Yes, Logitech could have done these things, resulting in a product that cost twice as much and half as convenient as what they currently sell. And someone would have found a way to snoop on them eventually anyway.
Utter bullshit. Put the cheapest possible one-wire jack on the keyboard, and on the dongle ($0.15 X 2). Supply a one-wire cable ($0.15) to go between them. When a special key combination is pressed ($0.00) the keyboard selects a new encryption key and transmits ($0.10) it down the wire. Use a high-edge-rate signal so you can use a capacitive return and only need one wire. Receive the key ($0.10) on the dongle and store it. The encryption algorithm can be something like Blowfish (IIRC it fits nicely in even limited microcontrollers and wouldn't add much cost, if any). The cable would only need to be connected during initial key setup. Total cost: $0.65, and my price estimates are extremely pessimistic. They're probably doing custom silicon anyway, where these functions could have been trivially implemented.
If you're at risk of having your keyboard sniffed, then you've got bigger concerns to begin with.
Again, utter bullshit. Once a large segment of an urban population installs a security hazard, it becomes cost effective to build snooping stations and collect passwords and credit card numbers. 25,000 yuppies checking their stocks and buying things over the Internet is an *awesome* target.
They don't seriously think that creative editing can get them out of this, do they?
Who is doing the creative editing? To quote from the GPL In addition, mere aggregation of another work not based on the Program with the Program... on a volume of a storage or distribution medium does not bring the other work under the scope of this License. Just because the program and DLL are distributed together DOES NOT MAKE THEM A SINGLE WORK FOR THE PURPOSES OF COPYRIGHT.
Actually dynamic linking to GPL code does make the linking program based on that code. There is no grey area when is come to the GPL.
True, but so what? This means that when the customer runs the program, they have created a derived work in the memory of their computer. As they do not have the full source used to create that memory image, they must refrain from doing a core dump and distributing copies of the dump.
This has to be a GPL violation otherwise there is no GPL.
Wrong. The GPL would still exist and still be legally enforceable. It just wouldn't have certain nonexistent properties that certain deluded zealots wish it did. If you want a different license that enforces your wishes, write one, but don't be surprised if nobody dares to touch your software.
Legally, yes. If it weren't, there would be no reason for the existence of the LGPL...
Wrong. In the case of dynamic linking, the only difference between the GPL and the LGPL is that the LGPL lets you incorporate header files and other glue in the main program, without having to release the source of the main program. This eliminates the onerous task of writing your own headers from scratch.
IMO, this case comes down to whether SloMedia used VirtualDub header files, or created their own headers from scratch. In the former case, they are screwed, and in the latter case they are not.
Also, you probably need GPL'ed header files to make use of the library.
This is probably the only potentially valid argument for saying the executable which links the dll is a derived work.
Copyright law only cares whether the included work is 1) copyrighted, and 2) included. Meet those criteria, and it's a derived work. A minor infringement is still an infringement.
OTOH, it's pretty easy to write down a list of function names and signatures, and have somebody else independently write new header files. Function names and parameter order are purely functional elements (like numbers in a phone book), and are not subject to copyright. If they actually did this, I think the judge would buy it.
It's (probably) a reference to Linus originally releasing his OS as 'Freax' (or somesuch). Somebody decided that that was too dumb and renamed it to the much snappier 'Linux'.
High UID kids these days -- no respect for their history.;-)
I was talking about internal DC currents. If you have a pair of CPUs drawing 100 W total and running at 1.0 V, that's 100 A. When they go from sleep to full-bore-computation, that's a hell of a sudden current change for the power supply to deal with.
Is it? If the device communicates on "The Internet" then it has a TCP/IP stack.
If the life critical component is not totally isolated from Internet influences, they are idiots. The only signal going back into the medical equipment should be "please dump data". The engineers who design things like pacemakers are totally in tune with the KISS principle, so this shouldn't be a problem.
and shortly thereafter the deaths when someone ownz the system and turns the cities light system all green simultaneously.
Not a worry: they have relays that pretty much guarantee that 4-way green can never happen. Otherwise people would be killed everytime the timer loses its mind. Of course, an attacker could still screw with the timing and cause gridlock, which means you still need security.
Your message sounds like the beginnings to so many ridiculously unsecure/vulnerable systems out there.
Nah. Define it as insecure right from the start and make the absence of encryption conspicuous. That way there's no false sense of security, which is often a worse problem than mere disclosure of information. Ignoring the low-benefit part means you have more resources and fewer distractions for the important part: digital signatures so that the physician can trust the data. You have to fit security to the application, and not just design in every possible security feature simply because you can.
