If we actually return to the moon, might a space elevator be more practical there? Could we do that now?
There's a whole wikipedia article on the topic, but in summary, a lunar elevator would be off the shelf. Not off the shelf at your local home depot, but more like Aircraft Spruce and Specialty. Not kidding, I checked their website and they sell kevlar49 spools at about ten cents per foot for 7100 denier.
It would be fun to try on a "smaller" asteroid. Then literally hardware store products would be good enough plus the operational training would be priceless.
How about we get to the point where we can build a bridge over a valley somewhere with carbon nanotubes first. Even that is a LONG ways out. Not any of our lifetimes. And that bridge is about 10000X easier to build than a space elevator.
Not really. CF is basically "really weak nanotube". CF doesn't burn too well in a vacuum, and there are not many vandals in space, or at least you can put a guard shack at the base and be done with that issue.
There are "many" CF bridges, at least in the USA. Mostly corrosion proof, you tend to find them up north in "road salt country". I can imagine, within my lifetime, we will no longer use steel rebar in concrete.
The problem with purely CF bridges, is to stop vandals with no more than a hunting knife from collapsing the bridge, or stop a simple vehicle fire from incinerating the bridge, you have to encapsulate the thing in concrete, which of course in freeze/thaw cycles delaminates from the CF, or sticks to the surface of the CF making the CF delaminate, etc.
People get real confused about CF and give it mystical properties it does not have. It has spectacular, record breaking tensile strength, yes. Surface hardness? No, not like diamond, in fact its about as hard as charcoal WRT to dulling cutting blades. Abrasion resistance? No, not like stainless steel, more like the plastic resin its bonded with. High temperature strength? No, not like tool steel, more like the plastic its bonded with. Bullet proof? Yeah about as bullet proof as an equal amount of plastic. It's just about unstretchable, but otherwise you may as well think of it like plastic.
Carbon based bridges are a huge mess, not because carbon based structures suck, but because vandals suck, and gasoline automobiles suck.
ooh cool. too bad the dev kit is $300 and the whole thing requires licenses. I suppose the license and its probably hideous fees is what prevents ladyada etc from offering a $20 kit.
Apple makes much more than just software. iPhones, iPads and Mac's are not software. I truly hate blanket statements like this. All the happened was the market cap of an information technology firm exceeded the market cap of an oil company. These numbers reflect the value of the companies not the value of the underlying commodities.
Apple does not make any of those devices. A company in China makes them. Apple supposedly designs them, and definitely Apple markets and distributes them.
Exxon does not "make" oil either. They're pretty hot stuff at finding it, and coordinating the work of subcontractors to pump it out, assuming the country owning the land allows them to work instead of using their own nationalized company (petrobras, etc). In a world of declining oil production, I'm not sure how useful Exxon is. Kind of like a middleman. Does transocean really need exxon anymore if all the worlds oil is already found, mostly remains in far away lands where exxon isn't allowed to work, and TO is already pumping it?
Both of those already exist. HIV in particular does both.
Yes that's exactly the point. Just like its no problem for healthy normal humans if 0.001% of staph bacteria are antibiotic resistant... The problem is when you kill ALL the other staph bacteria so the population becomes 100% antibiotic resistant. Then you got a big problem, especially if that resistance cross pollinates into other bacteria.
So if you wipe almost all viruses off the face of the earth except HIV, then pretty much all viruses, are going to take on the characteristics of HIV. Including the common cold. Then we're pretty much screwed.
Just sayin...it thinks all your braincells are viruses and turns you into a ZOMBIE.
Naah your blood/brain barrier is remarkably efficient. Thats the good news. The bad news is this probably won't cure viral meningitis unless you inject it directly into the spinal fluid.
The really bad situation is this kills virus infected cells. Well.. research some autoimmune based digestive system diseases: the simplified version is your digestive system eats a protein (gluten, soy, etc) that your immune system dislikes, so your immune system exerts great effort on destroying your digestive system. Luckily it usually doesn't work terribly well, and results at worst in long term malnutrition or starvation. My son is in that situation. As long as he does not eat the protein that agitates his immune system he's all good.
