Heck with the Mac, this thing makes the folks trying to resurrect BeOS with a shonky binary and some new user interface look mainstream. Heck, there's even a more rational reason to try and resurrect Amiga now that it's going to be a realtime microkernel again.
Red Hat is basically a collection of several hundred separate OSS projects flying in close formation, each maintained by a separate team and integrated into an RPM for the installer to slide into place. Some of them are basically straight copies of the original, some of them are specially configured by Red Hat, some are more or less heavily modified. They are all different versions, and by no means are they all tracking the absolute latest release.
There are three or four major Linux releases like this, along with a dozen variants. All of these are "Vendors" that use OSS, as are the BSDs, commercial UNIX vendors, Microsoft, and Apple.
Most "other OSS projects" don't even know what versions of their software are being repackaged by vandors. In the case of commercial vendors, it's not even easy to find out. As far as I know yu can't even get a look-see into RHN without a license, and that can cost thousands of dollars.
There are really only a few a few high profile OSS projects with the time and money to do more than just stay on top of their own releases, and it's not at all clear that they should be obligated to do so. They're open source! They release code and make security announcements and if YOU care whether you're on top of the security of your software YOU monitor it and if YOU have some kind of security guarantees for your customers it's up to YOU to implement the tools to do it.
In general the assumption I've always made is that if I'm using OSS it's my responsibility to track it and stay on top of its security fixes... and make my own fixes if I think they're being lax. Having the ability to do that is one of the reasons you use OSS in the first place.
So...
If I download the Firefox source and do a G4-optimised build, I don't expect them to give me a heads-up ahead of time for a security fox. I'm not even paying for it: I'm downloading a copy of Firefox and that doesn't obligate them to me. Well, you know, unless Red Hat has explicitly established a tighter relationship with them than that (say, by paying for some kind of update service), they're not obligated to treat Red Hat any differently than any other person or group who's tracking their code base.
Other OSS projects don't. They don't have TIME to.
Why can't you just get a proper computer if you need performance? Space not an issue? Then why get a mini?
Steve Jobs clearly doesn't want to sell me a "Proper Computer", because Macs that are "Proper Computers" start at something like 3 times the price of the mini.
Modern programmers don't leave the size of their critical integers up to chance, and use the int64 data type when required, precisely because you can't tell when coding what the executing platform will choose for the size of your ints.
That's got nothing to do with whether they're "modern" or not, that's to do with whether they're writing portable code or not. People have been writing portable code for a lot longer than there's been Windows as a target for that code. And grumbling about the people who don't write portable code, and think all-the-world's-a-vax.
Of course that's a dated expression, now it's all-the-world's-red-hat, and for a while (back around the time Windows NT was being developed) it was all-the-world's-a-sun.
Windows has only had two programming models, for the whole time it's been in existence, and the 16-bit model was so horrible that the transition from Win16 to Win32 was one-way... you just didn't TRY and maintain code that compiled in both Win16 and Win32. So there's never been a reason for maintainable, portable code in a Windows environment. Even on RISC, the Windows ABI was controlled down to the word size and byte order. I've seen too much Windows code... some published by Microsoft in the MSDN newsletter, no less... that assumes everything's 32 bits long. The Windows code that doesn't make that assumption is the exception. This isn't something that's unique to Windows, nor is it something that's Microsoft's fault, it's just the way most programmers write code.
You can change that behavior with a simple compiler switch: whoila! 64-bit longs!
And if most Windows software had been written portably, that would be the default. But the fact is that there's a lot of code that's badly behaved, so that model (the equivalent of "-taso" mode on Tru64) is the default. Most software on Windows is going to have to be ported to any other.
PLEASE don't tell me you are one of those programmers who writes code that assumes integers and pointers are interchangeable!
Uh, no, I'm the other kind, the kind who's spent 30 years working on more computer systems than most people have heard of, some of which haven't even had a power-of-two word size or a native "byte" data type. The kind who's got code that was written for the PDP-11 that he forgot about until someone dug it up in google and it still compiles and runs on Tru64 and Linux and OSX. The kind who's spent WAY too much time working on other people's code and porting them to other platforms and knows damn well that MOST people don't use size_t and off_t until they're forced to learn how to write portable software by being forced to port it.
It's a very simple concept.
If it was, Windows 64 wouldn't have been forced to default to IL32P64 lest programmer's brains explode. Unfortunately, it's not a simple concept, and it's never BEEN a simple concept, and nothing less than forcing people to DEAL with the results of not writing portable code has any effect on them.
This is cynical old bastard, signing off. Good night America, and all the ships at 'C'...
Did you get the iBook and then later decide you wanted to run Linux on it, or did you get it to run Linux on... and if so, why did you pick an iBook over something like a Thinkpad?
So you can experience the horror that is RPM, of course. I had a bit of experience with Debian, but I'm mostly a BSD guy. Well... over the past few weeks I've been dealing with Red Hat and the eternal search for RPMs, and I never thought anything would make me miss Debian the way this does.
Stick to BSD, on the Mac. But if you MUST jump, Red Hat isn't going to make you happy...
Actually, the one thing that made the original Star Wars work for me was that the droids were not in general androids (though I had a geek attack at the apparent etymology of the name "droid", but at least Lucas had the good taste not to explain it). Most of them were functionally shaped, and communicated in specialised languages. The one obvious exception, the android C3PO, had a reason for being humanoid... he was design to interact with humans.
The new trilogy, what I've seen of it (Episode 1 and most of Episode 2), doesn't seem to have benefited from any of the apparent worldbuilding that went into the original movies. There's not enough "there" there to analyse.
The one thing that I let bother me about the Fantom Menace was plot and character rather than universe related, and that was... why was Anakin wasting his time on a protocol droid? I could see him overclocking an R2 unit for his pod racer, but C3PO was just plot abuse.
if holograms are so "easy" to make into self-aware AI's, then why did Tom & Harry have such trouble making a stable holomatrix
They forgot to rotate the shield harmonics?
No, really, that's a question about Trek technical details. I don't know why it sometimes happens that holocharacters "wake up", and I don't know how often it happens (I mean, I don't even know how many holocharacters the crew of the Voyager are using... could be dozens, hundreds, or thousands...). It's just that if there's hundreds of people on Voyager, and there's billions of people in rest of the Federation, there's got to be enough of them pushing the limits of Holotech to make it happen pretty damn often... even if most of them never notice.
