I use a abit bp6 - 5 PCI, 1AGP (Obviously) and 2 ISA.
Yes, I've used motherboards of a similiar size as well. But the post I was responding too wanted not only that, but a NIC, 5.1 Sound, Firewire, SCSI, and an extra IDE controller. I have an MSI K7 Master (1 AGP, 5 PCI, 1 CNR), which has the sound and space for the SCSI [it just has empty spots where the Adaptec chip and the connectors would be], and it's already huge. Adding Firewire, a NIC, and an extra pair of IDE connectors would push it over ATX size specs for sure.
As part of the thread, yes, the ISA bus hangs off the PCI for keyboard, serial, parallel, etc.
Yeah, I went home and looked. Too bad, really. I've always found it amusing that, without some serious messiness, it's impossible to not compile ISA support into a Linux kernel (on x86, at least - I hope it's not getting compiled into my SPARC kernels!)
Of course, if it actually worked 1% as well as advertised, it likely would be pretty good. My experience with USB has been pretty sucky.
Really? Too bad. I just got a nice G4 (at work) that's all USB for keyboard/mouse and that's been going fine. And I recently started using an USB Zip drive (I still use PS/2 for my keyboard, because I love using those huge old IBMs and there's no way anyone is making a USB version of that!)
However, I can understand why people will want to get rid of ISA. Oh, well, maybe someday...
you do not get 10% greater performance by buying a motherboard that has ni ISA slots (like those Asus KT boards). Because the fact is that even if they have no ISA slots, they still have a ISA bus built in the southbridge to support legacy stuff like the printer/parrallel port, the serial port/s & the PS2 mouse & keyboard ports.
For the most part, I would be willing to beleive you on this. However, is it not possible that modern chipsets use PCI or a seperate bus that does not interfere with the PCI bus to communicate with those devices? I'm really not sure; I suppose I'll go home and run lspci on my AMD-760 based sytem to see if there is an ISA bus lurking there.
In any case, ISA is a waste of space. I would take another PCI slot anyday.
They recommend a pretty well generic (though above average) Antec case, but this is s'pose to be a ultimate Linux box.
I would have to disagree with you here. I spent about 3 weeks researching what case to get, and decided on the Antec SX830 because of it's excellent overall design, room for plenty of fans, and well built drive tray. The slide out motherboard tray in the cases you recommend is all but useless to me, personally (other may, of course, disagree). Others by Supermicro, etc, were also discarded, partly on (lack of) certain features, and also because they were overly large.
In this particular case, I would have gone with a nice 2U rackmount case; after all, the Thunder K7 was built for rackmounting.
Your desire for massive amounts of integrated hardware, plus 6 PCI + 1 ISA, plus SMP, would end up, I'm pretty sure, with a motherboard too big to fit into most ATX cases.
but the Linksys cards (I forget the chip they use but it's not one of the ones you listed) have been solid
I've been using Netgear FA310TX cards (which are, apparently, the same thing as some Linksys cards - I'm kind of confused about the exact relation but that's what seems to be the case). Anyway, they are tulips, and I've never seen any problems like those mentioned running on a 100 Mbit network with plenty of traffic over the last 2+ years. I am highly skeptical of this supposed problem.
Intel EtherExpress isn't bad either, it's built into the motherboards of some of the machines around here and they seem to be working out pretty well.
But seriously, if you have a perfectly good Compaq case and power supply holding a 486 board and you find a good deal on a Gateway Socket 7 (or the other way around brand wise), it's really annoying that one's got the video jack where the other has the keyboard and mouse or vise versa
Yeah, OEMs do love to play those kinds of games. Too bad, but I guess they like being the single source (and being able to charge amounts which match that status for spare parts).
By the way, just what kind of old computer junk can you use? (if you weren't kidding)
KAI is not compiler in the usual sense though, its a translator C->C++ (last time i checked anyway),
True. But I think about it as a a compiler, because it translates a complex language (C++) into a simple(r) language (C), just like gcc does (except it's C++ -> some form of assembler).
hence bound to be slow(compilation time wise).
Not really. Faster than gcc 3.0.1, anyway (I haven't done any timing, but I would estimate about 1.5x the speed).
Maybe it's time to replace the whole ISA-PCI-IDE started out as an 8-bit platform and got patched and kludged time and time again mess with something that anticipates that what now seem like big drives, big RAM sticks, and fast processors and video cards will soon be classed with 8088s and 64K ram chips.