Why is it any more or less relevant to a discussion of the power requirements of an Athlon motherboard than a hard drive??
It's not the peak value, it's the transients. For example, imagine you're doing video processing and it only takes a fraction of the CPU bandwidth. After a frame finishes, the CPUs will almost totally go to sleep and power consumption will drop to a few tens of amps. When it wakes back up, the FPUs, caches, buses, and RAMs will suddenly start drawing as much power as they ever will, which is potentionally 100 amps for a dual Athlon.
So the power supply has to go from 10 amps to 100 amps in a microsecond without drooping excessively. This takes a stout power supply, and once you've made it stout it's fairly cheap to go ahead and raise the maximum power output. Conversely, if you're looking for off-the-shelf power supplies with better transient response, they're typically the higher power models.
The first dual Athlons are intended for big servers anyway, where the hard drives will need lots of power, so there's no point in optimizing for transient response alone.
... but it is absurd that it will use the internet, and whoever is thinking that is a good idea should be fired and removed from anything even remotely technical.
This is pure uninformed hysteria, just like the/. story itself. Unless you're using physically-secure data links every step of the way (secure as in no wireless data, ciphers on every physical link, guaranteed QoS, and Marines ready to deploy when the intrusion-detection system finds something), then you're at the mercy of the public data networks. The Internet is not much less reliable than the phone system that supplies 911 emergency telephone services, and diagnostic data from cardiac equipment is rarely time critical anyway.
Sure maybe it'd be nice if these devices had an encrypted bluetooth/802.11
This is just pure ignorance. Heart waveforms aren't secrets and don't need encryption. All that is needed is simple authentication.
It's completely Dell's fault, as they were the ones who designed a system that can CATCH FIRE when used under normal circumstances, because they failed to do adequate research into the specs on the batteries.
Bullshit. For this particular model, Dell is just sales and support. The laptops (5000e at least -- don't know about the plain 5000) were designed and manufactured by Compal. The batteries are manufactured (and presumably designed) by Panasonic. The external power supply is made by Delta, a well-known P.S. vendor. Everything about this machine is private-labeled by other OEMs for Dell.
That said, Dell did offer a warranty and does seem to be making good on it.
You are right for conventional PPP-over-SSH-over-TCP tunneling, but there is still hope for tunneling over TCP.
If the tunneled connections don't do retransmission themselves, you can just carefully design the tunneling protocol to be very nonagressive about retransmissions. E.g., ask "did you get that" instead of retransmitting the whole packet, and using a steeper-than-TCP exponentional delay function.
And if you have to tunnel TCP over TCP, the tunneler could inspect packets, detect when the tunneled TCP is retransmitting, and simply drop the retransmission on the floor. This is just a tiny step beyond NAT. Of course, if you're tunneling arbitrary reliable protocols, you're screwed. (Although I suppose you could blindy bandwidth limit the tunneled protocol by dropping packets. If you did this agressively enough, the tunneled protocol could be convinced to sufficiently rate limit itself.)
Incidentally, I've been thinking about this because sometimes you don't have a choice about what kind of connection to use. Sometimes you are provided with an arbitrary stream-oriented, possibly reliable, connection and have to make do.
BTW, thanks for the link to the TCP-TCP web page. I can point people at that instead of explaining...
Next time I go on a long flight, I'll make sure the plane was designed and maintained by people who "played amongst beautiful things" instead of learning their jobs.... Of course it's necessary for creative technical people to play with ideas, but that's got nothing to do with printer repairmen.
I don't know about you, but *I* want mechanics that take everything apart just to see what it looks like inside. If you don't play with the tools of your profession just for the sake of playing with them, you're nothing but a drone with meaningless classroom learning.
(BTW, nice trolls. Much better than the usual lamers.)
We bought a new house but had the misfortune of getting it AFTER they'd finished all of the drywall and painting. Running Cat5 is going
to be a tremendous expense for us because they'd have to reopen the walls (and close them again afterwards).
If it's a single-story house, go up in the attic and drill holes to the inside of the walls, make little square holes for the outlets, and fish the wire down from the attic. (This works best for interior walls, as exterior walls have insulation.) You can probably avoid opening up any walls.
use shielded twisted pair ethernet wiring. Not only
will it reduce interference, but it usually is plenum rated to comply with building codes.
I have to agree with the other poster on these: 1) shielding can cause more problems than it solves unless the shielding is wide and unbroken throughout the entire system, and 2) plenum rated cable is expensive and only needed for cables inside air ducts.