Anyway, imagine instead of a weakened tired out immune system, this magic drug being administered, and it finds your intestinal walls have absorbed a cow / chicken / pig / tobacco virus... And this magic drug being quite strong... on a sy fy network movie special, this drug could liquify your entire digestive system if you ate a virus infested plant or animal. That could be a problem, yeah.
Our cells carry loads of genetic material picked up pretty much everywhere. They just go "hey look, some code, I wonder what will happen if I execute it".
There are now at least two valid interpretations of "Humans are DNA based windows users"
This could potentially save us Lots of money not having to continuously reprint on paper these schedules.
IF and only if the e-paper is cheaper than 260 sheets of regular paper. I'm guessing that's not gonna turn out well.
I'm struggling, struggling hard, to think of a use for this. Something gets rewritten dozens of times in its useful life and no one cares about the cost...
Democratizing app building, empowering kids, women, and underrepresented groups
I said this when it came out and I'll say it again - where is the real demand for this from these people the author is quoting?
Supposedly that demand is the result of anti-kid / anti-woman in other dev tools. Ah that must be emacs with its "kitchen sink" comparisons, you know, keep em barefoot, pregnant and in the KITCHEN. Of course then there is vi. My guess is vi is anti-child, because you hit escape about every 5th keystroke, and everyone knows from horror movies that some mass murder ESCAPEs and kills all the teenagers in the movie. As for perl, well you got the camel book, and camels are from the middle east, and they're not known for their feminist outlook on life.
Inductive charging will be a valuable asset when a room, a house, or even a city can be blanketed with the required EM field,
Grandiose. I'd be happy with a magic box under my work desk and bedroom dresser that turn the entire horizontal surface into a magic battery charger. Can't lose the charger if its bolted underneath my desk. Can't fumble for cables if there are no cables.
Interoperability would be nice.. In ye olden days before microusb or whatever its called, every phone had a special connector and pinout to chain the users to the expensive aftermarket mfgrs. Theoretically an interoperable system would mean my friends would not need a magic cable for a charge, all they need to do is drop their device on my table.
Could some EE closer to the inductive charge community comment on:
1) Are the chargers "smart" like if I drop my wedding ring on the charger does it heat up/melt or does the charger recognize the inductance / current draw is way outta whack and shut off? If it shuts off does an indicator of some type turn on, or does it just not charge?
2) Frequency of operation? I would really hate to hear anything below, say, 30 KHz. Even at my age, just years ago when CRTs still roamed the earth, I found horiz sync whine to be astoundingly annoying. I would really hate to hear a 60 hz inductor, any tropical fish owner / diaphragm air pump owner knows the annoying drone of 60 hz + harmonics.
3) Who can sell me an inductive receiver kit to power other stuff? I'm not talking about bolt and go, but ladyada / dangerousprototypes sort of places and products? Who makes this stuff, anyway? At a superficial glance the usual suspects in the analog power community don't seem to offer any specialized ICs for the task... unless the RX has no 2-way comm with the tx and literally is just any ole coil feeding a bridge rect and a switcher.
4) I'm sadly picturing some kind of hideous DRM where the expensive charger and expensive device need to negotiate a RSA key across bluetooth to light up the charger... Please tell me it isn't so? A generation of interoperability would be awesome.
Because PoE gives you about 25W (up to 50W if you don't care about specs, standards and safety) at, usually, 47V. Converting that down to 19V probably takes quite a bit of efficiency so you'll be lucky to get 10W.
It would take a miracle to get less than 10W because thats what you'd get out of a linear regulator, the least efficient form of commercially available regulators.
lots/most AC switchers start by rectifying the incoming AC... the question is, can you get it to start off only 47 volts. A working switcher would deliver near 25W out with 25W in. You can't dump 15 watts in those tiny little things, they'd literally melt.
