Just because it's considered to be acceptable, does not mean that it is efficient.
Newsflash: My new-ish router only has 128MB of RAM, and runs Linux. My previous router also ran Linux on 32MB. And in times not-so-distant, I had a mail server, DNS, IRC server, router and multi-user general-purpose shell machine with 24MB, that even from time-to-time hosted a remote X app or three. It worked fine.
There's never an excuse to be wasteful of resources. Never. Suppose it costs a developer 50% more time to create resource-efficient code. And suppose that it costs users an extra $50 to allow inefficient code to run well on their computer.
$50 * millions of users == $$$. (I'll let you estimate how much optimization is worth in terms of developer-hours vs. end-user expenditure, if you choose. Myself, I can see quite plainly that whatever the figure is, it's well beyond my paygrade...and that's without factoring energy.)
I don't mind much that Kroger knows what I buy. What I do mind is that my loyalty card and my credit card are just a database JOIN away from being marketable, personal data, which can be aggregated with other retailers, and subsequently used to form a database of everything that I buy, wherever I buy it from. It's easy to extrapolate that into where I was, when I was there, and what I might have been up to at the time.
Use a different loyalty card along with the credit card at a different store? Linked. And now my cash purchases are trackable as well, across two retailers.
The rationalization is that this likely eventuality is simply too close to big brother for my liking. That it is driven by corporate greed instead of government overseers does not change my opinion of this.
Not that I'm accusing Kroger (or any other retailer) of doing this at the present time. But without strong consumer privacy laws (such as what the EU has), it will remain impossible to know what Kroger knows, or to tell them that I'd like it to be deleted, or preclude them from selling my data to a third party, or even know if that's what they're up to.
What I find more annoying than that, is magstripes and 2D barcodes on drivers licenses. If I walk to the corner store and buy a beer with cash, I don't want my information stored in a database. There's simply no need for it, and zero merit for me: Even with a loyalty card, there are no discounts or points for beer, so I don't even get the illusion of a good deal by giving them my data. But the clerk will, about half the time, swipe my ID through their card reader, and that information is presumptively stored in a database. (Not that it works on mine; neodymium magnets do an adequate job of nuking magstripes, but still.)
If it behaves anything like a "high-performance" VNC, I'm not interested.
VNC is not a network-transparent display. It is a glorified remote framebuffer.
I thought we gave up on dumb framebuffers a long, long time ago, with the first accelerated X server (or Windows 2 or 3 or whatever it was).
One can still configure an X server today to work as a dumb framebuffer, or Windows for that matter Nobody does that. It is lousy and slow. Nobody wants this shit if they have a choice in the matter.
If its networked display options were described as a "high-peformance RDP," I'd be interested in learning more. But it is not.
(For those still living in Mom's basement, both RDP and X are tremendously more useable over a LAN than VNC will ever be capable of.)
I would say that for consumers, its not so much the motherboard. They usually are fast enough, but the routers in between often aren't. So I think the first thing we need is a drop in price for gigabit routers.
Asus RT-N16 is by no means new tech, is less than $100, and has gigabit ports.
Dunno how fast it can actually route, since as a midwestern American, my 12Mbps connection is considered "fast," and the absolute maximum I can get is 18Mbps.
Yes. I understand how DNS works. I understand that it is hierarchical. I said as much.
If you step outside of the box that you're thinking yourself into, I think you'll find a number of different ways that DNS can be used to contain lists of all manner of stuff.
Or are you too thick to realize that *.emei-blacklist.eff.org (or similar) is a thing that could exist? Or shall I suppose that all those DNS-based RBLs that have been in use since forever are impossible?
First offense: Cruel, unusual. (I can provide citations.)
Second offense: Noise ordinance violation.
Third offense: Cruel, unusual. (Again.)
Fourth offense: 10 years suspension? I'd rather they spend a week at Ft. Leavenworth making big rocks into little rocks. After that, two weeks. Then four.
Not if you're a service tech at a modern car dealer: For some makes, having the car have an Internet connection (to communicate with the manufacturer) is a currently useful diagnostic aid during a test drive when attempting to sort out driveability issues.