[Conduit] can be grounded independently to a grounding rod to reduce noise, and avoid ground loops.
This is dangerously incomplete advice. Large, grounded metal structures, such as conduit, must be interconnected to the power grounding system using an appropriately-large conductor. This is especially important when a second grounding rod is used. (E.g., if a fault energizes the conduit ground rod, and it isn't connected to the power grounding system, large voltages would be present between two ostensibly grounded bare metal objects. This is a recipe for electrocution.)
Plan for expansion! Run at least 4 lines to each room of the house, and pool them all together with a fast ethernet switch, stored in a
cool, dry place.
Now that I agree with. Especially if you're doing a labor intensive install in an old building, the extra cable is cheap compared to the hassle and effort of installing it. I'd add that you should install some 75 ohm coax with F connectors, for television signals. And in an older building, I'd rewire all the phones that can be conveniently gotten to (you can use CAT4 to save a little money).
THere is absolutely no reason why I should run 9million feet of expensive fragile and hard to fish plenum cable just because the drop ceiling doubles as a cold air return in my 2 story building!
PVC == poly-vinyl chloride. Chloride == releases chlorine compounds when it burns. Chlorine compounds == coughing up pieces of your lungs before you die several weeks later. For even a small building, you could easily end up with 50+ pounds of PVC in the ducts, which is the smoke equivalent of, say, 500+ pounds of smoldering wood chips.
Regarding the ductwork being improperly handled in a firewall, that's just foolish. With a proper firewall, half the building can burn down without hurting the other half (much).
Half the building inspectors out there are brain dead anyway.
I think you're underestimating.;-)
And you didn't mention my pet peeve: the building codes that require ethernet to be installed in metal conduit. What, to keep people from driving a nail into the Ethernet and being accidentally digitized?! I mean, how many people were killed by Ethernet that they had to make this law?
Just like everything else: there are good people and non-good people.
This past August - October I worked for Aurora Casket for 2 months without contract.
Employment is an interesting area of the law. It is possible, especially in lieu of a written contract, to be an employee without ever signing anything. In general, if you work at the company facilities, using company equipment, when the company tells you to, using methods specified by the company, you are an employee.
They paid me my first 2 weeks pay in a company check right on time.
Thus creating a paper trail demonstrating the existence and fulfillment of an employer-employee or client-contractor relationship.
During this time management thought things out and changed their ideas... and have this new kid start over with a better well thought out idea.
If no contract or employment existed, then they are wrongfully using your trade secrets and your copyrights. Cease-and-desist 'em, let 'em scream about how they have a right because your were an employee or contractor, then stick 'em with breach of contract. They can't have their cake and eat it too.
End of year I had to pay taxes on that first bit of money they gave me ($2500)
If you're in the U.S., have a look at the IRS's "twenty questions" test for determining whether you are an employee. This seems to be a good page on the subject. (Google found it for me -- I didn't read it closely.) If you answer "yes" to more than 10 of the questions, you are an employee, and the company should probably have paid taxes. Moreover, they should probably also have paid other things, like unemployment insurance, worker's comp, yadda yadda yadda.
On the other hand, it's always a good idea to keep aware of your tax situation even if someone is supposed to pay for it. The IRS commandos don't care whose door they break down. Of course, you probably know this now.
... meanwhile being shorted the $10,000 they still owed me for all my other hours work... I don't have the grounds to fight them in court and by that point I was so screwed for cash there was no way I could afford to.
Firstly, don't assume you cannot afford to fight them in court. Small claims court isn't terribly expensive (for either court fees or amount of lawyering needed). Even if small-claims has a $1500 or $2500 limit, you might could get them to settle out-of-court for more just to avoid the hassle. Even if you just get the small-claims limit, well, that's more money than you had. Secondly, if you are sure they owe you a debt (check with a lawyer first) there are other ways to attack them. For instance, a creditor can file a lien against the debtor's assets: land, buildings, bank accounts, and so forth. This would make them unable to, for instance, mortgage their property until the lien is resolved.
I'd recommend to get all your tangible evidence together (notebooks, bank records showing deposited check, names of people you worked with), and go talk to a lawyer. It isn't that expensive, and even if you don't litigate, a good attorney can help you avoid this sort of situation in the future.
Maybe archie provided this capability... but if so where are the archie servers today?
http killed ftp for browseable information, and the www search engines killed archie for indexing. The point is, that popular olden-day files were duplicated across numerous ftp servers, you used the closest/fastest ftp server, and archie searched by (IIRC) file name/short description. Very limiting, but very useful for certain kinds of information.