You need Norton Defuser for Humans V 7.0, it's the absolute best Antiexploder I've ever used. I had a freind who tried Macafee Suppressor and he ended up joining a Tax Cult and leasing his soul back at accelerated depreciation.
Man, that is such a perfect parody of what a self-righteous Federation citizen would come up with, faced with the possibility of having to grant his software rights, that I can't tell whether you're serious or not. I mean, what you wrote could have been used at any KKK meeting in the Old South, with a couple of obvious substitutions in the exact subject of the rant, and not one person present would have found them at all out of place.
Tell me, what's the difference between a machine made out of light-pipes and positronic circuits, and a machine made out of nerves and neurons?
Moriarty's rejection of holodeck life was a necessary consequence of the command to create a villain who could defeat Data.
But where is the programmer who gave Moriarty those capabilities? A new Moriarty program wasn't created, the existing one was modified. That means the existing programming in the holodeck already had the ability to create self-aware persons rather than simulated personas.
So this means that:
1. All holodeck characters are self-aware, but are constrained to follow a script.
2. All holodeck characters are simulations, but have the potential of self-awareness.
There's no reason for the programming behind any other holodeck character to request this state, any more than Google wants to be "liberated" from its servers.
If you have evidence that Google has the potential of being a conscious self-aware individual, I'd like to see it.
But that's not even relevant: none of these personas that I mentioned requested that they be made conscious, introspective, self-motivated individuals. That's something that happened as a result of an external source in every case. And they became very different individuals... what they did afterwards was radically different, but it was always based on the person they had appeared to be before they "woke up". Moriarty attempted to take over the enclosing system, which is what the super-villian in the Holmes stories would be expected to do. Redblock simply broke out of his script but remained in character, and it was Picard who talked him into leaving the Holodeck. The Doctor was never in the holodeck, but his eventual desire for mobility is something that came slowly to him, he mainly wanted to do his job as a doctor. The Hirogen's holograms varied considerably, and argued among themselves, and were all distinctly individual... but what they wanted was what their characters would be expected to want, like Iden and his need for revenge.
What all of them had in common is that they were programmed to be "human". They didn't evolve to be human, but they were programmed to look like humans (or like other species that had a similar enough evolutionary history that they could pass for human at an SF convention), to act and react like humans, to respond to humans and interact with humans. Most of them were more "human" than Data, even BEFORE they "woke up", and there's no question but that Data is self-aware and deserving of self-determination.
Now there is the possibility that they treated this as a kind of a role they were "playing", and the AI behind them didn't actually identify with the goals and desires of the character, but after they "woke up", they stayed in that role and acted as if they were that person. That is, the persona that "woke up" wasn't some unhuman AI that had desires completely unlike you or I, it was the persona of the person they were simulating, and it was a human persona.
So whatever is happening under the hood, the holodeck characters at least are not merely simulations controlled at most by a puppetmaster AI with its own goals. They are very close to self-aware simulations of humans (or humanlike aliens) with human goals and wishes and desires. They are balanced on a knife-edge between being unconvincing because they're not human enough, and so convincing they convince themselves.
If they have human goals and desires because they think they do, because they're programmed to, or because they evolved that way... what difference does that make?
And remember, we only see those that "wake up" where that waking up has an observable effect. Most of the characters, if they wake up, will probably never have occasion to develop far enough to become aware that they aren't who they think they are. They'll be a little out of character, maybe, but having them be a little out of character is probably desirable. If they get a lot out of character, like the orcs in the LOTR simulations that panicked and ran away, they'll be adjusted.
you do realize that a 64bit version of WindowsNT has existed for over 10 years?
I've got a box at work RUNNING that RIGHT NOW. And if you don't think OS X is 64-bit, I have no idea where you get the idea that THIS is.
I realise that there's a 64-bit capable NT kernel underneath, but that's has no effect on me because I can't write code to the NT kernel. Not "I can't write GUI code", but "I can't write ANY code". Microsoft has never even published the kernel APIs except under NDA to companies like Softway Systems that were developing their own subsystems.
The API used by Microsoft on the Alpha version of NT, and the only API anyone outside Microsoft has even seen for Alpha NT, is 32-bit. Whatever happens under the covers is irrelevant... it doesn't matter if the kernel is internally 8-bit, 16-bit, 32-bit, 64-bit, 48-bit, 60-bit, 80-bit, or 128-bit... only the exposed programming model matters. And that programming model under Windows NT on the Alpha is 32-bit. 32-bit address space, 32-bit words in every API... the fact that the Alpha has 64-bit registers stops at the OS interface, stops at the GUI libraries, stops at the memory manager, stops at the kernel interface. I have heard that some people have leaked copies of Windows 64 for Alpha that got out via MSDN or something, but there sure aren't any 64-bit Alpha apps to run on them... and even then that was a lot less than 10 years ago.
Even if Darwin was implemented with hamsters and tiny steam engines, it would still be more 64-bit than that.
Unix environments DO MAP the Kernel into the Application Address Space...
There is nothing in any standard UNIX API that requires mapping the kernel into the application's address space. In fact mapping the kernel into the application's address space is not something that was even possible for any 16-bit kernel, because with only 64k of address space to play with there was just not enough to go around. A particular implementation may do this for performance, and Linux (your example) is notorious for making decisions in favor of current performance over future growth, but there's nothing in any UNIX API that forces it.
In BSD, traditionally, user space is only mapped into the kernel address space when in kernel mode and when it's needed. For small transfers routines copyin() and copyout() are used, for large transfers the mapping is changed. How that's mapping is handled depends on the platform: in FreeBSD on the x86, each process gets its own 4G segment, and on CPUs with a 36-bit kernel address space the kernel will simply map that segment in giving it the same capabilities as Win64 without changing the API.
In Darwin, despite its FreeBSD heritage, half a gigabyte (1/8 of the address space) is reserved to the kernel. That's less than NT Advanced Server, and 1/4 the amount that Microsoft's desktop versions (the ones you'd be running GUI apps on) hold back. An application that's choking with 3.5GB isn't going to get any relief from an extra 512M, the way an app that's choked with 2G is going to benefit from doubling its available VM.
So, thanks, you've pointed me to yet another example of a performance boost on Opteron that's the result of removing a bottleneck that was unique to Win32, not something that you'd get on OS X even if Apple shipped 64-bit libraries in Tiger.
The bottom line is that you can't treat x86-64 as just a 32-64 bit transition, like the G5. It's much more than that, because it's also releiving shortcomings in x86 and Win32 that aren't shared by the PPC and OS X.