I assumed that when parent said he'd been dealing with RS-232 since the early 80s, he knew what the fuck he meant when he assigned that term to that timeframe.
It seems to me that this is a wonderful application for DNS as it stands.
I mean, what do we need in an IMEI blacklist? Something simple, hierarchical, efficiently cached, distributed, low-bandwidth? Oh, yeah: DNS does that already.
Of course this implies that the database is accessible by the public...but I don't see any harm in that.
It seems I'm the only one who questions such things, but:
On whose authority was this action pursued?
Since when does the FBI or MSFT or RIAA or MPAA or North Korea or Anonymous or [etc] have a right to diddle with others computers?
What gives them (for any incarnation of "them") the authority to modify privately-owned computers?
If it's for the indiscriminate greater good, then that seems more like military action...which I don't think the FBI is authorized to deal with, and certainly not any private US-based company.
(To be clear: I'm happy whenever I hear about a botnet being destroyed. But I'm unhappy whenever I see the government or anyone else assuming authority where none has been granted.)
In your first case, have you tried getting a PCMCIA (preferred) or Cardbus (less-preferred) or ExpressCard (possibly useless) RS-232 interface?
In the first two instances, generally: Proper voltage levels. Real serial hardware with real hardware interrupts.
ExpressCard is iffy only because ExpressCard exposes both a PCI-Express bus and a USB interface, the latter of which might be used instead of the PCIe bus (depending on the particular serial adapter)....which means you might get a USB adapter in a different form factor.
And why not just run DOS on the machine, anyway? It's not like you need Linux and DOSBox to drive a serial port, even in 2013. Last I looked, new cardbus/PCMCIA RS-232 interfaces still came with DOS "drivers" (more like at-boot configurators, but whatever), and should work every bit as well now as they did back then: Assign it to COM1, run the industrial control program, and move on with life.
In the second instance: If it were a ground issue, it wouldn't be solved by using USB, as the ground is continuous throughout.
In the first instance: Who the fuck knows. There are too many dodgy API layers to immediately blame wiring. But grounding (and any loops) would stay the same, as long as the replacement laptop were plugged into the same outlet as the new (native RS-232-less) laptop: Switching things around should not create new ground issues.
The "standard" is not interesting here. If his old equipment does not support rs-232 to the letter, then any new communication equipment have to work with that. The problem at hand is to keep using that old equipment - because replacing it might be extremely expensive. You don't replace industrial machinery just because it has a quirk in the rs-232 communication.
The "standard" is very interesting here, because if the antique device does not follow it to the letter and this creates a problem, then you have justification to create a better solution.
A MAX232 chip might be part of that solution. A +/- 12VDC supply and some transistors might be a different part of it.
Whatever the case, if the old gear isn't within spec, you can say "Yeah, boss: It should with with 3.3V input, like EIA-TIA says, but it doesn't.. So we need to get a special interface to make the modern widget work with the old antiquated widget."
And then you give him a price. And then he writes the check. And then you hook it up. And then you can go back to reading Slashdot.
This. I've been doing RS232 since the late 80s, and have only come across one relic that requires more than three wires for basic operation: An old TI SilentWriter thernal teletype that I scored in barely-used condition a few years ago.
It -needs- DSR, DTR, CTS, RTS, or nothing happens. And, obviously, RXD, TXD, and GND. Getting 2-way chitchat going on this thing with fewer than 7 wires is a complete non-starter.
Slashdot Mirror
Just because it's considered to be acceptable, does not mean that it is efficient.
Newsflash: My new-ish router only has 128MB of RAM, and runs Linux. My previous router also ran Linux on 32MB. And in times not-so-distant, I had a mail server, DNS, IRC server, router and multi-user general-purpose shell machine with 24MB, that even from time-to-time hosted a remote X app or three. It worked fine.
There's never an excuse to be wasteful of resources. Never. Suppose it costs a developer 50% more time to create resource-efficient code. And suppose that it costs users an extra $50 to allow inefficient code to run well on their computer.