Say what you like about IA-64, at least Intel does seem intested in killing off all that crap fairly soon. I don't see why you group PCI in with ISA and IDE, however. PCI is a "real" bus, and while it's not perfect, it's high-performance enough for most things, and it's reasonably well designed. I imagine that 3GIO is going to be, for the most part, an extension of PCI to higher clock speeds, rather than completly reworking it (I haven't been able to find any actual info about anywhere, so I guess I could be wrong). Also going out over the next few years are PS/2 [I hope I can find a PS/2 -> USB adapter; I love my keyboard and want to keep using it], floppy drives, serial ports [which is rather unfortunate, lots of old Unix boxen can or have to use the serial port to get an install going], and parellel IDE, replaced by Serial ATA. It's kind of kludgy internally, but I guess from the user's perspective it will be a lot nicer. Similiar to firewire drives, but >4x faster.
un-upgradeable except by pitching them into the landfill and buying a whole new system.
It's funny that you seem to dislike the bad old standards so much, yet you also don't want to throw any of it way. I'm confused.:) You've got to pick, keep the crap, or dump it and everything related to it. But please, don't toss your old computers into a landfill. Give them to me!
With 3 or more new fast ATA/100 drives (limit 1 per cable) and cheap IDE controllers, you can run out of PCI bus bandwidth.
Potentially, anyway. I have a 7200 RPM ATA/66 drive, and usually only get about 23-26 Mbytes/sec when reading directly off the disk (no caching). Some fast IBM ATA/100 disks manage up to 37 Mbytes, which would saturate the bus if 4 of them were working non-stop. How realistic any of these sitautions are, is, of course, depends on the situation.
Of course, there's always 64-bit 66 Mhz PCI, 3GIO [someday], and firewire drives. USB 2.0 too, I guess.
I belive it when I see the numbers, although Borland got a reputation of isssuing quite fast compilers..
True enough, but what I worry about much more is whether it will be (roughly) C++ compliant or not. It would kind of suck if a vast majority of C++ apps couldn't be built with it.
So far I have heard mixed reports from people about ICC effiency in terms of code generated, speed of binaries, size of binaries etc (slashdot users who use ICC - please post your conclusions).
There are, in fact, several other C/C++ compilers for Linux besides those two, including KAI C++ and Portland Group C++ (I believe you mentioned PGCC in you post, actually). I'm quite certain that KAI, especially, is more widely used than Intel's compiler, because it runs on everything [Linux/x86 all the way up to Crays and S/390s], and has been around for some years.
I haven't used Intel's compiler, and I can't say much for C, but I will say this for C++: compared to gcc 2.95.2 and gcc 3.0.1, KAI C++ usually generates code that is faster (2x-3x) and smaller (about half). Since Intel recently acquired KAI, I imagine those properties (well, besidesthe portablity) have bled over to some extent into Intel's compiler. Portland Group C++, I've had nothing but trouble with.
Is MetroWerks a "real" compiler on Linux now? Last time I looked it was just a pretty IDE for gcc.
1. GCC compatible - you'll need it if you want to be used by open source users OR to allow developers to move their apps which used GCC to your compiler.
I don't think so; how many applications really use the GCC extensions? Not very many. All it has to be is ANSI C compatible, so it can use the normal glibc functions, and for the most part the problem is gone.
Hey, totally off topic, but your nick is based of the main character of Weis & Hickman's Deathgate Cycle, right? Just checking my memory, I'm not quite sure.
Uh... to help avoid getting modded as off-topic, yeah, LSB compliant systems would certainly help quite a bit. Fortunately, you know that on most Linux systems, and indeed on most Unix systems in general, you've got a/home,/var, etc. One cool thing for this hypothetical mover software could do would be to parse/etc/fstab and handle that for you when it installs the new distro.
One nice idea I had a while back as far as copying/backing up the passwd/group files is using something like LDAP to handle it. I've had to do what you did a few times (upgrading the NIS server, for example), and it's a real PITA (especially when there's ~50 users and a bunch of groups, all of which are vital to the organizations functioning). Unfortunately, in the main, most software doesn't really deal with having passwords anyplace but in the usual spots. It really would be nice if there were a better seperation between system and "real user" accounts, but, alas...
I one time forgot about the mail spools when moving over. Thank god this was my home machine and not something which a bunch of people would would be out for blood for. Though I did feel pretty dumb (soon after I changed my procmail config to dump _everything_ into a file in my home directory).
There has been a lot of talk that PGP is the only unbreakable encryption method out there
Which is complete bullshit, for 2 reasons:
1) PGP, like virtually any other form of cryptography, is breakable. It may not be breakable in a reasonable amount of time, but it is certainly breakable. Do not use "unbreakable" unless something is, in fact, unbreakable (within whatever environmentt you're placing it).
2) Why would PGP be particularly strong over other protocols? It's not. Hell, it doesn't even include MACs in the messages, which brings up all kinds of problems.
maybe because the RSA has finally found a way to break the PGP encryption?
Umm... wait, what? Perhaps you meant the NSA???
Anyway, maybe it was because they realized that they didn't have any kind of proof that PRZ had exported the code, that if they pursued it with criminal charges, more likely than not the unconstitutional restrictions on crypto would have been declared as such and lifted, and PGP was already widely available outside the US so they weren't getting much out of it.