I just hope it becomes a standard because PoE was nowhere to be found.
The mark up for PoE switches was/is spectacular, because the marketing guys told them to price that feature at "just below the cost of hiring a union electrician to run a dedicated AC line next to the wall plug". Which, it turns out, is a heck of a lot of money.
The marketing people forgot about extension cords. So, most of the real world uses extension cords instead. Whoops. PoE was a cool idea, but too sabotaged to ever make it.
I find it very hard to believe that this feedback loop exists in real life to any significant degree. If it really was true, the professional sports athletes would prefer walking and biking over driving their cars, and the sport stars seem to be preferring their luxury sports cars today.
IBM's model must be missing one or more important variables to why people choose cars over walking.
They're talking about old people. Old people who can't walk, don't. Of course it might take 70 years for changes made today to have an effect on the 80-year olds of the future.
Ever try that in SimCity? Unless you have some massive money built up it usually fails. Turns out rail is not all that cheap.
That's the unrealistic part about simcity. In the real world, rail pays the govt, not the other way around. That being why the rail infrastructure is in such poor shape in this country... If they invested, the govt would tax the tar out of them, which they can't afford.
"Computers need to be very fast to emulate something slow enough to be accurate."
Yeah, I don't see anything wrong with that theory. It's kind of like how I always drive my Ferrari at 5 miles an hour because it's more accurate than riding a bicycle.
A much better/. car analogy is the "best" most "modern" way to emulate a ford model T, oddly enough involves a very elaborate and expensive CNC machine shop, because you can clone replica parts. If the engine block casting had a 0.0001 warp, the CNC mill can include that "mistake" in the new part. If there was a slag inclusion in the piston, the CNC lathe can take a little divot out to screw up the balance to maintain "prototype" vibration levels. If the prototype way to build something was dumb or dangerous, the robot will do it exactly.
Attempting to rely on today's multi-core processors leads to all kinds of timing problems.
No, not at all, unless you're practically trying intentionally to screw it up. I've done a lot of work on emulators, admittedly mostly for DEC and IBM big iron, and the other extreme of emulating embedded microcontrollers, rather than video games. But video games are pretty simple compared to a large virtual mainframe installation.
The way you use multi-cores is the same way you use multi-cpus. One core/cpu does all the "thinking" and the other cores/cpus do ALL the behind the scenes heavy lifting. For example, the cpu emulator process/thread/hardware cpu does NOT handle local file access, it does NOT handle the real machine's "audio interface of the month", it does NOT handle whatever video-graphical-console interface the real machine provides, it does NOT interact with the real machines window manager, it does NOT process the real world machine's network packets. If your "CPU emulator" core/proc is also responding to pings from the real LAN, you're doin it wrong (or at least you need to buy more cores/procs, etc).
If the emulated machine had multiple processors or some peripheral processor, then you can run the FPU or keyboard controller or graphics proc or whatever on a separate real core/processor. Each call involves a tick counter, and each call ends with a loop waiting for the tick counter to hit the correct value before returning "done" status to the main.
Even the main machine needs a tick counter, and some sleepy calls. Crude systems do some arithmetic on the RTC to see how many ticks have been processed vs should have been processed and then sleep till then, maybe every X ticks. Advanced systems have a sleepy routine each and every tick, and a synthetic PID controller that adjusts the sleepy routine rather often. The fanciest systems which are usually too slow for videogames or realtime use "real time" for each instruction, including modeling of the current position of the disk drive arm and the suction level of the tape drive pumps and such.
Someone who's challenged by the concept of atomic transactions is going to be confused by them in emulators, just like they'll be confused by them in database transactions or any other app. Same deal with semaphores, fixing deadlocks, etc. There's nothing "really tricky" about programming emulators vs anything else.
Is Coda still developed? It seemed dead to me last time I went looking into the alternatives (some years ago, I'll admit).
Its dead; although no one develops it, supposedly it works. There's a mailing list with very light traffic, meaning either it never breaks or no one uses it, hard to say.
