have been available to the vast majority of people.
i agree that, on the surface, content should always trump presentation. that seems to follow from oft-repeated sayings like "it's the thought that counts." but - for better or worse - we live in a world where we are not only judged by what we say and do, but how we say it and how we look while saying it. you seem to acknowledge this fact while disagreeing with it, so this obviously isn't new information for you.
i think you'd be hard-pressed to find an example where poor spelling/grammar is publicly rejected in such a manner. of course, you'll see responders criticising the initial poster on his/her writing, but in general those are just people with too much time on their hands and probably some issues of their own. at the very least, a dose of perfectionism.
that may be the case, but i would bet that at the time, many of these people didn't have the benefit of consistent dictionaries, thesuaruses, or grammar guides. today, you can pick up a variety of any of these at a local bookstore.
here i have to disagree. there's the unrelated rebuttal: presentation and content are often tightly bound in an organisational sense depending on the subject matter.
i think "penchant for mimicry" is a bit strong. that's like saying "writing web pages that comply with standards is a sign that you have a penchant for mimicry." no, you don't. you simply want to make it easy for users of your product (web pages aka your writing) to be able to understand what's going on and view the information as you intend.
true, but this also depends upon the disposition of the reader. one of my pet peeves happens to be poor spelling and grammar. perhaps that is because, for
and on that, i agree. i personally don't want MS/SCO/(insert evil corporation of the week) taking my hard work, packaging it in a black box, and selling it for who knows how much. but this has nothing to do with 'freedom'. at least not freedom for _real people_.
information does not 'want' to be free. it just is. it has no feelings, no emotion, no desires. it doesn't 'care' if it's free. _people_ care, because the level of freedom determines how free they are to use/modify/redistribute the code. how is that a problem? you're ok with requiring that your source code be distributed along with binaries of your software, but you think it's too much require that others give credit where it's due? that's reeks of hypocrisy. they don't need to 'excplicitly exclude' the X libraries. they just need to not license them under their new license, which is what they've done. granted, i believe the exact wording was that they've "deferred" a decision to relicense the core X libs until later. regardless, i believe there's an implementation of libX11 maintained by freedesktop.org. if the xfree86 maintainers choose to relicense _their_ X11 libraries with the new license, binary distributors should be able to 'get around' the advertising clause by linking against the freedesktop xlibs, as long as they remain binary-compatible. (i presume they can do this, but IANAL.) personally, i'd like to see a shift toward something like XCB, but that's something that's a bit of a ways away i'm sure.
the fact of the matter is that the new xfree86 license is more free than most other licenses out there, including the GPL, when you consider giving _actual people_ the freedom to do what they wish with the code. personally, i'd only license my stuff under the GPL. i'm too much of a control freak to consider something like this new xfree86 license, or bsd, or ($DEITY forbid) public domain. but, unlike others, i'm not going to knock those that are more selfless than i am and will license their code under terms less restrictive than the GPL.
on a side note, given RMS's constant desire for attribution of work (example: "it's GNU/Linux, not Linux"), i'm utterly confused as to why the GPL doesn't contain something of an advertising clause.
it's hypocritical to accuse someone of using the same practices in which you engage. it's not hypocritical to have a model, and accuse others of attempting to use that same model in a manner that is incompatible with the rest of their system (when it works well in your system).
at the very least, i believe metacity has done this from its inception - pressing ctrl+alt+arrows displays a row of desktops in the middle of the screen while allowing you to navigate. when did the first release of metacity come out? MS applied for the patent in april of 2002, so that might be close... regardless, i would be very surprised if there wasn't an older WM that did this as well.
the ABIs include such things as the integers that correspond to each syscall, the actual integers behind error codes that we think of as ENOMEM, EAGAIN, etc., the numbers used for each signal (SIGBUS, SIGTERM, etc.). also you'd include in the ABI e.g. the format of the ioctl() call - what arguments it takes, in what order, and what each argument means and does.
ever seen a run-time linker error when you try to run something compiled against an incompatible glibc? that's a user-space ABI incompatibility at work.
think about the words that make up the acronym 'ABI'. it's an 'interface' (function calls, defined constants) that 'applications' use. the 'binary' portion just means that differing ABIs don't impact how the code is written (i.e., you can have correctly-compiling code that is ABI-incompatible with a particular library or whatever). assuming that all else was the same (binary format, linker format, calling conventions, etc.), you could theoretically compile a binary on BSD or SCO UNIX that would run on linux. unfortunately, it wouldn't actually do anything - it would probably crash almost immediately - because, for example, when linux sends a signal to the app with number 14, linux is sending SIGALRM. but the app, compiled against a different ABI, may believe that 14 means something else, say, SIGIO (contrived example).
when we're talking about user-space, ABI incompatibilities usually manifest themselves as link time errors (compile-time or run-time). when we're talking about the user/kernel interface, ABI incompatibilities manifest themselves as crashes, because the application erroneously believes it is doing one thing, but the kernel believes it's trying to do something else (possibly in violation of what it's allowed to do).
but the point that you made is still correct - much of linux's ABI does differ from SCO UNIX (and BSD), so i don't see where SCO is going with this...