It's standard practice that once you know how to break someone's code, you don't ever let them know which guarantees that you can keep on reading all their transmissions.
Wait... huh? From what I can get out of this, it would imply that the NSA even if the NSA could break the algorithms in PGP, they would keep up bothering Phil forever, to give people the impression that PGP was still a big threat to them.
I think you might have meant ``faster than a deterministic Turing machine'', which I wouldn't argue with.
Holy crap I'm dumb. What's really embarressing is that I took a complexity theory class last semester that went over all of this stuff; you'd think I could remember something as simple as that for at least 6 months.
BTW, I don't think your post was being pedantic at all; my original post was quite plainly false, and I thank you for pointing out my error.
JHU has one main one (jhu.edu), and most everyone has subdomains off of that. There is another one for the med school (jhmi.edu), and a few others for various organizations (jhuisi, jhuacm, etc).
However, invoking whois, I see that in fact jhu also owns jhu.{com,org}, jhuisi.{com,org}, jhmi.{com,org}, and various correct and incorrect spellings of johnshopkins.{com,org,edu} (though johnhopkins.net is for sale, it seems). Presumably they did this to prevent people from using them to make the university look bad, or whatever.
So I'm starting to guess that a lot of schools do this sort of thing, just like the various companies will squat on domains they have no real rights to (see the story on the Register about Reuters). But don't worry, even if they weren't doing this, I'm sure your college could find other ways to waste your tuition on frivoulous things. I know mine does.:)
That's what I don't trust--the hype and hubbub. Superpositional computation has tremendous theoretical possibilities, but superpositional computation in practice is... nowhere near useful. Ask me again in five years what I think and you might get a different answer.:)
OK, yeah, that's a pretty reasonable opinion to take at this point. OTOH, it does seem like at least some progress is being made; even though the ones made so far are tiny and pretty useless, the theories that they would be able to compute faster than a Turing machine model computer have been shown to be valid. But if someone can ever build one that actually does something a normal computer can't do... that is the question.
I've read some papers about getting quantum key exchange working; it's nasty. I don't know enough physics to understand about 60% of it, but I can get enough out of it to know that getting those quanta to do the right thing is a real PITA.
Overall, I would disagree with your statement. Fertilizer is pretty damn common, at least in rural areas where I grew up (if you own a farm, it's not too hard to convince someone to sell you a bunch of fertilizer). Knives are heavily regulated? What? Ever been to a steakhouse?
You cannot simple say banning==bad freedom==good unless your definition of good is anarchy.
True, but "No Constitutional Rights == bad". Banning encryption (or banning strong encryption, at least), would infringe upon my rights of free speech (at least according do, IIRC 9th circuit court of appeals, which held that software is speech, at least in some circumstances). Additionally, at least I believe that my right to free speech allows me to speak in encrypted form (how is that different from using a private slang to talk to a friend?). I don't know if that has been tested in any court, however.
Perhaps books about number theory, information theory, and network security should be banned as well? All can be used as tools to build encryption, just like if I wanted to buy detonators for a bomb.
Encryption is not a thing, it is an idea. Just like terrorism, actually. You think if every terrorist in the world was killed, more wouldn't start showing up? Get rid of the PLO, Hezbollah, and Islamic Jihad, and Palestinians will start being happy about having their houses bulldozed by Israel?
I don't trust QC of either sort because it depends on so much knowledge of physics and technical savvy that, were it to be fielded today, it would be hideously insecure by virtue of its implementation being so difficult to get right.
Quantum Crypto, I agree, is way to new to be trusted (not like anyone has figured out how to implement it in a practical manner yet anyway). But quantum computation: what is there to trust or not trust? It's just like an Athlon, only different; either it works, or it doesn't. Either it factors an RSA key, or it doesn't. It's not like it spits out an answer that you can't check easily.
I don't trust ECC, even though the Taniyama-Shimura Conjecture has been proven, because all of the good elliptic curves have been patented by Certicom and the remainder are either untrustworthy or too slow for practical use.
Well, that's a good reason not to use it, but not a good reason not to trust it. I mean, you (presumably) trusted some form of public key crypto prior to 1997, and all forms were covered by the RSA and Hellman-Merkle patents up until then (or at least RSADSI claimed they were covered, and was willing to sue anyone who disagreed). Anyway, IIRC Certicom only (or mostly) patented stuff using prime fields, leaving GF(2^N) wide open. Sure, you need a larger field for security, decreasing the benefits, but it's still faster and smaller than RSA or Elgamal.
It amazes me to see all these people on Ask Slashdot (I know these kinds of questions have been asked at least twice in the last year) assuming that new == good. So overall I would say you make some good points.