I've heard it compared to XFS; Its very non-mainstream. Its users like it because of its special abilities. Sometimes it blows up very spectacularly and no one can figure out why. I have nothing against those characteristics; in my youth I dated women just like that. Just know what you're getting into.
I just looked at AFS on Wikipedia and it looks very interesting.
What implementation of AFS do you use, server side and client side?
Do you have any books or documentation online you could recommend?
Sorry for responding late but thats how it goes. The plain ole off the shelf Debian packages. Debian infrastructure uses LDAP internally, they eat their own dogfood, if there is one piece of infrastructure I'd trust it would be the Debian LDAP packages...
The openafs guys make installable packages for mac and windows clients. Really boring, install and it goes. I am told its a stereotypical windows/mac experience in that either it "just works" or it is simply beyond untroubleshootable. Therefore since its known the clients work great on a happy network, a problem with the clients would indicate some kind of server problems. So the best way to troubleshoot the server is a linux (debian) client. Get the linux clients working, then get the mac clients working, then get the windoze clients working.
Anyway for the new guy, I think the best place to start is getting LDAP working:
I kid you not, look at the web page, match it to your screen, type what they say, it just works.
I have no connection w/ spinlock other than respect for their good work.
There is some circularity; as I recall, after setting up AFS you obviously get to go back to LDAP and change your homedirs to "/afs/something.com/users/you". When you set up LDAP, you'll change your PAM to use LDAP "semi-secure" passwords; then when you set up Kerberos you'll change PAM to use Kerberos passwords instead; Its circular-ish, not like set up LDAP and never touch it again ever. Also there are optional/bonus things like entering DNS SRV records so clients become "mostly self configuring" which is pretty cool.
One final note, is due to replication, and presumably a decent backup config, this is literally something you do one time and one time only. So you have kerberos "admins" like me who set up a long time ago and frankly can't provide much help on new installs, and the other side of the ecosystem is total noobs who just tried setting it up last week and therefore ALSO can't provide much (useful) advice. Another example is the last time I added a user to LDAP at home was after my youngest daughter was born; that is getting to be a long time ago; so asking me about adding LDAP users is like asking me about the fine details of anything I did during the early Bush years; "I donno". The moral of the story is go slow, think about it, and assume you'll never be able to change this again. For example, its kind of a PITA to change your Kerberos domain name after you set it up, so get it right the first time, you only have to get it right once.
AFS, for a start, is massively too complex for most home users or small business users,
Maybe if you set it up by reverse engineering the source code or watching protocol analyzer streams.
There's enough recipes out there, that if you can bake a cake, you can set up LDAP/Kerberos/AFS. In fact there are recipes that are so detailed they show screenshots and exactly what to type at each stage. Its literally easier than setting up a webserver (I've certainly done both enough times...). You know how you open a quality mouse hardware, and there's a printed poster with a dozen steps of how to install the mouse? Yeah thats the kind of overkill I'm talking about.
If you can't google, you're screwed wrt to figuring out how to install AFS; but lets face it, if you can't google, you're pretty much screwed wrt to life in general, so its not much of an excuse.
Also I suppose it depends on the distro. The Debian packages are up to snuff, thats for sure. I acknowledge it could be a PITA to compile on gentoo, donno.
there's no telling what might be crawling around in an old system
Which is why a lot of charities wipe by default. What's their liability if you install windows, a windows keylogger, donate to them, and now you identify theft someone?
I personally switched because the alternatives finally worked better, did more of what I wanted, better cross platform support... Its the end of an era where the best LAN protocol for a new deployment is SMB...
I acknowledge that switching is easier when you're only talking about a couple dozen boxes instead of 25K boxes. I suppose they'll always be legacy environments. That does not mean I'm actively interested in new advances in token ring cabling, or IPX/SPX, or decnet.
Slashdot Mirror
If we actually return to the moon, might a space elevator be more practical there? Could we do that now?