i would agree with this, except for one caveat. (this may not be an issue, as i haven't RTFAed, but...)
kerberos is designed with the concept of a single authoritative authentication directory in mind. that seems to be pretty much at odds with jabber's goals here.
now, it _is_ possible to form "trust relationships" between disparate kerberos realms, but that isn't really an oft-used feature, and it seems to me to be something that was almost tacked on last minute without being truly designed into the system.
now, what i'd really like to see is a fundamental redesign of kerberos (version 6, anyone?) that takes into account some of the open, decentralised concepts that jabber seems to be pushing here.
of course, the final issue is application support. despite being around forever, kerberos still has little to no support in web browsers. certainly none of the major browsers (moz, IE, safari) support kerberos auth. other kerberised apps are hard to come by - sure, the krb5 distribution comes with kerberised telnet, ftp, rsh, etc., but GUI clients are hard to come by. and they want to add _yet another_ authentication protocol? who's going to support it?
i think something like this is a great idea, but unfortunately i believe that you'd need some major corporate backing to get this into current applications.
so they _could_ try to log what goes on on the network, but their picture of it would be far from the whole picture. and even then, with the current infrastructure they could only find out what's on the network, not what people are actually downloading. they could get a sample of what people are downloading by hosting a multitude of content, and then collecting statistics on what people get from them, but otherwise that's about it. at any rate, any such stats-gathering mission is severely limited based on the scope of the material they offer.
now, they can figure out what's being shared simply by doing a bunch of searches. but, as i said, the lists they get back will be far from complete, and will change very rapidly.
that's correct, but i think you're missing the point. this article is about uncapping your modem so it is capable of transferring data at speeds _above_ what the cable company advertises. if you're not even getting what you're _supposed_ to get, uncapping the modem isn't going to change anything.
certainly, if you're promised 2/256, that's what you should get. but that's a separate issue.
sure, you can do that, and they have every right to refuse you service when you do, since most of these blanket contracts usually have a clause in them that basically say they can terminate your service whenever they damn well please.
no, online games really do have low bandwidth requirements. the sticking point is the latency of the connection. if all the building is doing is playing halo with the outside world all day, a 2-3mbit cable line should be sufficient. if people are saturating the connection with downloads and uploads, however, you'll run into problems. wireless is also a bit more latent than wired, so on a loaded network there might be some issues as well.
contrary to somewhat-popular belief, a house that you live in, especially when it is mortgaged, is a liability, not an asset.
furthermore, if he's going to be putting down something in the realm of $10k on a house, he's going to have a huge mortgage. the monthly payments (depending upon where he lives) will likely exceed his current monthly rent, and a large portion of those mortgage payments are going to be on interest (more money wasted).
if he's into real estate, i'd suggest putting that money into real estate that's rented out itself. nice bit of supplemental income if you do it right and are willing to accept the risk and take the time and research to do it properly.
however, this guy didn't ask _what_ to do with his money. he already has a plan. everything need not be about saving and planning. sometimes people just want to buy stuff, and there's nothing wrong with that. besides, i'd bet he can get them a respectable setup for under a grand, and that's $6k he can invest or put away or whatever.
as for the actual question of network setup, that unfortunately depends quite a bit on the house itself. if the walls/ceilings/floors aren't too thick, you might get away with two access points and a little bit of wiring to connect the APs to the wired network. worst case, you put two APs on each floor, one on either side. depending on what standard you go with (802.11{a,b,g}), that will run you anywhere between $1000 and $1500. if you luck out and the two-AP approach works (or even a one-per-floor approach), you'll finish with a lot of money to spare.
i can't really recommend what brands are the best, because my experience is limited. but i'd suggest that after you pick an AP model, get two of them, and see what kind of coverage you can get. every building is different, and experimentation is the only way you'll learn what your particular building needs. if you need more APs, go out and buy them, but after experimenting you should have a pretty fair idea of how many you'll need in total. for a 4-story house i'd say two is the bare minimum if you're lucky, but you may want more because the quality of the connection at the fringes of range may not be acceptable to you.
-1, overly nitpicky, pedantic, and annoying
the point has nothing to do with the government interfering with business. the point is that when the government contracts a corporation to do work for them, if that corporation is offshoring the work, then the government is essentially paying the wages of non-US citizens. it's like importing IT from overseas. i'm not saying there's anything wrong with imports, but when you have a perfectly willing workforce available domestically, and you're a government that pledges to act in the best interests of your constitutents, you choose the domestic labor, not the foreign.
just my opinion, of course, tho it also happens to be the right one ^_~.
are you trying to use the rivafb framebuffer driver? if so, don't. it doesn't play nice with nvidia's X drivers. use vesafb instead.
oh please. if you think _any_ company supports linux for reasons other than that they think it will make them money, you are sadly deluded.
and the amusing fact is, although there are some GPL zealots on the lkml, linus torvalds himself is not one of them. i've seem him quoted to the effect that the GPL worked for him at the time to further his goals as an OS developer, and that's why he used it - not because he felt an overwhelming desire to preach about how software should be free. hell, he even uses bitkeeper, a proprietary SCM, to manage the kernel. why? because it does what he needs it to, simple as that.