Just assemble and ship the product. If you don't know how to use it, don't buy it from us. A side benefit of this idea would be that "hard-core" computer geeks can stop wandering from site to site to build their computers. Since there's little over-head to cover, prices will stay cheap.
I'm skeptical that they could do anything, that, for example, mwave can't. It's a rare piece of (new) hardware that I don't buy from them. You can pay a little extra to get a full box assembled before they ship, but it's better to get it as is, IMHO. And you can, of course, order anything else you might want - full flexibility. Split your computer purchase over 2 orders (which is, in fact, what I'm doing right now - the video card, DVD-ROM, and sound card won't be ordered until next month). Get stuff that Dell would never sell you: you want 3 PCI Voodoo 5 5500s in your new box? - go for it [actually I think their stock of Voodoo5s ran out a while back, but you get the idea]. And get it damn cheap, at least compared to Dell, et. al.
Their policies are pretty much what you describe; their return procedures are fine [if it's broken, they'll replace it], they sell quality stuff, and there is no real support. But if you have questions, they will answer them, to the best of their ablities, in a quick and courteous manner. And they ship fast (like, same day, in most cases - next day, if you want them to assemble stuff for you).
Your hypothetical company could offer more products, etc, but that's about the only benefit I would see to it. And on the down side: QA problems [is there a warranty? If so, how do your make sure that your wide selection of stuff can't be mixed together to form something that doesn't work due to conflicts.], slower shipments than parts places (you have to built each one by hand, because there are a big selection of different parts - you can't pre-build common selections very well), and a new start-up is going to have a hard time getting business of that sort, because 90% of the people I know who are building a PC order everything from mwave.
Currently sitting in a room with 9 mwave machines...
and bars that take cash only are a relic of days gone by
I would say you simply don't like to hang out in hole in the wall dives much. Most of the "best" bars in the Baltimore area, for example, will only take cash (the stupid college-bar places will take CCs, though). I have seen vending machines that will take CCs, but never a pool table. Nor will most jukeboxes that I see. And the local liquor store: cash only.
Also, there are other places that strongly prefer cash, for example, tattoo places. They'll take plastic or check, but they like cash (who doesn't want to make $100/hour tax free?).
Your point about "Stripper Dollars" is interesting, but I think I'll stick to 5s for now.:)
"Monday 9/24, noon, at the Mattin Center: U.S. Representative Constance Morella (8th District of Maryland) will talk regarding Information Security and Privacy."
The Mattin Center is the new arts building on the campus of Johns Hopkins University. I'll be showing up and a hope others will as well.
If you want directions or more info please respond to this post.
IANAL, but aren't there already laws making it a felony to use the telephone or postal system, or to meet face to face to discuss illegal activies ("Conspiracy to...").
Indeed there are. Not to mention laws that allow a judge to ban you from talking to some set of other people (with a felony penalty if you are later caught with them). Kind of like a restraining order, but the other person doesn't want you to keep away from them.
Also, in a few cities, a police officer can prohibit you from talking to someone else, just for the hell of it (for example: you're talking with your friend in a coffee shop, and a cop, can, legally, go up and say "you - stop talking with that guy" and haul your ass of to jail if you refuse). These may have been deemed unconstitutional by now, but I'm not sure. They were nominally for anti-gang stuff, but they were apperantly mis-applied (the police were going on the theory: black -> in gang).
I think you misunderstood. I just meant that Java itself isn't slow. It's the class libraries - like the String class - that are the cause of most of the slowdown in your average application. If you are writing a large scale app, implementing a fast string class is completely trivial.
No, I understand what was meant. However, the libraries are fairly intrinsic to the language, especially those in the basic package. If the libraries are slow, or otherwise deficient, the language itself suffers (IMO), because either you take a performance hit (in terms of real application performance), or you have to reimplement basic functionality. Of course, I can reimplement a string type - but why should I have to? Just because Java can do things not involving it's standard library relatively fast doesn't mean a whole lot, if it's libraries are really slow, and thus making virtually every application actually written in Java slower than it needs to be.
For that matter, why doesn't someone just speed up the damn String class? Has Sun not heard of code profiling?
And even if you could build your own computer without Windows, you still can't build a laptop.
Actually, I believe one manufactuer (Acer?) is selling laptops without disks or memory. So no Windows, either. I think I saw it on arstechnica, or possibly anandtech. Or then again, maybe toms. One of the 3, anyway.
Slashdot Mirror
I use a abit bp6 - 5 PCI, 1AGP (Obviously) and 2 ISA.
Yes, I've used motherboards of a similiar size as well. But the post I was responding too wanted not only that, but a NIC, 5.1 Sound, Firewire, SCSI, and an extra IDE controller. I have an MSI K7 Master (1 AGP, 5 PCI, 1 CNR), which has the sound and space for the SCSI [it just has empty spots where the Adaptec chip and the connectors would be], and it's already huge. Adding Firewire, a NIC, and an extra pair of IDE connectors would push it over ATX size specs for sure.