There's a whole wikipedia article on the topic, but in summary, a lunar elevator would be off the shelf. Not off the shelf at your local home depot, but more like Aircraft Spruce and Specialty. Not kidding, I checked their website and they sell kevlar49 spools at about ten cents per foot for 7100 denier.
It would be fun to try on a "smaller" asteroid. Then literally hardware store products would be good enough plus the operational training would be priceless.
"if we could build one"
How about we get to the point where we can build a bridge over a valley somewhere with carbon nanotubes first. Even that is a LONG ways out. Not any of our lifetimes. And that bridge is about 10000X easier to build than a space elevator.
Not really. CF is basically "really weak nanotube". CF doesn't burn too well in a vacuum, and there are not many vandals in space, or at least you can put a guard shack at the base and be done with that issue.
There are "many" CF bridges, at least in the USA. Mostly corrosion proof, you tend to find them up north in "road salt country". I can imagine, within my lifetime, we will no longer use steel rebar in concrete.
The problem with purely CF bridges, is to stop vandals with no more than a hunting knife from collapsing the bridge, or stop a simple vehicle fire from incinerating the bridge, you have to encapsulate the thing in concrete, which of course in freeze/thaw cycles delaminates from the CF, or sticks to the surface of the CF making the CF delaminate, etc.
People get real confused about CF and give it mystical properties it does not have. It has spectacular, record breaking tensile strength, yes. Surface hardness? No, not like diamond, in fact its about as hard as charcoal WRT to dulling cutting blades. Abrasion resistance? No, not like stainless steel, more like the plastic resin its bonded with. High temperature strength? No, not like tool steel, more like the plastic its bonded with. Bullet proof? Yeah about as bullet proof as an equal amount of plastic. It's just about unstretchable, but otherwise you may as well think of it like plastic.
Carbon based bridges are a huge mess, not because carbon based structures suck, but because vandals suck, and gasoline automobiles suck.
ooh cool. too bad the dev kit is $300 and the whole thing requires licenses. I suppose the license and its probably hideous fees is what prevents ladyada etc from offering a $20 kit.
Ariane is not a suborbital vehicle, so not quite understanding why it would be relevant.
Oh they've certainly made plenty of suborbital Ariane launches, plenty. Not intentionally, of course, but...
Seriously though, anything orbital inherently has an absolutely whopping suborbital capability.
Apple makes much more than just software. iPhones, iPads and Mac's are not software. I truly hate blanket statements like this. All the happened was the market cap of an information technology firm exceeded the market cap of an oil company. These numbers reflect the value of the companies not the value of the underlying commodities.
Apple does not make any of those devices. A company in China makes them. Apple supposedly designs them, and definitely Apple markets and distributes them.
Exxon does not "make" oil either. They're pretty hot stuff at finding it, and coordinating the work of subcontractors to pump it out, assuming the country owning the land allows them to work instead of using their own nationalized company (petrobras, etc). In a world of declining oil production, I'm not sure how useful Exxon is. Kind of like a middleman. Does transocean really need exxon anymore if all the worlds oil is already found, mostly remains in far away lands where exxon isn't allowed to work, and TO is already pumping it?
Both of those already exist. HIV in particular does both.
Yes that's exactly the point. Just like its no problem for healthy normal humans if 0.001% of staph bacteria are antibiotic resistant... The problem is when you kill ALL the other staph bacteria so the population becomes 100% antibiotic resistant. Then you got a big problem, especially if that resistance cross pollinates into other bacteria.
So if you wipe almost all viruses off the face of the earth except HIV, then pretty much all viruses, are going to take on the characteristics of HIV. Including the common cold. Then we're pretty much screwed.
Just sayin...it thinks all your braincells are viruses and turns you into a ZOMBIE.
Naah your blood/brain barrier is remarkably efficient. Thats the good news. The bad news is this probably won't cure viral meningitis unless you inject it directly into the spinal fluid.