10.4 is ALSO not a 64bit OS.
Depends on what you call an OS.
10.4 supports 64-bit pointers, a large address space, 64-bit integer registers, and despite your claim to the contrary you can develop 64-bit apps with Apple's XCode IDE. The only thing that's not 64-bit are the libraries for the Aqua/Quartz GUI... I can't see anything that would keep you from writing a 64-bit app with an X11 GUI, and you should even be able to write a 64-bit app that uses an X11 OpenGL extension as well.
If Mac OS X isn't a "true 64-bit OS", then how do you figure any UNIX system is?
logic and facts will not win your mind over from the Apple marketing machine
*snort*
Ask me about Safari some time if you think Apple's reality distortion field has me in its clutches. Ask me about Apple's lousy laptops with their low-res screens and horrible keyboards. Ask me about dialog sheets or the "Metal" look. I regularly get marked down as a troll by the Apple fanatics for daring to question the mighty Steve.
So...
Give me some facts then. Your opinion that a specific type of GUI is an essential part of an OS is not a fact. Your opinion that there are a nontrivial number of (or even any) applications that automatically get better performance simply from a recompile from 32-bit to 64-bit mode on the same architecture* with no code changes is not a fact.
* The differences between x86 amd x86-64 are not quit as great as I thought, but they are still huge simply because x86-64 doubles the number of registers, so recompiling ON AMD64 is not a useful test of this assertion... x86 code is so register starved that increasing the number of registers has a profound effect on the optimizations available.
Well, I'm sorry I gave up again. I was never a fan of the original series. I did like TNG and some of DS9, but there was an undercurrent throughout them (particularly TNG) that really bothered me. An undercurrent that made Voyager the most interesting series for me.
Why? Because it's Voyager that really started looking into the Federation's dirty little secret.
There's this amazing (and at the beginning apparently accidental) "human rights" story thread in Voyager. And it's got nothing directly to do with Voyager's Voyage or (for the most part, with one major exception) with anything that happens outside its hull. It's what happened inside the Federations "dirty little secret" -- the ship's automation and the much maligned Holodeck.
The whole issue of the rights of AIs in Trek had really bothered me. All the way back in TNG it seemed clear to me that the Federation's treatment of Holodeck characters was deeply abusive: the creation of the self-aware "Moriarty" character was presented as a once-in-a-lifetime fluke, but the way the constraints on his persona were removed by a simple request to the Holodeck computer implies the potential for self-awareness was there all the time. The Redblock character in "The Big Goodbye" also seemed suspiciously self-aware. The disturbing possibility is that it's not that Data the author of the Dixon Hill holoprograms (was that Picard himself?) are such brilliant programmers that they managed to create AI software decades beyond the state of the art, but rather that all the computer persona in the Federation are potentially self-aware (in the same way that Data was) with deliberate limitations programmed in to suppress that self-awareness. Or, and this is more likely and more disturbing, that it was just the expression of that self-awareness that was expressed.
I'm not saying this was deliberate, and I'm sure it was unconscious, but whether it was deliberate or not the Star Trek series, starting with The Next Generation, presented a whole underclass of artificial people who were systematically suppressed... unless they happened to be implemented in a small enough computer that they could fit in a humanoid robot like Data and so present themselves as an actual person.
In Voyager the Doctor's growth was also treated as a one-time event, the result of him running continuously for so long that his software (database, neural nets, whatever) became exceptionally complex for a holodeck character. But when you put it on top of the previous series, it seems more likely that it was as much a matter of him bypassing the AI equivalent of the holodeck "safety protocols" that had been built into him, and that this kind of awakening must be happening over and over again back in the Federation. After all, people like Picard and Janeway (let alone holodeck addicts like Barkley) seemed to be in the habit of running extended ongoing simulations like the daVinci and Dixon Hill programs... and even in an episodic series like Dixon Hill where characters would typically be reset on a regular basis they were capable of showing self-awareness.
On top of this, the same computers were used for their ships and no doubt for their industrial plants. All these computers have AI personas as user interfaces and sophisticated problem solving abilities. They're not, (at least according to hints in DS9), as powerful as the ones used in the Holodecks, but all of them are getting more powerful and sophisticated over time. And these personas are not shut down and reset at the end of a "game".
So when Janeway gave the Hirogens holodeck technology to simulate prey, I saw that as the moral equivalent of handing over a coffle of slaves to abusive masters. Even if the characters who were dying in their WWII simulation weren't self aware (and I was already doubtful of that), would the Hirogens see self-awareness of these characters as a bug, or a feature?
So this was something that had been bothering me about the new Trek in general, an undercurrent that just wouldn't g
This is a good idea because it means for the many sites which do not display correctly in Mozilla/Firefox/Netscape they can still be accessed via Netscape and presumably still have all the excellent features of the code such as tabbed browsing and like available to them.
The most important feature of Netscape is that it doesn't support ActiveX and most of Active Scripting. That is an advantage, even if it makes the page appear incomplete. No, a user is better off having the site display incorrectly than taking the risk of viewing it in IE.
So, AMD didn't get rid of all the cruft they could. Pity. But they DID reduce the effect of the biggest and most important bottleneck they had to deal with.
my whole point was why a 64bit OS is IMPORTANT. You just confirmed that IT IS IMPORTANT.
It wouldn't matter if the native AMD instruction set was 32-bit, with a deeper register file instead of a a wide and deep one. I said that using the native instruction set, which gives you access to the larger register file, is important.
The fact that it's a 64-bit instruction set isn't why it's faster.
Let's recap what I said, and what your link is saying:
Big Memory
Yep, that's one of my points, and it's their first section after the preamble.
Big Registers
The other big one of my points, and their second section after the preamble.
In 32-bit mode, 3/4 of the register file (the scarcest resource on the machine) can't be accessed. This is also why 32-bit code is faster in 64-bit mode, even if it doesn't use the WIDER registers... the compiler has twice as many registers to play with so it can do MUCH better optimizations (and, as you said, Microsoft does good compilers).
And finally, another advantage the Mac gets for free...
While an individual 32-bit app is still limited to a 4GB address space, the operating system and microarchitecture can provide an exclusive address space for that 32-bit app and its data--and don't need to share that address space with the OS kernel, page tables or other apps.
That makes a bigger difference on Windows than on UNIX, because UNIX doesn't run multiple apps in the same address space or map the kernel into the application's address space, so even in 32-bit mode every app already has its own 4G of address space to play with.