As part of the thread, yes, the ISA bus hangs off the PCI for keyboard, serial, parallel, etc.
Yeah, I went home and looked. Too bad, really. I've always found it amusing that, without some serious messiness, it's impossible to not compile ISA support into a Linux kernel (on x86, at least - I hope it's not getting compiled into my SPARC kernels!)
Of course, if it actually worked 1% as well as advertised, it likely would be pretty good. My experience with USB has been pretty sucky.
Really? Too bad. I just got a nice G4 (at work) that's all USB for keyboard/mouse and that's been going fine. And I recently started using an USB Zip drive (I still use PS/2 for my keyboard, because I love using those huge old IBMs and there's no way anyone is making a USB version of that!)
However, I can understand why people will want to get rid of ISA. Oh, well, maybe someday...
you do not get 10% greater performance by buying a motherboard that has ni ISA slots (like those Asus KT boards). Because the fact is that even if they have no ISA slots, they still have a ISA bus built in the southbridge to support legacy stuff like the printer/parrallel port, the serial port/s & the PS2 mouse & keyboard ports.
For the most part, I would be willing to beleive you on this. However, is it not possible that modern chipsets use PCI or a seperate bus that does not interfere with the PCI bus to communicate with those devices? I'm really not sure; I suppose I'll go home and run lspci on my AMD-760 based sytem to see if there is an ISA bus lurking there.
In any case, ISA is a waste of space. I would take another PCI slot anyday.
They recommend a pretty well generic (though above average) Antec case, but this is s'pose to be a ultimate Linux box.
I would have to disagree with you here. I spent about 3 weeks researching what case to get, and decided on the Antec SX830 because of it's excellent overall design, room for plenty of fans, and well built drive tray. The slide out motherboard tray in the cases you recommend is all but useless to me, personally (other may, of course, disagree). Others by Supermicro, etc, were also discarded, partly on (lack of) certain features, and also because they were overly large.
In this particular case, I would have gone with a nice 2U rackmount case; after all, the Thunder K7 was built for rackmounting.
Your desire for massive amounts of integrated hardware, plus 6 PCI + 1 ISA, plus SMP, would end up, I'm pretty sure, with a motherboard too big to fit into most ATX cases.
but the Linksys cards (I forget the chip they use but it's not one of the ones you listed) have been solid
I've been using Netgear FA310TX cards (which are, apparently, the same thing as some Linksys cards - I'm kind of confused about the exact relation but that's what seems to be the case). Anyway, they are tulips, and I've never seen any problems like those mentioned running on a 100 Mbit network with plenty of traffic over the last 2+ years. I am highly skeptical of this supposed problem.
Intel EtherExpress isn't bad either, it's built into the motherboards of some of the machines around here and they seem to be working out pretty well.
But seriously, if you have a perfectly good Compaq case and power supply holding a 486 board and you find a good deal on a Gateway Socket 7 (or the other way around brand wise), it's really annoying that one's got the video jack where the other has the keyboard and mouse or vise versa
Yeah, OEMs do love to play those kinds of games. Too bad, but I guess they like being the single source (and being able to charge amounts which match that status for spare parts).
By the way, just what kind of old computer junk can you use? (if you weren't kidding)
I think this sums it up: link.
KAI is not compiler in the usual sense though, its a translator C->C++ (last time i checked anyway),
True. But I think about it as a a compiler, because it translates a complex language (C++) into a simple(r) language (C), just like gcc does (except it's C++ -> some form of assembler).
hence bound to be slow(compilation time wise).
Not really. Faster than gcc 3.0.1, anyway (I haven't done any timing, but I would estimate about 1.5x the speed).
Maybe it's time to replace the whole ISA-PCI-IDE started out as an 8-bit platform and got patched and kludged time and time again mess with something that anticipates that what now seem like big drives, big RAM sticks, and fast processors and video cards will soon be classed with 8088s and 64K ram chips.
:) You've got to pick, keep the crap, or dump it and everything related to it. But please, don't toss your old computers into a landfill. Give them to me!
Say what you like about IA-64, at least Intel does seem intested in killing off all that crap fairly soon. I don't see why you group PCI in with ISA and IDE, however. PCI is a "real" bus, and while it's not perfect, it's high-performance enough for most things, and it's reasonably well designed. I imagine that 3GIO is going to be, for the most part, an extension of PCI to higher clock speeds, rather than completly reworking it (I haven't been able to find any actual info about anywhere, so I guess I could be wrong). Also going out over the next few years are PS/2 [I hope I can find a PS/2 -> USB adapter; I love my keyboard and want to keep using it], floppy drives, serial ports [which is rather unfortunate, lots of old Unix boxen can or have to use the serial port to get an install going], and parellel IDE, replaced by Serial ATA. It's kind of kludgy internally, but I guess from the user's perspective it will be a lot nicer. Similiar to firewire drives, but >4x faster.
un-upgradeable except by pitching them into the landfill and buying a whole new system.