Speaking of real-estate, what's the purpose of putting buttons on the units anyway?
Right-click & center-click are the biggies. A "fire" button might be nice for games. A scroll wheel -- which I think I remember seeing on a PDA -- makes reading long things much easier.
(BTW, you screwed up with your opening rant tag. Perhaps you need a copy of Sarcastic Faux HTML For Dummies.;-)
Slashdot Mirror
Amen. The only way RMS will be able to distinguish userland-loaded binaries from kernel-loaded binaries is if the judge doesn't understand him.
IMO, this case comes down to whether SloMedia used VirtualDub header files, or created their own headers from scratch. In the former case, they are screwed, and in the latter case they are not.
OTOH, it's pretty easy to write down a list of function names and signatures, and have somebody else independently write new header files. Function names and parameter order are purely functional elements (like numbers in a phone book), and are not subject to copyright. If they actually did this, I think the judge would buy it.
High UID kids these days -- no respect for their history. ;-)
I was talking about internal DC currents. If you have a pair of CPUs drawing 100 W total and running at 1.0 V, that's 100 A. When they go from sleep to full-bore-computation, that's a hell of a sudden current change for the power supply to deal with.
Brings back memories of sitting up all night at a terminal cluster reading Usenet...
So the power supply has to go from 10 amps to 100 amps in a microsecond without drooping excessively. This takes a stout power supply, and once you've made it stout it's fairly cheap to go ahead and raise the maximum power output. Conversely, if you're looking for off-the-shelf power supplies with better transient response, they're typically the higher power models.
The first dual Athlons are intended for big servers anyway, where the hard drives will need lots of power, so there's no point in optimizing for transient response alone.
That said, Dell did offer a warranty and does seem to be making good on it.
Can Red Hat do this? ;-)
If the tunneled connections don't do retransmission themselves, you can just carefully design the tunneling protocol to be very nonagressive about retransmissions. E.g., ask "did you get that" instead of retransmitting the whole packet, and using a steeper-than-TCP exponentional delay function.
And if you have to tunnel TCP over TCP, the tunneler could inspect packets, detect when the tunneled TCP is retransmitting, and simply drop the retransmission on the floor. This is just a tiny step beyond NAT. Of course, if you're tunneling arbitrary reliable protocols, you're screwed. (Although I suppose you could blindy bandwidth limit the tunneled protocol by dropping packets. If you did this agressively enough, the tunneled protocol could be convinced to sufficiently rate limit itself.)
Incidentally, I've been thinking about this because sometimes you don't have a choice about what kind of connection to use. Sometimes you are provided with an arbitrary stream-oriented, possibly reliable, connection and have to make do.
BTW, thanks for the link to the TCP-TCP web page. I can point people at that instead of explaining...
I tried that, but the PS/2 connector kept getting snagged on conduit junctions.
Regarding the ductwork being improperly handled in a firewall, that's just foolish. With a proper firewall, half the building can burn down without hurting the other half (much).
I think you're underestimating.And you didn't mention my pet peeve: the building codes that require ethernet to be installed in metal conduit. What, to keep people from driving a nail into the Ethernet and being accidentally digitized?! I mean, how many people were killed by Ethernet that they had to make this law?
On the other hand, it's always a good idea to keep aware of your tax situation even if someone is supposed to pay for it. The IRS commandos don't care whose door they break down. Of course, you probably know this now.
Firstly, don't assume you cannot afford to fight them in court. Small claims court isn't terribly expensive (for either court fees or amount of lawyering needed). Even if small-claims has a $1500 or $2500 limit, you might could get them to settle out-of-court for more just to avoid the hassle. Even if you just get the small-claims limit, well, that's more money than you had. Secondly, if you are sure they owe you a debt (check with a lawyer first) there are other ways to attack them. For instance, a creditor can file a lien against the debtor's assets: land, buildings, bank accounts, and so forth. This would make them unable to, for instance, mortgage their property until the lien is resolved.I'd recommend to get all your tangible evidence together (notebooks, bank records showing deposited check, names of people you worked with), and go talk to a lawyer. It isn't that expensive, and even if you don't litigate, a good attorney can help you avoid this sort of situation in the future.
http killed ftp for browseable information, and the www search engines killed archie for indexing. The point is, that popular olden-day files were duplicated across numerous ftp servers, you used the closest/fastest ftp server, and archie searched by (IIRC) file name/short description. Very limiting, but very useful for certain kinds of information.
(BTW, you screwed up with your opening rant tag. Perhaps you need a copy of Sarcastic Faux HTML For Dummies. ;-)