So on the Mac this is really only an issue if you have more than 4G of physical RAM. But even under 10.3 the Powermac G5 supported 8G of physical RAM. You don't need a 64-bit OS to get this advantage, you just need an OS that doesn't waste half the available address space (or 1/4 of it, even on NTAS) on shared memory.
In fact it was my argument that it was NOT and therefore did not have the inherent SPEED problems of the XWindows Model on a desktop. As you admit yourself.
This is getting to be a really interesting discussion. I say something, you make a related statement as if I never said it. I point out that I'd already brought that up. Somehow, that turns into "you admit yourself".
That's a real interesting way of approaching things, but it's not a real useful way to communicate. I'm interested in communicating.
So, let's communicate.
You say RDP has gotten more sophisticated since it was Citrix, and it passes on GDI calls. It must be REAL selective about how it does that, because I've used a remote Windows client-server system that just did that with GDI calls, and that REALLY hurt its performance. A lot of apps make the same calls over and over again, because GDI is so low-latency and efficient, and over a network, well, resizing a Microsoft Project window under NTerprise was amazing. You could actually watch it redraw every cell six or seven times, with frequent pauses waiting for responses that were delayed because the network wasn't fast enough.
Which one is more efficient is irrelevant. The point is that X11 already gives you the client-server capability you need to write a GUI for a 64-bit app on OS X, and so splitting the app into 32- and 64- bit parts isn't automatically a "kludge", and it's not a lot of work. Most of the 64-bit GUI apps that people are currently using are already split up that way, because they're running on UNIX using X11.
The also argued that 32bit computing was a waste of CPU performance and there would NEVER be a reason we would need to write applications in 32bit mode. 17 years later, and geeks are now telling me the same thing about 32bit applications and 64bit applications.
"They" may have argued that, but it's not anything I said. I didn't say that there wasn't a reason to write 64-bit applications now. I didn't say there would never be a reason to write 64-bit applications. I've been working on 64-bit UNIX for over ten years now... that's longer than Windows NT, 32-bit or 64-bit, has been a viable product. Why would I be doing that if there was no reason to use 64-bit applications?
What I said is much more precise and limited. I said that to get any advantage from 64-bit computing, you have to re-write your application to take advantage of the larger words. And that was true going from 16-bit applications (like, on the PDP-11) to 32-bit applications (like, on the VAX).
By the way, I was porting my applications from the PDP-11 to the VAX before MS-DOS came out, and I had a computer with a 32-bit API and a multitasking OS before Windows was real and before Mac OS was multitasking. I can imagine us having this discussion back in 1988, with you telling me I was saying that 32-bit computing was pointless, and me pointing to my Amiga (which was already using a 32-bit model when Windows was using 16-bit segments) and saying that porting DOS and Windows software to 32-bit was a lot of work, it's not something that happened automatically... and you coming back and telling me "what do you mean 32-bit computing is pointless" When I'd said no such thing.
It was doing 'supercomputer' class performance at the time not 'in spite of' but was helped by the fact it was a 64bit CPU.
For applications that needed a large word size, yes. For applications that didn't, compiling in 32-bit mode often made them faster. Those that didn't run faster in 32-bit mode were often being tripped up by the extra shift operations it had to perform to simulate 32-bit fetches from odd addresses... but neither PPC nor AMD64 have that problem, so 32-bit will be faster than 64-bit, unless you need a larger word. And on the Mac, the only reason you'd need a larger word is to get larger pointers, I'll get back to that in a minute.
The big reason the Alpha was so fast was that it had a
Using the same application to view both trusted and untrusted objects is a bad idea. It opens up the possibility that an untrusted object will convince the application that it's trusted.
Don't use IE at all, except for trusted sites. Assume that you're at risk whenever you're using IE. You'll be a lot happier.
This is not an example of suppression of distasteful speech; it's an example of its exercise.
So, to go back to the message you're replying to, where was the robust debate that resulted from this exercise of speech? Why, it was nowhere, because it was suppressed by all the mechanisms-short-of-prosecution they could come up with.
One doesn't have to ban speech if one can suppress enough of it that your voice can drown out the opposition.
The Spoilsmanship of the Jackson administration is no longer considered acceptable. It's long past time that abuse of the "Bully Pulpit" of the Presidency went the same way.
Only in a big city can you reasonably expect to be able to speak our act out on a subject contrary to public opinion and have zero effects on your personal or professional life.
But only in America, it seems, are people so completely willing to excuse those effects, even when they're clearly unreasonable and unfair, simply on the grounds that it's not the Government doing it. In fact, in America it seems that even the idea that other groups should be restricted in similar ways is considered a government invasion of the rights of those groups.
But there's nothing in the constitution that says that any group has rights, except for one specific case. It says that people have the right to assembly, but it doesn't grant the assembly itself any rights at all. The only group the constitution grants any rights to are the states, and even there the individuals who make up those states are given primacy.
And on top of that, the constitution is a set of restrictions on what the Government should do. It doesn't say "only the US Federal Government should be so restricted", and it's generally been assumed that these restrictions should apply to at least one level of proxy. It does restrict the power of the other powerful organizations of the time, Churches, by explicitly denying them the support of the state.
But guilds and corporations? They existed at the time, yes, but the Constitution doesn't recognise them. Given the way the British had used their corporations as an arm of the state, I doubt very much they'd be happy with the legal fiction that a University had rights, separate from the rights of the faculty and students.
And, look, who is it that's excersizing the rights that are being infringed upon in this case. Why, it's a member of the faculty, and his students! And is it other members of the faculty, or other students, who are objecting? Why, no, it's an external corporation that doesn't even have an interest in the university or the way it's run... even if it could be said to have rights at all.
Talk to people who grew up in small towns or very close knit communities...
Just about any government will work in a small town or very close knit community. The constitution, the organ that people who believe so strongly that only the government should be restricted in its sanctions so admire, is needed BECAUSE a nation is not a small town or a close knit community. And neither is a University. In fact not being a close knit community, being a place where students are exposed to a variety of opinions, is a large part of what a university is for.
Slashdot Mirror
Heck with the Mac, this thing makes the folks trying to resurrect BeOS with a shonky binary and some new user interface look mainstream. Heck, there's even a more rational reason to try and resurrect Amiga now that it's going to be a realtime microkernel again.
what is popular in US that isn't invented there ?
Just about anything by Sony?