It's funny that you seem to dislike the bad old standards so much, yet you also don't want to throw any of it way. I'm confused.
With 3 or more new fast ATA/100 drives (limit 1 per cable) and cheap IDE controllers, you can run out of PCI bus bandwidth.
Potentially, anyway. I have a 7200 RPM ATA/66 drive, and usually only get about 23-26 Mbytes/sec when reading directly off the disk (no caching). Some fast IBM ATA/100 disks manage up to 37 Mbytes, which would saturate the bus if 4 of them were working non-stop. How realistic any of these sitautions are, is, of course, depends on the situation.
Of course, there's always 64-bit 66 Mhz PCI, 3GIO [someday], and firewire drives. USB 2.0 too, I guess.
I belive it when I see the numbers, although Borland got a reputation of isssuing quite fast compilers..
True enough, but what I worry about much more is whether it will be (roughly) C++ compliant or not. It would kind of suck if a vast majority of C++ apps couldn't be built with it.
So far I have heard mixed reports from people about ICC effiency in terms of code generated, speed of binaries, size of binaries etc (slashdot users who use ICC - please post your conclusions).
There are, in fact, several other C/C++ compilers for Linux besides those two, including KAI C++ and Portland Group C++ (I believe you mentioned PGCC in you post, actually). I'm quite certain that KAI, especially, is more widely used than Intel's compiler, because it runs on everything [Linux/x86 all the way up to Crays and S/390s], and has been around for some years.
I haven't used Intel's compiler, and I can't say much for C, but I will say this for C++: compared to gcc 2.95.2 and gcc 3.0.1, KAI C++ usually generates code that is faster (2x-3x) and smaller (about half). Since Intel recently acquired KAI, I imagine those properties (well, besidesthe portablity) have bled over to some extent into Intel's compiler. Portland Group C++, I've had nothing but trouble with.
Is MetroWerks a "real" compiler on Linux now? Last time I looked it was just a pretty IDE for gcc.
1. GCC compatible - you'll need it if you want to be used by open source users OR to allow developers to move their apps which used GCC to your compiler.
I don't think so; how many applications really use the GCC extensions? Not very many. All it has to be is ANSI C compatible, so it can use the normal glibc functions, and for the most part the problem is gone.
Hey, totally off topic, but your nick is based of the main character of Weis & Hickman's Deathgate Cycle, right? Just checking my memory, I'm not quite sure.
/home, /var, etc. One cool thing for this hypothetical mover software could do would be to parse /etc/fstab and handle that for you when it installs the new distro.
Uh... to help avoid getting modded as off-topic, yeah, LSB compliant systems would certainly help quite a bit. Fortunately, you know that on most Linux systems, and indeed on most Unix systems in general, you've got a
One nice idea I had a while back as far as copying/backing up the passwd/group files is using something like LDAP to handle it. I've had to do what you did a few times (upgrading the NIS server, for example), and it's a real PITA (especially when there's ~50 users and a bunch of groups, all of which are vital to the organizations functioning). Unfortunately, in the main, most software doesn't really deal with having passwords anyplace but in the usual spots. It really would be nice if there were a better seperation between system and "real user" accounts, but, alas...
I one time forgot about the mail spools when moving over. Thank god this was my home machine and not something which a bunch of people would would be out for blood for. Though I did feel pretty dumb (soon after I changed my procmail config to dump _everything_ into a file in my home directory).
There has been a lot of talk that PGP is the only unbreakable encryption method out there
Which is complete bullshit, for 2 reasons:
1) PGP, like virtually any other form of cryptography, is breakable. It may not be breakable in a reasonable amount of time, but it is certainly breakable. Do not use "unbreakable" unless something is, in fact, unbreakable (within whatever environmentt you're placing it).
2) Why would PGP be particularly strong over other protocols? It's not. Hell, it doesn't even include MACs in the messages, which brings up all kinds of problems.
maybe because the RSA has finally found a way to break the PGP encryption?
Umm... wait, what? Perhaps you meant the NSA???
Anyway, maybe it was because they realized that they didn't have any kind of proof that PRZ had exported the code, that if they pursued it with criminal charges, more likely than not the unconstitutional restrictions on crypto would have been declared as such and lifted, and PGP was already widely available outside the US so they weren't getting much out of it.
It's standard practice that once you know how to break someone's code, you don't ever let them know which guarantees that you can keep on reading all their transmissions.
Wait... huh? From what I can get out of this, it would imply that the NSA even if the NSA could break the algorithms in PGP, they would keep up bothering Phil forever, to give people the impression that PGP was still a big threat to them.