$50 * millions of users == $$$. (I'll let you estimate how much optimization is worth in terms of developer-hours vs. end-user expenditure, if you choose. Myself, I can see quite plainly that whatever the figure is, it's well beyond my paygrade...and that's without factoring energy.)
It's just a fear of data.
From my own perspective:
I don't mind much that Kroger knows what I buy. What I do mind is that my loyalty card and my credit card are just a database JOIN away from being marketable, personal data, which can be aggregated with other retailers, and subsequently used to form a database of everything that I buy, wherever I buy it from. It's easy to extrapolate that into where I was, when I was there, and what I might have been up to at the time.
Use a different loyalty card along with the credit card at a different store? Linked. And now my cash purchases are trackable as well, across two retailers.
The rationalization is that this likely eventuality is simply too close to big brother for my liking. That it is driven by corporate greed instead of government overseers does not change my opinion of this.
Not that I'm accusing Kroger (or any other retailer) of doing this at the present time. But without strong consumer privacy laws (such as what the EU has), it will remain impossible to know what Kroger knows, or to tell them that I'd like it to be deleted, or preclude them from selling my data to a third party, or even know if that's what they're up to.
What I find more annoying than that, is magstripes and 2D barcodes on drivers licenses. If I walk to the corner store and buy a beer with cash, I don't want my information stored in a database. There's simply no need for it, and zero merit for me: Even with a loyalty card, there are no discounts or points for beer, so I don't even get the illusion of a good deal by giving them my data. But the clerk will, about half the time, swipe my ID through their card reader, and that information is presumptively stored in a database. (Not that it works on mine; neodymium magnets do an adequate job of nuking magstripes, but still.)
From the article it appears that he hasn't been charged with anything. At all.
Just sayin'.
You still haven't come up with an explanation of how the hierarchical nature of DNS is detrimental.
I guess because you can't.
Seriously?
Well, then, allow me to retort that I don't like X much these days, either.
FFS.
Inspired, I went and actually looked.
Apparently, the Asus RT-N16 is good for about 141Mbps of actual routing from the WAN interface to an internal LAN.
There are others in the ~$100 range listed which can do many times this much.
The fastest in that list is a visually-hideous and rather expensive D-Link box, which manages a charming 924Mbps.
And thanks for reminding me about the difference between routing and switching, you patronizing shit. GTFO my lawn.
If it behaves anything like a "high-performance" VNC, I'm not interested.
VNC is not a network-transparent display. It is a glorified remote framebuffer.
I thought we gave up on dumb framebuffers a long, long time ago, with the first accelerated X server (or Windows 2 or 3 or whatever it was).
One can still configure an X server today to work as a dumb framebuffer, or Windows for that matter Nobody does that. It is lousy and slow. Nobody wants this shit if they have a choice in the matter.
If its networked display options were described as a "high-peformance RDP," I'd be interested in learning more. But it is not.
(For those still living in Mom's basement, both RDP and X are tremendously more useable over a LAN than VNC will ever be capable of.)
Asus RT-N16 is by no means new tech, is less than $100, and has gigabit ports.
Dunno how fast it can actually route, since as a midwestern American, my 12Mbps connection is considered "fast," and the absolute maximum I can get is 18Mbps.
Yes. I understand how DNS works. I understand that it is hierarchical. I said as much.
If you step outside of the box that you're thinking yourself into, I think you'll find a number of different ways that DNS can be used to contain lists of all manner of stuff.
Or are you too thick to realize that *.emei-blacklist.eff.org (or similar) is a thing that could exist? Or shall I suppose that all those DNS-based RBLs that have been in use since forever are impossible?
What court?
What warrant?
Who?
(No, it's not obvious.)
But my Dude won't run weird stuff on his phone, and my neighbors don't even know what a router is except that it gives them Teh Interwebs.
I'm convinced of the usefulness of these tools just on their very basis, but creating locally-useful momentum is not so easy in this sea of inertia.
First offense: Cruel, unusual. (I can provide citations.)
Second offense: Noise ordinance violation.
Third offense: Cruel, unusual. (Again.)