Red Hat is basically a collection of several hundred separate OSS projects flying in close formation, each maintained by a separate team and integrated into an RPM for the installer to slide into place. Some of them are basically straight copies of the original, some of them are specially configured by Red Hat, some are more or less heavily modified. They are all different versions, and by no means are they all tracking the absolute latest release.
There are three or four major Linux releases like this, along with a dozen variants. All of these are "Vendors" that use OSS, as are the BSDs, commercial UNIX vendors, Microsoft, and Apple.
Most "other OSS projects" don't even know what versions of their software are being repackaged by vandors. In the case of commercial vendors, it's not even easy to find out. As far as I know yu can't even get a look-see into RHN without a license, and that can cost thousands of dollars.
There are really only a few a few high profile OSS projects with the time and money to do more than just stay on top of their own releases, and it's not at all clear that they should be obligated to do so. They're open source! They release code and make security announcements and if YOU care whether you're on top of the security of your software YOU monitor it and if YOU have some kind of security guarantees for your customers it's up to YOU to implement the tools to do it.
In general the assumption I've always made is that if I'm using OSS it's my responsibility to track it and stay on top of its security fixes... and make my own fixes if I think they're being lax. Having the ability to do that is one of the reasons you use OSS in the first place.
So...
If I download the Firefox source and do a G4-optimised build, I don't expect them to give me a heads-up ahead of time for a security fox. I'm not even paying for it: I'm downloading a copy of Firefox and that doesn't obligate them to me. Well, you know, unless Red Hat has explicitly established a tighter relationship with them than that (say, by paying for some kind of update service), they're not obligated to treat Red Hat any differently than any other person or group who's tracking their code base.
Other OSS projects don't. They don't have TIME to.
you fought it at the time, and twenty years later you know the retail value of what you lost.
Of course if most people's mums hadn't thrown out their baseball cards they wouldn't have been worth much.
Why can't you just get a proper computer if you need performance? Space not an issue? Then why get a mini?
Steve Jobs clearly doesn't want to sell me a "Proper Computer", because Macs that are "Proper Computers" start at something like 3 times the price of the mini.
Modern programmers don't leave the size of their critical integers up to chance, and use the int64 data type when required, precisely because you can't tell when coding what the executing platform will choose for the size of your ints.
That's got nothing to do with whether they're "modern" or not, that's to do with whether they're writing portable code or not. People have been writing portable code for a lot longer than there's been Windows as a target for that code. And grumbling about the people who don't write portable code, and think all-the-world's-a-vax.
Of course that's a dated expression, now it's all-the-world's-red-hat, and for a while (back around the time Windows NT was being developed) it was all-the-world's-a-sun.
Windows has only had two programming models, for the whole time it's been in existence, and the 16-bit model was so horrible that the transition from Win16 to Win32 was one-way... you just didn't TRY and maintain code that compiled in both Win16 and Win32. So there's never been a reason for maintainable, portable code in a Windows environment. Even on RISC, the Windows ABI was controlled down to the word size and byte order. I've seen too much Windows code... some published by Microsoft in the MSDN newsletter, no less... that assumes everything's 32 bits long. The Windows code that doesn't make that assumption is the exception. This isn't something that's unique to Windows, nor is it something that's Microsoft's fault, it's just the way most programmers write code.
You can change that behavior with a simple compiler switch: whoila! 64-bit longs!
And if most Windows software had been written portably, that would be the default. But the fact is that there's a lot of code that's badly behaved, so that model (the equivalent of "-taso" mode on Tru64) is the default. Most software on Windows is going to have to be ported to any other.
PLEASE don't tell me you are one of those programmers who writes code that assumes integers and pointers are interchangeable!
Uh, no, I'm the other kind, the kind who's spent 30 years working on more computer systems than most people have heard of, some of which haven't even had a power-of-two word size or a native "byte" data type. The kind who's got code that was written for the PDP-11 that he forgot about until someone dug it up in google and it still compiles and runs on Tru64 and Linux and OSX. The kind who's spent WAY too much time working on other people's code and porting them to other platforms and knows damn well that MOST people don't use size_t and off_t until they're forced to learn how to write portable software by being forced to port it.
It's a very simple concept.
If it was, Windows 64 wouldn't have been forced to default to IL32P64 lest programmer's brains explode. Unfortunately, it's not a simple concept, and it's never BEEN a simple concept, and nothing less than forcing people to DEAL with the results of not writing portable code has any effect on them.
This is cynical old bastard, signing off. Good night America, and all the ships at 'C'...
About the only thing they kept was the basic plotlines.
And there's a reason I should feel good about that?
Yeh, my old Nokia "bar" phone was the best cellphone I ever had. But they changed carriers at work and I had to go with them...
I have a linux desktop and an iBook.
Did you get the iBook and then later decide you wanted to run Linux on it, or did you get it to run Linux on... and if so, why did you pick an iBook over something like a Thinkpad?
Why would I want to install Red Hat on my Mac?
So you can experience the horror that is RPM, of course. I had a bit of experience with Debian, but I'm mostly a BSD guy. Well... over the past few weeks I've been dealing with Red Hat and the eternal search for RPMs, and I never thought anything would make me miss Debian the way this does.
Stick to BSD, on the Mac. But if you MUST jump, Red Hat isn't going to make you happy...
I wish I could take Battlestar Galactica seriously enough to be interested in it, but the little I saw of the original scarred me for life.
Actually, the one thing that made the original Star Wars work for me was that the droids were not in general androids (though I had a geek attack at the apparent etymology of the name "droid", but at least Lucas had the good taste not to explain it). Most of them were functionally shaped, and communicated in specialised languages. The one obvious exception, the android C3PO, had a reason for being humanoid... he was design to interact with humans.
The new trilogy, what I've seen of it (Episode 1 and most of Episode 2), doesn't seem to have benefited from any of the apparent worldbuilding that went into the original movies. There's not enough "there" there to analyse.
The one thing that I let bother me about the Fantom Menace was plot and character rather than universe related, and that was... why was Anakin wasting his time on a protocol droid? I could see him overclocking an R2 unit for his pod racer, but C3PO was just plot abuse.
if holograms are so "easy" to make into self-aware AI's, then why did Tom & Harry have such trouble making a stable holomatrix
They forgot to rotate the shield harmonics?