I think you might have meant ``faster than a deterministic Turing machine'', which I wouldn't argue with.
Holy crap I'm dumb. What's really embarressing is that I took a complexity theory class last semester that went over all of this stuff; you'd think I could remember something as simple as that for at least 6 months.
BTW, I don't think your post was being pedantic at all; my original post was quite plainly false, and I thank you for pointing out my error.
300 seems, uh... mildly excessive.
:)
JHU has one main one (jhu.edu), and most everyone has subdomains off of that. There is another one for the med school (jhmi.edu), and a few others for various organizations (jhuisi, jhuacm, etc).
However, invoking whois, I see that in fact jhu also owns jhu.{com,org}, jhuisi.{com,org}, jhmi.{com,org}, and various correct and incorrect spellings of johnshopkins.{com,org,edu} (though johnhopkins.net is for sale, it seems). Presumably they did this to prevent people from using them to make the university look bad, or whatever.
So I'm starting to guess that a lot of schools do this sort of thing, just like the various companies will squat on domains they have no real rights to (see the story on the Register about Reuters). But don't worry, even if they weren't doing this, I'm sure your college could find other ways to waste your tuition on frivoulous things. I know mine does.
That's what I don't trust--the hype and hubbub. Superpositional computation has tremendous theoretical possibilities, but superpositional computation in practice is... nowhere near useful. Ask me again in five years what I think and you might get a different answer. :)
OK, yeah, that's a pretty reasonable opinion to take at this point. OTOH, it does seem like at least some progress is being made; even though the ones made so far are tiny and pretty useless, the theories that they would be able to compute faster than a Turing machine model computer have been shown to be valid. But if someone can ever build one that actually does something a normal computer can't do... that is the question.
I've read some papers about getting quantum key exchange working; it's nasty. I don't know enough physics to understand about 60% of it, but I can get enough out of it to know that getting those quanta to do the right thing is a real PITA.
Overall, I would disagree with your statement. Fertilizer is pretty damn common, at least in rural areas where I grew up (if you own a farm, it's not too hard to convince someone to sell you a bunch of fertilizer). Knives are heavily regulated? What? Ever been to a steakhouse?
You cannot simple say banning==bad freedom==good unless your definition of good is anarchy.
True, but "No Constitutional Rights == bad". Banning encryption (or banning strong encryption, at least), would infringe upon my rights of free speech (at least according do, IIRC 9th circuit court of appeals, which held that software is speech, at least in some circumstances). Additionally, at least I believe that my right to free speech allows me to speak in encrypted form (how is that different from using a private slang to talk to a friend?). I don't know if that has been tested in any court, however.
Perhaps books about number theory, information theory, and network security should be banned as well? All can be used as tools to build encryption, just like if I wanted to buy detonators for a bomb.
Encryption is not a thing, it is an idea. Just like terrorism, actually. You think if every terrorist in the world was killed, more wouldn't start showing up? Get rid of the PLO, Hezbollah, and Islamic Jihad, and Palestinians will start being happy about having their houses bulldozed by Israel?
ROT 13. Plus DMCA. Plus Attack Lawyers.
Someone actually suggested this in a class I'm in (High Assurance Systems), for securing a system. You don't go to JHU, do you?
I don't trust QC of either sort because it depends on so much knowledge of physics and technical savvy that, were it to be fielded today, it would be hideously insecure by virtue of its implementation being so difficult to get right.
Quantum Crypto, I agree, is way to new to be trusted (not like anyone has figured out how to implement it in a practical manner yet anyway). But quantum computation: what is there to trust or not trust? It's just like an Athlon, only different; either it works, or it doesn't. Either it factors an RSA key, or it doesn't. It's not like it spits out an answer that you can't check easily.
I don't trust ECC, even though the Taniyama-Shimura Conjecture has been proven, because all of the good elliptic curves have been patented by Certicom and the remainder are either untrustworthy or too slow for practical use.
Well, that's a good reason not to use it, but not a good reason not to trust it. I mean, you (presumably) trusted some form of public key crypto prior to 1997, and all forms were covered by the RSA and Hellman-Merkle patents up until then (or at least RSADSI claimed they were covered, and was willing to sue anyone who disagreed). Anyway, IIRC Certicom only (or mostly) patented stuff using prime fields, leaving GF(2^N) wide open. Sure, you need a larger field for security, decreasing the benefits, but it's still faster and smaller than RSA or Elgamal.
It amazes me to see all these people on Ask Slashdot (I know these kinds of questions have been asked at least twice in the last year) assuming that new == good. So overall I would say you make some good points.
Just assemble and ship the product. If you don't know how to use it, don't buy it from us. A side benefit of this idea would be that "hard-core" computer geeks can stop wandering from site to site to build their computers. Since there's little over-head to cover, prices will stay cheap.
