Hello Everyone!
in [General Linux]
|
From: André Gillibert on 2 Dec 2009 07:32 Aragorn <aragorn(a)chatfactory.invalid> wrote: > > BTW, you can put a Linux system on a FAT file system too, with 777 > > permissions on all files & folders, and a tmpfs mounted on /dev. That > > can be done with an initrd. > > Yes, but if it is a pure FAT (or FAT32), then it wouldn't work > correctly, even if only because file ownerships and other extended UNIX > properties cannot be saved to the filesystem at shutdown - there is no > way to synchronize the information of the virtual filesystem layer with > the structure of the physical filesystem underneath. The lack of symbolic link may be a problem, but, with 777 permissions, I guess it would boot. Permissions are not saved on disk, but it's the same for FAT32 on WinNT... Programs can call the ACL API, but it has no effect. WinNT doesn't have umsdos, but, people are supposed to use NTFS. The main difference I see is that Linux was never designed to run on a FAT partition, and, if it can, that's purely accidental, while WinNT had to run on FAT to be compatible with Windows 3.x and MS-DOS. > Yes, but only with NetBEUI. The TCP/IP stack was "borrowed" from BSD, > and using the two together proved catastrophical for security, albeit > that the DOS-based Windows versions - i.e. Win95 and successors - had > both of them active at the same time, and it was left to the user to > disable NetBEUI. And maybe NetBIOS over IPX too. > Well, you obviously interpret the term "multi-user" as inclusive of the > diskspace-sharing - as for instance in a Novell Netware context - and > you are not the only person who interprets the term in this way, but > that is not what I myself mean by "multi-user". When I speak of > a "multi-user context", then I really do mean "the execution of > multiple processes owned by multiple users in the processor(s) and > memory of one and the same physical machine". > In that case, I agree. That can be seen in Windows network organization. In early nineties, the computer network of my university (Université de Caen Basse Normandie) was UNIX based. User workstations were light terminals (I don't know if they were text or graphical, I wasn't there at this time) connected to central servers, so that, user programs were all executing on these servers. The network has been replaced by WinXP machines. There's one domain controller, and all machines are complete personal computers with hard drives, 3D video card and high end CPU. When a user logs out, all his programs are exited. AFAIK, there's no way for a user to launch a daemon. This show the different network architecture of UNIX and WinXP. One admin resists to the Windows invasion. The sciences department has Linux based computers! > > Proprietary video drivers suck, but, even with free drivers I got a > > few issues. That's also sad to see that there is no nVidia free driver > > with textured 3D support. > > The "nouveau" project is working on that, but alas, they're not making > great advancements due to a lack of collaboration from nVidia and the > fact that it's an entirely volunteer-driven project, meaning that they > also don't have all the hardware available to reverse-engineer. > AFAIK, the only 3D features it has (in 2009) is untextured 3D for some cards. Reverse-engineering is fastidious. -- André Gillibert
From: André Gillibert on 2 Dec 2009 10:15 Peter Köhlmann <peter-koehlmann(a)t-online.de> wrote: > Aragorn wrote: > > 64-bit long mode does > > not have real mode compatibility via a "virtual 86" mode. > > Wrong. How come my VT-less 64bit AMD runs 16bit windows apps? > And 16bit DOS apps, under DOSemu as well as DOSBox? > > Unless you declare that I am lying, you have some real explanation to do > here. And no, those apps are *not* running in any emulation mode, they run > full speed > You're lucky, because I ran a test that showed how slow dosemu is on my Athlon64 3000+ computer with Arch Linux 64 bits edition, and how fast it is on my older 32 bits Pentium IV 2.67Ghz. I created the test program from debug: 0100 xor ax,ax 0102 xor cx,cx 0104 inc ax 0105 jnz 104 0107 inc cx 0108 jnz 104 010A mov ax,4C00 010D int 21 These are two nested loops, with 65536 steps in each. I saved it as test.com On my 32 bits Pentium IV, it runs in 2.8 second (+/- 0.2 because of my timing inaccuracy). On my 64 bits athlon with a 64 bits kernel and 64 bits dosemu, it runs in 56 seconds. -- André Gillibert
From: Stan Bischof on 2 Dec 2009 23:07 Wanna-Be Sys Admin <sysadmin(a)example.com> wrote: > > I don't think it's appropriate to always act like or claim it's all only > ever down to opinion, some things are simply better than others > (sometimes in a few ways, only one way, or in every way), though > sometimes they are not (in the case of Ford vs. Toyota and Windows vs. I'd have to disagree here. As soon as I ask "Is A better than B?" I have inherently created a situation in which the answer is almost surely a matter of opinion. Why? Because I have not in any way specified a criteria for the comparison so anyone answering can use any criteria they want to. "Is a Ford better than a Toyota" is a matter of opinion. "Is a Ford 4-door sedan more reliable over a typical 5-year lifetime than a comparibly priced Toyota" can actually be answered by something other than opinion. Unless you find an "A" that is demonstrably better than "B" in _every_ conceivable manner you end up with opinion if you do not specifiy a criterion. Such a comparison may exist but it would be darn rare and certainly wouldn't relate to anything complex like a car or an OS. Anyhow, back to it! Stan PS- sure a shame that there's so little real Linux information here these days. COLM used to have quite high SNR. Oh well.