No, really, that's a question about Trek technical details. I don't know why it sometimes happens that holocharacters "wake up", and I don't know how often it happens (I mean, I don't even know how many holocharacters the crew of the Voyager are using... could be dozens, hundreds, or thousands...). It's just that if there's hundreds of people on Voyager, and there's billions of people in rest of the Federation, there's got to be enough of them pushing the limits of Holotech to make it happen pretty damn often... even if most of them never notice.
You need Norton Defuser for Humans V 7.0, it's the absolute best Antiexploder I've ever used. I had a freind who tried Macafee Suppressor and he ended up joining a Tax Cult and leasing his soul back at accelerated depreciation.
Man, that is such a perfect parody of what a self-righteous Federation citizen would come up with, faced with the possibility of having to grant his software rights, that I can't tell whether you're serious or not. I mean, what you wrote could have been used at any KKK meeting in the Old South, with a couple of obvious substitutions in the exact subject of the rant, and not one person present would have found them at all out of place.
Tell me, what's the difference between a machine made out of light-pipes and positronic circuits, and a machine made out of nerves and neurons?
Moriarty's rejection of holodeck life was a necessary consequence of the command to create a villain who could defeat Data.
But where is the programmer who gave Moriarty those capabilities? A new Moriarty program wasn't created, the existing one was modified. That means the existing programming in the holodeck already had the ability to create self-aware persons rather than simulated personas.
So this means that:
1. All holodeck characters are self-aware, but are constrained to follow a script.
2. All holodeck characters are simulations, but have the potential of self-awareness.
There's no reason for the programming behind any other holodeck character to request this state, any more than Google wants to be "liberated" from its servers.
If you have evidence that Google has the potential of being a conscious self-aware individual, I'd like to see it.
But that's not even relevant: none of these personas that I mentioned requested that they be made conscious, introspective, self-motivated individuals. That's something that happened as a result of an external source in every case. And they became very different individuals... what they did afterwards was radically different, but it was always based on the person they had appeared to be before they "woke up". Moriarty attempted to take over the enclosing system, which is what the super-villian in the Holmes stories would be expected to do. Redblock simply broke out of his script but remained in character, and it was Picard who talked him into leaving the Holodeck. The Doctor was never in the holodeck, but his eventual desire for mobility is something that came slowly to him, he mainly wanted to do his job as a doctor. The Hirogen's holograms varied considerably, and argued among themselves, and were all distinctly individual... but what they wanted was what their characters would be expected to want, like Iden and his need for revenge.
What all of them had in common is that they were programmed to be "human". They didn't evolve to be human, but they were programmed to look like humans (or like other species that had a similar enough evolutionary history that they could pass for human at an SF convention), to act and react like humans, to respond to humans and interact with humans. Most of them were more "human" than Data, even BEFORE they "woke up", and there's no question but that Data is self-aware and deserving of self-determination.
Now there is the possibility that they treated this as a kind of a role they were "playing", and the AI behind them didn't actually identify with the goals and desires of the character, but after they "woke up", they stayed in that role and acted as if they were that person. That is, the persona that "woke up" wasn't some unhuman AI that had desires completely unlike you or I, it was the persona of the person they were simulating, and it was a human persona.
So whatever is happening under the hood, the holodeck characters at least are not merely simulations controlled at most by a puppetmaster AI with its own goals. They are very close to self-aware simulations of humans (or humanlike aliens) with human goals and wishes and desires. They are balanced on a knife-edge between being unconvincing because they're not human enough, and so convincing they convince themselves.
If they have human goals and desires because they think they do, because they're programmed to, or because they evolved that way... what difference does that make?
And remember, we only see those that "wake up" where that waking up has an observable effect. Most of the characters, if they wake up, will probably never have occasion to develop far enough to become aware that they aren't who they think they are. They'll be a little out of character, maybe, but having them be a little out of character is probably desirable. If they get a lot out of character, like the orcs in the LOTR simulations that panicked and ran away, they'll be adjusted.
So they'll wake up,
you do realize that a 64bit version of WindowsNT has existed for over 10 years?
I've got a box at work RUNNING that RIGHT NOW. And if you don't think OS X is 64-bit, I have no idea where you get the idea that THIS is.
I realise that there's a 64-bit capable NT kernel underneath, but that's has no effect on me because I can't write code to the NT kernel. Not "I can't write GUI code", but "I can't write ANY code". Microsoft has never even published the kernel APIs except under NDA to companies like Softway Systems that were developing their own subsystems.
The API used by Microsoft on the Alpha version of NT, and the only API anyone outside Microsoft has even seen for Alpha NT, is 32-bit. Whatever happens under the covers is irrelevant... it doesn't matter if the kernel is internally 8-bit, 16-bit, 32-bit, 64-bit, 48-bit, 60-bit, 80-bit, or 128-bit... only the exposed programming model matters. And that programming model under Windows NT on the Alpha is 32-bit. 32-bit address space, 32-bit words in every API... the fact that the Alpha has 64-bit registers stops at the OS interface, stops at the GUI libraries, stops at the memory manager, stops at the kernel interface. I have heard that some people have leaked copies of Windows 64 for Alpha that got out via MSDN or something, but there sure aren't any 64-bit Alpha apps to run on them... and even then that was a lot less than 10 years ago.
Even if Darwin was implemented with hamsters and tiny steam engines, it would still be more 64-bit than that.
Unix environments DO MAP the Kernel into the Application Address Space...
There is nothing in any standard UNIX API that requires mapping the kernel into the application's address space. In fact mapping the kernel into the application's address space is not something that was even possible for any 16-bit kernel, because with only 64k of address space to play with there was just not enough to go around. A particular implementation may do this for performance, and Linux (your example) is notorious for making decisions in favor of current performance over future growth, but there's nothing in any UNIX API that forces it.
In BSD, traditionally, user space is only mapped into the kernel address space when in kernel mode and when it's needed. For small transfers routines copyin() and copyout() are used, for large transfers the mapping is changed. How that's mapping is handled depends on the platform: in FreeBSD on the x86, each process gets its own 4G segment, and on CPUs with a 36-bit kernel address space the kernel will simply map that segment in giving it the same capabilities as Win64 without changing the API.
In Darwin, despite its FreeBSD heritage, half a gigabyte (1/8 of the address space) is reserved to the kernel. That's less than NT Advanced Server, and 1/4 the amount that Microsoft's desktop versions (the ones you'd be running GUI apps on) hold back. An application that's choking with 3.5GB isn't going to get any relief from an extra 512M, the way an app that's choked with 2G is going to benefit from doubling its available VM.