Fourth offense: 10 years suspension? I'd rather they spend a week at Ft. Leavenworth making big rocks into little rocks. After that, two weeks. Then four.
Not if you're a service tech at a modern car dealer: For some makes, having the car have an Internet connection (to communicate with the manufacturer) is a currently useful diagnostic aid during a test drive when attempting to sort out driveability issues.
I guess I'll keep driving and maintaining my existing cars until I die.
(And for the record: I'm OK with this; my cars suit me perfectly.)
...unless there is more than one of "it", in which case DNS is still viable.
(I hate replying to myself.)
Who owns the control nodes? Who determines whether or not they are end-user machines?
What authority do they have to disrupt them?
(Also: In the US, corporations may not petition for warrants. If you think otherwise, I'm done with this conversation with you.)
Or vendors. Or activist groups.
It must be run by someone, mustn't it?
Me neither, kid.
I assumed that when parent said he'd been dealing with RS-232 since the early 80s, he knew what the fuck he meant when he assigned that term to that timeframe.
You, however, assumed he was an idiot.
Good luck out there, kid.
Like DNS?
It seems to me that this is a wonderful application for DNS as it stands.
I mean, what do we need in an IMEI blacklist? Something simple, hierarchical, efficiently cached, distributed, low-bandwidth? Oh, yeah: DNS does that already.
Of course this implies that the database is accessible by the public...but I don't see any harm in that.
It seems I'm the only one who questions such things, but:
On whose authority was this action pursued?
Since when does the FBI or MSFT or RIAA or MPAA or North Korea or Anonymous or [etc] have a right to diddle with others computers?
What gives them (for any incarnation of "them") the authority to modify privately-owned computers?
If it's for the indiscriminate greater good, then that seems more like military action...which I don't think the FBI is authorized to deal with, and certainly not any private US-based company.
(To be clear: I'm happy whenever I hear about a botnet being destroyed. But I'm unhappy whenever I see the government or anyone else assuming authority where none has been granted.)
Interesting.
In your first case, have you tried getting a PCMCIA (preferred) or Cardbus (less-preferred) or ExpressCard (possibly useless) RS-232 interface?
In the first two instances, generally: Proper voltage levels. Real serial hardware with real hardware interrupts.
ExpressCard is iffy only because ExpressCard exposes both a PCI-Express bus and a USB interface, the latter of which might be used instead of the PCIe bus (depending on the particular serial adapter)....which means you might get a USB adapter in a different form factor.
And why not just run DOS on the machine, anyway? It's not like you need Linux and DOSBox to drive a serial port, even in 2013. Last I looked, new cardbus/PCMCIA RS-232 interfaces still came with DOS "drivers" (more like at-boot configurators, but whatever), and should work every bit as well now as they did back then: Assign it to COM1, run the industrial control program, and move on with life.
In the second instance: If it were a ground issue, it wouldn't be solved by using USB, as the ground is continuous throughout.
In the first instance: Who the fuck knows. There are too many dodgy API layers to immediately blame wiring. But grounding (and any loops) would stay the same, as long as the replacement laptop were plugged into the same outlet as the new (native RS-232-less) laptop: Switching things around should not create new ground issues.
The "standard" is very interesting here, because if the antique device does not follow it to the letter and this creates a problem, then you have justification to create a better solution.
A MAX232 chip might be part of that solution. A +/- 12VDC supply and some transistors might be a different part of it.
Whatever the case, if the old gear isn't within spec, you can say "Yeah, boss: It should with with 3.3V input, like EIA-TIA says, but it doesn't.. So we need to get a special interface to make the modern widget work with the old antiquated widget."
And then you give him a price. And then he writes the check. And then you hook it up. And then you can go back to reading Slashdot.
*shrug*
This. I've been doing RS232 since the late 80s, and have only come across one relic that requires more than three wires for basic operation: An old TI SilentWriter thernal teletype that I scored in barely-used condition a few years ago.
It -needs- DSR, DTR, CTS, RTS, or nothing happens. And, obviously, RXD, TXD, and GND. Getting 2-way chitchat going on this thing with fewer than 7 wires is a complete non-starter.