I'm skeptical that they could do anything, that, for example, mwave can't. It's a rare piece of (new) hardware that I don't buy from them. You can pay a little extra to get a full box assembled before they ship, but it's better to get it as is, IMHO. And you can, of course, order anything else you might want - full flexibility. Split your computer purchase over 2 orders (which is, in fact, what I'm doing right now - the video card, DVD-ROM, and sound card won't be ordered until next month). Get stuff that Dell would never sell you: you want 3 PCI Voodoo 5 5500s in your new box? - go for it [actually I think their stock of Voodoo5s ran out a while back, but you get the idea]. And get it damn cheap, at least compared to Dell, et. al.
Their policies are pretty much what you describe; their return procedures are fine [if it's broken, they'll replace it], they sell quality stuff, and there is no real support. But if you have questions, they will answer them, to the best of their ablities, in a quick and courteous manner. And they ship fast (like, same day, in most cases - next day, if you want them to assemble stuff for you).
Your hypothetical company could offer more products, etc, but that's about the only benefit I would see to it. And on the down side: QA problems [is there a warranty? If so, how do your make sure that your wide selection of stuff can't be mixed together to form something that doesn't work due to conflicts.], slower shipments than parts places (you have to built each one by hand, because there are a big selection of different parts - you can't pre-build common selections very well), and a new start-up is going to have a hard time getting business of that sort, because 90% of the people I know who are building a PC order everything from mwave.
Currently sitting in a room with 9 mwave machines...
and bars that take cash only are a relic of days gone by
:)
I would say you simply don't like to hang out in hole in the wall dives much. Most of the "best" bars in the Baltimore area, for example, will only take cash (the stupid college-bar places will take CCs, though). I have seen vending machines that will take CCs, but never a pool table. Nor will most jukeboxes that I see. And the local liquor store: cash only.
Also, there are other places that strongly prefer cash, for example, tattoo places. They'll take plastic or check, but they like cash (who doesn't want to make $100/hour tax free?).
Your point about "Stripper Dollars" is interesting, but I think I'll stick to 5s for now.
I got this email on Friday:
"Monday 9/24, noon, at the Mattin Center: U.S. Representative Constance Morella (8th District of Maryland) will talk regarding Information Security and Privacy."
The Mattin Center is the new arts building on the campus of Johns Hopkins University. I'll be showing up and a hope others will as well.
If you want directions or more info please respond to this post.
IANAL, but aren't there already laws making it a felony to use the telephone or postal system, or to meet face to face to discuss illegal activies ("Conspiracy to...").
Indeed there are. Not to mention laws that allow a judge to ban you from talking to some set of other people (with a felony penalty if you are later caught with them). Kind of like a restraining order, but the other person doesn't want you to keep away from them.
Also, in a few cities, a police officer can prohibit you from talking to someone else, just for the hell of it (for example: you're talking with your friend in a coffee shop, and a cop, can, legally, go up and say "you - stop talking with that guy" and haul your ass of to jail if you refuse). These may have been deemed unconstitutional by now, but I'm not sure. They were nominally for anti-gang stuff, but they were apperantly mis-applied (the police were going on the theory: black -> in gang).
(Or you use disposable phones, face to face meetings, mail drops and personals ads like they actually do...)
Don't worry, all those will be banned shortly.
I wish I could say this entirely as a joke.
Just ditch this piece of shit Tyan and buy something that works properly... MBs only cost like $100
Ummm... I hope you're joking. Tyan Thunder == Dual Athlon == Very expensive == Only one on market right now.
I think you misunderstood. I just meant that Java itself isn't slow. It's the class libraries - like the String class - that are the cause of most of the slowdown in your average application. If you are writing a large scale app, implementing a fast string class is completely trivial.
No, I understand what was meant. However, the libraries are fairly intrinsic to the language, especially those in the basic package. If the libraries are slow, or otherwise deficient, the language itself suffers (IMO), because either you take a performance hit (in terms of real application performance), or you have to reimplement basic functionality. Of course, I can reimplement a string type - but why should I have to? Just because Java can do things not involving it's standard library relatively fast doesn't mean a whole lot, if it's libraries are really slow, and thus making virtually every application actually written in Java slower than it needs to be.
For that matter, why doesn't someone just speed up the damn String class? Has Sun not heard of code profiling?
You haven't played at a dual 800 G4 it seems. It's a bit on the expencive side, but damn fast.
I just got one for work. It's nice. And damn sure, no problems with speed.
And even if you could build your own computer without Windows, you still can't build a laptop.
Actually, I believe one manufactuer (Acer?) is selling laptops without disks or memory. So no Windows, either. I think I saw it on arstechnica, or possibly anandtech. Or then again, maybe toms. One of the 3, anyway.