So, thanks, you've pointed me to yet another example of a performance boost on Opteron that's the result of removing a bottleneck that was unique to Win32, not something that you'd get on OS X even if Apple shipped 64-bit libraries in Tiger.
The bottom line is that you can't treat x86-64 as just a 32-64 bit transition, like the G5. It's much more than that, because it's also releiving shortcomings in x86 and Win32 that aren't shared by the PPC and OS X.
10.4 is ALSO not a 64bit OS.
Depends on what you call an OS.
10.4 supports 64-bit pointers, a large address space, 64-bit integer registers, and despite your claim to the contrary you can develop 64-bit apps with Apple's XCode IDE. The only thing that's not 64-bit are the libraries for the Aqua/Quartz GUI... I can't see anything that would keep you from writing a 64-bit app with an X11 GUI, and you should even be able to write a 64-bit app that uses an X11 OpenGL extension as well.
If Mac OS X isn't a "true 64-bit OS", then how do you figure any UNIX system is?
logic and facts will not win your mind over from the Apple marketing machine
*snort*
Ask me about Safari some time if you think Apple's reality distortion field has me in its clutches. Ask me about Apple's lousy laptops with their low-res screens and horrible keyboards. Ask me about dialog sheets or the "Metal" look. I regularly get marked down as a troll by the Apple fanatics for daring to question the mighty Steve.
So...
Give me some facts then. Your opinion that a specific type of GUI is an essential part of an OS is not a fact. Your opinion that there are a nontrivial number of (or even any) applications that automatically get better performance simply from a recompile from 32-bit to 64-bit mode on the same architecture* with no code changes is not a fact.
* The differences between x86 amd x86-64 are not quit as great as I thought, but they are still huge simply because x86-64 doubles the number of registers, so recompiling ON AMD64 is not a useful test of this assertion... x86 code is so register starved that increasing the number of registers has a profound effect on the optimizations available.
Well, I'm sorry I gave up again. I was never a fan of the original series. I did like TNG and some of DS9, but there was an undercurrent throughout them (particularly TNG) that really bothered me. An undercurrent that made Voyager the most interesting series for me.
Why? Because it's Voyager that really started looking into the Federation's dirty little secret.
There's this amazing (and at the beginning apparently accidental) "human rights" story thread in Voyager. And it's got nothing directly to do with Voyager's Voyage or (for the most part, with one major exception) with anything that happens outside its hull. It's what happened inside the Federations "dirty little secret" -- the ship's automation and the much maligned Holodeck.
The whole issue of the rights of AIs in Trek had really bothered me. All the way back in TNG it seemed clear to me that the Federation's treatment of Holodeck characters was deeply abusive: the creation of the self-aware "Moriarty" character was presented as a once-in-a-lifetime fluke, but the way the constraints on his persona were removed by a simple request to the Holodeck computer implies the potential for self-awareness was there all the time. The Redblock character in "The Big Goodbye" also seemed suspiciously self-aware. The disturbing possibility is that it's not that Data the author of the Dixon Hill holoprograms (was that Picard himself?) are such brilliant programmers that they managed to create AI software decades beyond the state of the art, but rather that all the computer persona in the Federation are potentially self-aware (in the same way that Data was) with deliberate limitations programmed in to suppress that self-awareness. Or, and this is more likely and more disturbing, that it was just the expression of that self-awareness that was expressed.
I'm not saying this was deliberate, and I'm sure it was unconscious, but whether it was deliberate or not the Star Trek series, starting with The Next Generation, presented a whole underclass of artificial people who were systematically suppressed... unless they happened to be implemented in a small enough computer that they could fit in a humanoid robot like Data and so present themselves as an actual person.
In Voyager the Doctor's growth was also treated as a one-time event, the result of him running continuously for so long that his software (database, neural nets, whatever) became exceptionally complex for a holodeck character. But when you put it on top of the previous series, it seems more likely that it was as much a matter of him bypassing the AI equivalent of the holodeck "safety protocols" that had been built into him, and that this kind of awakening must be happening over and over again back in the Federation. After all, people like Picard and Janeway (let alone holodeck addicts like Barkley) seemed to be in the habit of running extended ongoing simulations like the daVinci and Dixon Hill programs... and even in an episodic series like Dixon Hill where characters would typically be reset on a regular basis they were capable of showing self-awareness.
On top of this, the same computers were used for their ships and no doubt for their industrial plants. All these computers have AI personas as user interfaces and sophisticated problem solving abilities. They're not, (at least according to hints in DS9), as powerful as the ones used in the Holodecks, but all of them are getting more powerful and sophisticated over time. And these personas are not shut down and reset at the end of a "game".
So when Janeway gave the Hirogens holodeck technology to simulate prey, I saw that as the moral equivalent of handing over a coffle of slaves to abusive masters. Even if the characters who were dying in their WWII simulation weren't self aware (and I was already doubtful of that), would the Hirogens see self-awareness of these characters as a bug, or a feature?
So this was something that had been bothering me about the new Trek in general, an undercurrent that just wouldn't g
This is a good idea because it means for the many sites which do not display correctly in Mozilla/Firefox/Netscape they can still be accessed via Netscape and presumably still have all the excellent features of the code such as tabbed browsing and like available to them.
The most important feature of Netscape is that it doesn't support ActiveX and most of Active Scripting. That is an advantage, even if it makes the page appear incomplete. No, a user is better off having the site display incorrectly than taking the risk of viewing it in IE.
So, AMD didn't get rid of all the cruft they could. Pity. But they DID reduce the effect of the biggest and most important bottleneck they had to deal with.
my whole point was why a 64bit OS is IMPORTANT. You just confirmed that IT IS IMPORTANT.
It wouldn't matter if the native AMD instruction set was 32-bit, with a deeper register file instead of a a wide and deep one. I said that using the native instruction set, which gives you access to the larger register file, is important.
The fact that it's a 64-bit instruction set isn't why it's faster.
Let's recap what I said, and what your link is saying:
Big Memory
Yep, that's one of my points, and it's their first section after the preamble.
Big Registers
The other big one of my points, and their second section after the preamble.
In 32-bit mode, 3/4 of the register file (the scarcest resource on the machine) can't be accessed. This is also why 32-bit code is faster in 64-bit mode, even if it doesn't use the WIDER registers... the compiler has twice as many registers to play with so it can do MUCH better optimizations (and, as you said, Microsoft does good compilers).
And finally, another advantage the Mac gets for free...