The really bad situation is this kills virus infected cells. Well.. research some autoimmune based digestive system diseases: the simplified version is your digestive system eats a protein (gluten, soy, etc) that your immune system dislikes, so your immune system exerts great effort on destroying your digestive system. Luckily it usually doesn't work terribly well, and results at worst in long term malnutrition or starvation. My son is in that situation. As long as he does not eat the protein that agitates his immune system he's all good.
Anyway, imagine instead of a weakened tired out immune system, this magic drug being administered, and it finds your intestinal walls have absorbed a cow / chicken / pig / tobacco virus... And this magic drug being quite strong... on a sy fy network movie special, this drug could liquify your entire digestive system if you ate a virus infested plant or animal. That could be a problem, yeah.
Our cells carry loads of genetic material picked up pretty much everywhere. They just go "hey look, some code, I wonder what will happen if I execute it".
There are now at least two valid interpretations of "Humans are DNA based windows users"
A real /.er would figure out the limit. I'm guessing its something super lame for PR purposes, like "limit one per address".
This could potentially save us Lots of money not having to continuously reprint on paper these schedules.
IF and only if the e-paper is cheaper than 260 sheets of regular paper. I'm guessing that's not gonna turn out well.
I'm struggling, struggling hard, to think of a use for this. Something gets rewritten dozens of times in its useful life and no one cares about the cost...
I said this when it came out and I'll say it again - where is the real demand for this from these people the author is quoting?
Supposedly that demand is the result of anti-kid / anti-woman in other dev tools. Ah that must be emacs with its "kitchen sink" comparisons, you know, keep em barefoot, pregnant and in the KITCHEN. Of course then there is vi. My guess is vi is anti-child, because you hit escape about every 5th keystroke, and everyone knows from horror movies that some mass murder ESCAPEs and kills all the teenagers in the movie. As for perl, well you got the camel book, and camels are from the middle east, and they're not known for their feminist outlook on life.
Inductive charging will be a valuable asset when a room, a house, or even a city can be blanketed with the required EM field,
Grandiose. I'd be happy with a magic box under my work desk and bedroom dresser that turn the entire horizontal surface into a magic battery charger. Can't lose the charger if its bolted underneath my desk. Can't fumble for cables if there are no cables.
Interoperability would be nice.. In ye olden days before microusb or whatever its called, every phone had a special connector and pinout to chain the users to the expensive aftermarket mfgrs. Theoretically an interoperable system would mean my friends would not need a magic cable for a charge, all they need to do is drop their device on my table.
Could some EE closer to the inductive charge community comment on:
1) Are the chargers "smart" like if I drop my wedding ring on the charger does it heat up/melt or does the charger recognize the inductance / current draw is way outta whack and shut off? If it shuts off does an indicator of some type turn on, or does it just not charge?
2) Frequency of operation? I would really hate to hear anything below, say, 30 KHz. Even at my age, just years ago when CRTs still roamed the earth, I found horiz sync whine to be astoundingly annoying. I would really hate to hear a 60 hz inductor, any tropical fish owner / diaphragm air pump owner knows the annoying drone of 60 hz + harmonics.
3) Who can sell me an inductive receiver kit to power other stuff? I'm not talking about bolt and go, but ladyada / dangerousprototypes sort of places and products? Who makes this stuff, anyway? At a superficial glance the usual suspects in the analog power community don't seem to offer any specialized ICs for the task... unless the RX has no 2-way comm with the tx and literally is just any ole coil feeding a bridge rect and a switcher.
4) I'm sadly picturing some kind of hideous DRM where the expensive charger and expensive device need to negotiate a RSA key across bluetooth to light up the charger... Please tell me it isn't so? A generation of interoperability would be awesome.
Because PoE gives you about 25W (up to 50W if you don't care about specs, standards and safety) at, usually, 47V. Converting that down to 19V probably takes quite a bit of efficiency so you'll be lucky to get 10W.
It would take a miracle to get less than 10W because thats what you'd get out of a linear regulator, the least efficient form of commercially available regulators.
lots/most AC switchers start by rectifying the incoming AC ... the question is, can you get it to start off only 47 volts. A working switcher would deliver near 25W out with 25W in. You can't dump 15 watts in those tiny little things, they'd literally melt.