From: Wanna-Be Sys Admin on 3 Dec 2009 02:10 Stan Bischof wrote: > Wanna-Be Sys Admin <sysadmin(a)example.com> wrote: >> >> I don't think it's appropriate to always act like or claim it's all >> only ever down to opinion, some things are simply better than others >> (sometimes in a few ways, only one way, or in every way), though >> sometimes they are not (in the case of Ford vs. Toyota and Windows >> vs. > > I'd have to disagree here. As soon as I ask "Is A better than B?" > I have inherently created a situation in which the answer is > almost surely a matter of opinion. Why? Because I have not > in any way specified a criteria for the comparison so > anyone answering can use any criteria they want to. > > "Is a Ford better than a Toyota" is a matter of opinion. No, it's not, or else no question about what product is better (performing, more reliable, etc.) can never be answered under the claim it's all going to be "opinion". That is utterly untrue. > "Is a Ford 4-door sedan more reliable over a typical > 5-year lifetime than a comparibly priced Toyota" can > actually be answered by something other than opinion. > > Unless you find an "A" that is demonstrably better than "B" > in _every_ conceivable manner you end up with opinion > if you do not specifiy a criterion. Such a comparison may > exist but it would be darn rare and certainly > wouldn't relate to anything complex like a car or an OS. The question is broad, so the answer is as well. That doesn't invalidate the quality of the generalized/broad answer. If an auto maker has every sedan and truck out perform another maker's sedans and trucks, in every single case, wouldn't it then be prudent to say that, yes, the one is better than the other? If the Linux OS offers the control, reliability and overall performance advantage over another (Windows), then it's "better". I don't see the point of suggesting or claiming that no one can ever answer a question about what one is better. Of course it can depend on the person and their application of the OS, but if you take two people that know the two operating systems equally, then Linux is better. > > PS- sure a shame that there's so little real Linux > information here these days. COLM used to have > quite high SNR. Oh well. Well, it's been like that for a while, in fact on most groups. This is about as close as we see to a Linux topic nowadays, sadly. Perhaps that's why I bothered to reply to this topic. In closing, it is my opinion that answering questions about cars and OS's being better than the other is not opinion. If someone states a fact, suggests people research the data about the comparable features to conclude which is better ("hands down" in many cases), then there's still no way to prevent people from saying "Well, it's all just opinion anyway". Balderdash, I say. What's the point of ever having a question or answer. It's my "opinion" that the Holocaust did happen, I would call it a fact, but there are those that call it opinion. Lame. -- Not really a wanna-be, but I don't know everything.
From: Loki Harfagr on 3 Dec 2009 13:08
Wed, 02 Dec 2009 23:10:09 -0800, Wanna-Be Sys Admin did cat : > In closing, it is my opinion that answering questions about cars and OS's > being better than the other is not opinion. Congratulations, you just introduced dependancy hell into the usage of question/answer ;-) > If someone states a fact, > suggests people research the data about the comparable features to > conclude which is better ("hands down" in many cases), then there's > still no way to prevent people from saying "Well, it's all just opinion > anyway". Correct, there's always that problem, there's no direct implication when someone asks a question that he'd not be an idiot. Well, just give him the facts again and wait ,-> > Balderdash, I say. Now that's a good question about metaphors, is 'Boulderdash' superior to 'The colossal cave' and was 'Civilization' better than 'Balance of Power' ? |