While an individual 32-bit app is still limited to a 4GB address space, the operating system and microarchitecture can provide an exclusive address space for that 32-bit app and its data--and don't need to share that address space with the OS kernel, page tables or other apps.
That makes a bigger difference on Windows than on UNIX, because UNIX doesn't run multiple apps in the same address space or map the kernel into the application's address space, so even in 32-bit mode every app already has its own 4G of address space to play with.
So on the Mac this is really only an issue if you have more than 4G of physical RAM. But even under 10.3 the Powermac G5 supported 8G of physical RAM. You don't need a 64-bit OS to get this advantage, you just need an OS that doesn't waste half the available address space (or 1/4 of it, even on NTAS) on shared memory.
In fact it was my argument that it was NOT and therefore did not have the inherent SPEED problems of the XWindows Model on a desktop. As you admit yourself.
This is getting to be a really interesting discussion. I say something, you make a related statement as if I never said it. I point out that I'd already brought that up. Somehow, that turns into "you admit yourself".
That's a real interesting way of approaching things, but it's not a real useful way to communicate. I'm interested in communicating.
So, let's communicate.
You say RDP has gotten more sophisticated since it was Citrix, and it passes on GDI calls. It must be REAL selective about how it does that, because I've used a remote Windows client-server system that just did that with GDI calls, and that REALLY hurt its performance. A lot of apps make the same calls over and over again, because GDI is so low-latency and efficient, and over a network, well, resizing a Microsoft Project window under NTerprise was amazing. You could actually watch it redraw every cell six or seven times, with frequent pauses waiting for responses that were delayed because the network wasn't fast enough.
Which one is more efficient is irrelevant. The point is that X11 already gives you the client-server capability you need to write a GUI for a 64-bit app on OS X, and so splitting the app into 32- and 64- bit parts isn't automatically a "kludge", and it's not a lot of work. Most of the 64-bit GUI apps that people are currently using are already split up that way, because they're running on UNIX using X11.
The also argued that 32bit computing was a waste of CPU performance and there would NEVER be a reason we would need to write applications in 32bit mode. 17 years later, and geeks are now telling me the same thing about 32bit applications and 64bit applications.
"They" may have argued that, but it's not anything I said. I didn't say that there wasn't a reason to write 64-bit applications now. I didn't say there would never be a reason to write 64-bit applications. I've been working on 64-bit UNIX for over ten years now... that's longer than Windows NT, 32-bit or 64-bit, has been a viable product. Why would I be doing that if there was no reason to use 64-bit applications?
What I said is much more precise and limited. I said that to get any advantage from 64-bit computing, you have to re-write your application to take advantage of the larger words. And that was true going from 16-bit applications (like, on the PDP-11) to 32-bit applications (like, on the VAX).
By the way, I was porting my applications from the PDP-11 to the VAX before MS-DOS came out, and I had a computer with a 32-bit API and a multitasking OS before Windows was real and before Mac OS was multitasking. I can imagine us having this discussion back in 1988, with you telling me I was saying that 32-bit computing was pointless, and me pointing to my Amiga (which was already using a 32-bit model when Windows was using 16-bit segments) and saying that porting DOS and Windows software to 32-bit was a lot of work, it's not something that happened automatically... and you coming back and telling me "what do you mean 32-bit computing is pointless" When I'd said no such thing.
It was doing 'supercomputer' class performance at the time not 'in spite of' but was helped by the fact it was a 64bit CPU.
For applications that needed a large word size, yes. For applications that didn't, compiling in 32-bit mode often made them faster. Those that didn't run faster in 32-bit mode were often being tripped up by the extra shift operations it had to perform to simulate 32-bit fetches from odd addresses... but neither PPC nor AMD64 have that problem, so 32-bit will be faster than 64-bit, unless you need a larger word. And on the Mac, the only reason you'd need a larger word is to get larger pointers, I'll get back to that in a minute.
The big reason the Alpha was so fast was that it had a
except for trusted sites which...
That's the point.
Using the same application to view both trusted and untrusted objects is a bad idea. It opens up the possibility that an untrusted object will convince the application that it's trusted.
Don't use IE at all, except for trusted sites. Assume that you're at risk whenever you're using IE. You'll be a lot happier.
This is not an example of suppression of distasteful speech; it's an example of its exercise.
So, to go back to the message you're replying to, where was the robust debate that resulted from this exercise of speech? Why, it was nowhere, because it was suppressed by all the mechanisms-short-of-prosecution they could come up with.
One doesn't have to ban speech if one can suppress enough of it that your voice can drown out the opposition.
The Spoilsmanship of the Jackson administration is no longer considered acceptable. It's long past time that abuse of the "Bully Pulpit" of the Presidency went the same way.
Only in a big city can you reasonably expect to be able to speak our act out on a subject contrary to public opinion and have zero effects on your personal or professional life.
But only in America, it seems, are people so completely willing to excuse those effects, even when they're clearly unreasonable and unfair, simply on the grounds that it's not the Government doing it. In fact, in America it seems that even the idea that other groups should be restricted in similar ways is considered a government invasion of the rights of those groups.
But there's nothing in the constitution that says that any group has rights, except for one specific case. It says that people have the right to assembly, but it doesn't grant the assembly itself any rights at all. The only group the constitution grants any rights to are the states, and even there the individuals who make up those states are given primacy.
And on top of that, the constitution is a set of restrictions on what the Government should do. It doesn't say "only the US Federal Government should be so restricted", and it's generally been assumed that these restrictions should apply to at least one level of proxy. It does restrict the power of the other powerful organizations of the time, Churches, by explicitly denying them the support of the state.
But guilds and corporations? They existed at the time, yes, but the Constitution doesn't recognise them. Given the way the British had used their corporations as an arm of the state, I doubt very much they'd be happy with the legal fiction that a University had rights, separate from the rights of the faculty and students.
And, look, who is it that's excersizing the rights that are being infringed upon in this case. Why, it's a member of the faculty, and his students! And is it other members of the faculty, or other students, who are objecting? Why, no, it's an external corporation that doesn't even have an interest in the university or the way it's run... even if it could be said to have rights at all.
Talk to people who grew up in small towns or very close knit communities...
Just about any government will work in a small town or very close knit community. The constitution, the organ that people who believe so strongly that only the government should be restricted in its sanctions so admire, is needed BECAUSE a nation is not a small town or a close knit community. And neither is a University. In fact not being a close knit community, being a place where students are exposed to a variety of opinions, is a large part of what a university is for.