I just hope it becomes a standard because PoE was nowhere to be found.
The mark up for PoE switches was/is spectacular, because the marketing guys told them to price that feature at "just below the cost of hiring a union electrician to run a dedicated AC line next to the wall plug". Which, it turns out, is a heck of a lot of money.
The marketing people forgot about extension cords. So, most of the real world uses extension cords instead. Whoops. PoE was a cool idea, but too sabotaged to ever make it.
Ever try that in SimCity? Unless you have some massive money built up it usually fails. Turns out rail is not all that cheap.
That's the unrealistic part about simcity. In the real world, rail pays the govt, not the other way around. That being why the rail infrastructure is in such poor shape in this country... If they invested, the govt would tax the tar out of them, which they can't afford.
"Computers need to be very fast to emulate something slow enough to be accurate."
Yeah, I don't see anything wrong with that theory. It's kind of like how I always drive my Ferrari at 5 miles an hour because it's more accurate than riding a bicycle.
A much better /. car analogy is the "best" most "modern" way to emulate a ford model T, oddly enough involves a very elaborate and expensive CNC machine shop, because you can clone replica parts. If the engine block casting had a 0.0001 warp, the CNC mill can include that "mistake" in the new part. If there was a slag inclusion in the piston, the CNC lathe can take a little divot out to screw up the balance to maintain "prototype" vibration levels. If the prototype way to build something was dumb or dangerous, the robot will do it exactly.
Attempting to rely on today's multi-core processors leads to all kinds of timing problems.
No, not at all, unless you're practically trying intentionally to screw it up. I've done a lot of work on emulators, admittedly mostly for DEC and IBM big iron, and the other extreme of emulating embedded microcontrollers, rather than video games. But video games are pretty simple compared to a large virtual mainframe installation.
The way you use multi-cores is the same way you use multi-cpus. One core/cpu does all the "thinking" and the other cores/cpus do ALL the behind the scenes heavy lifting. For example, the cpu emulator process/thread/hardware cpu does NOT handle local file access, it does NOT handle the real machine's "audio interface of the month", it does NOT handle whatever video-graphical-console interface the real machine provides, it does NOT interact with the real machines window manager, it does NOT process the real world machine's network packets. If your "CPU emulator" core/proc is also responding to pings from the real LAN, you're doin it wrong (or at least you need to buy more cores/procs, etc).
If the emulated machine had multiple processors or some peripheral processor, then you can run the FPU or keyboard controller or graphics proc or whatever on a separate real core/processor. Each call involves a tick counter, and each call ends with a loop waiting for the tick counter to hit the correct value before returning "done" status to the main.
Even the main machine needs a tick counter, and some sleepy calls. Crude systems do some arithmetic on the RTC to see how many ticks have been processed vs should have been processed and then sleep till then, maybe every X ticks. Advanced systems have a sleepy routine each and every tick, and a synthetic PID controller that adjusts the sleepy routine rather often. The fanciest systems which are usually too slow for videogames or realtime use "real time" for each instruction, including modeling of the current position of the disk drive arm and the suction level of the tape drive pumps and such.
Someone who's challenged by the concept of atomic transactions is going to be confused by them in emulators, just like they'll be confused by them in database transactions or any other app. Same deal with semaphores, fixing deadlocks, etc. There's nothing "really tricky" about programming emulators vs anything else.
Is Coda still developed? It seemed dead to me last time I went looking into the alternatives (some years ago, I'll admit).
Its dead; although no one develops it, supposedly it works. There's a mailing list with very light traffic, meaning either it never breaks or no one uses it, hard to say.
I've heard it compared to XFS; Its very non-mainstream. Its users like it because of its special abilities. Sometimes it blows up very spectacularly and no one can figure out why. I have nothing against those characteristics; in my youth I dated women just like that. Just know what you're getting into.
FTP? Webdav? For the leaching, mythtv-web assuming thats what you're doing, or simple http server?
I just looked at AFS on Wikipedia and it looks very interesting.
What implementation of AFS do you use, server side and client side?
Do you have any books or documentation online you could recommend?
Sorry for responding late but thats how it goes. The plain ole off the shelf Debian packages. Debian infrastructure uses LDAP internally, they eat their own dogfood, if there is one piece of infrastructure I'd trust it would be the Debian LDAP packages...
The openafs guys make installable packages for mac and windows clients. Really boring, install and it goes. I am told its a stereotypical windows/mac experience in that either it "just works" or it is simply beyond untroubleshootable. Therefore since its known the clients work great on a happy network, a problem with the clients would indicate some kind of server problems. So the best way to troubleshoot the server is a linux (debian) client. Get the linux clients working, then get the mac clients working, then get the windoze clients working.
Anyway for the new guy, I think the best place to start is getting LDAP working:
http://techpubs.spinlocksolutions.com/dklar/ldap.html
Then convert from ldap stored passwords to kerberos passwords as seen at:
http://techpubs.spinlocksolutions.com/dklar/kerberos.html
That gives you the required infrastructure to run afs:
http://techpubs.spinlocksolutions.com/dklar/afs.html
I kid you not, look at the web page, match it to your screen, type what they say, it just works.
I have no connection w/ spinlock other than respect for their good work.
There is some circularity; as I recall, after setting up AFS you obviously get to go back to LDAP and change your homedirs to "/afs/something.com/users/you".
When you set up LDAP, you'll change your PAM to use LDAP "semi-secure" passwords; then when you set up Kerberos you'll change PAM to use Kerberos passwords instead; Its circular-ish, not like set up LDAP and never touch it again ever.
Also there are optional/bonus things like entering DNS SRV records so clients become "mostly self configuring" which is pretty cool.
One final note, is due to replication, and presumably a decent backup config, this is literally something you do one time and one time only. So you have kerberos "admins" like me who set up a long time ago and frankly can't provide much help on new installs, and the other side of the ecosystem is total noobs who just tried setting it up last week and therefore ALSO can't provide much (useful) advice. Another example is the last time I added a user to LDAP at home was after my youngest daughter was born; that is getting to be a long time ago; so asking me about adding LDAP users is like asking me about the fine details of anything I did during the early Bush years; "I donno". The moral of the story is go slow, think about it, and assume you'll never be able to change this again. For example, its kind of a PITA to change your Kerberos domain name after you set it up, so get it right the first time, you only have to get it right once.
AFS, for a start, is massively too complex for most home users or small business users,
Maybe if you set it up by reverse engineering the source code or watching protocol analyzer streams.
There's enough recipes out there, that if you can bake a cake, you can set up LDAP/Kerberos/AFS. In fact there are recipes that are so detailed they show screenshots and exactly what to type at each stage. Its literally easier than setting up a webserver (I've certainly done both enough times...). You know how you open a quality mouse hardware, and there's a printed poster with a dozen steps of how to install the mouse? Yeah thats the kind of overkill I'm talking about.
If you can't google, you're screwed wrt to figuring out how to install AFS; but lets face it, if you can't google, you're pretty much screwed wrt to life in general, so its not much of an excuse.
Also I suppose it depends on the distro. The Debian packages are up to snuff, thats for sure. I acknowledge it could be a PITA to compile on gentoo, donno.
there's no telling what might be crawling around in an old system
Which is why a lot of charities wipe by default. What's their liability if you install windows, a windows keylogger, donate to them, and now you identify theft someone?
I personally switched because the alternatives finally worked better, did more of what I wanted, better cross platform support... Its the end of an era where the best LAN protocol for a new deployment is SMB...
I acknowledge that switching is easier when you're only talking about a couple dozen boxes instead of 25K boxes. I suppose they'll always be legacy environments. That does not mean I'm actively interested in new advances in token ring cabling, or IPX/SPX, or decnet.