Posts by Stoneshop 5951 publicly visible posts • joined 8 Oct 2009
Reminds me of being drafted in to a development department for a comms processor.
A real software tester* I knew always started his test run by whacking both hands, flat, on the keyboard. In an (un)surprising number of cases, this was enough to bugger up the software being tested.
* I can personally attest to this species having existed, but they have, alas, seem to have gone extinct.
"The last thing I want is one hackable device talking to another hackable device on my doorstep while my futuristic local burglar of 2020, who has just hacked into my fridge to make it think I’ve run out of milk, patiently waits at my front porch for the imminent drone delivery to save him the trouble of bringing a crowbar."
Of course, the well-prepared futuristic burglar of 2020 will have a one-click purchase link on his phone, for a crowbar*, for immediate drone delivery to said front porch.
To prevent leaving electronic traces, it will be ordered through the targeted house owner's account.
* I prefer the somewhat more versatile Stanley FuBar, especially the XXL, for stuff that needs to be separated sans regard for the aesthetics afterwards.
I installed a nice female English voice onto it, which for some reason I find comforting. (That's "English" as in "from England".) I derive endless amusement from her calling the on- or off-ramp a "slip road", and other English phrasings.
Ours is set to Afrikaans. It's far more different from Dutch than English is from American, in grammar, vocabulary as well as pronunciation, but for us it's perfectly comprehensible in general, and driving instructions are a very limited set of sentences anyway.
For maps I've changed to Openstreetmap.
Luckily, the Marketing Department have a solution. They ship loads of crappy products as free samples to dishonest and greedy "reviewers" who then write glowing "independent" reports about how wonderful the thing is. And we all believe them and assume that if stuff (as described above) doesn't work, it's our fault or failure.
"it is very easy to be blinded to the essential uselessness of them by the sense of achievement you get from getting them to work at all. In other words – and this is the rock-solid principle on which the whole of the Corporation's Galaxywide success is founded – their fundamental design flaws are completely hidden by their superficial design flaws."
It does indeed, but after opening this article I can view the picture full-size.
A Twitter link from there has a link to an article stating the case has been closed: Tesla found no technical problems leading to the crash, it's a one-sided accident with a single fatality so no liability to sort out.
See pictures - Driver exceeding 100Km/h on a road that would better be termed a woodland foot path.
If that's hyperbole, it's seriously out of place. And if you took the path in the background in this photo to be the road he was on, you're utterly wrong; it's just a path leading into the wood from the road he was on (N415). Which is properly paved, marked, is as good as straight for nearly its entire length with two very minor bends, a set of traffic lights on that stretch, and has a speed limit of 80km/h. Which he exceeded, as per Tesla, by almost 100%.
Most stunning here is that on ElReg we see so many ill advised folks *debating* the situation.
I see very little of that, mostly with regard to battery chemistry, where it's you just guessing whether it's a LHD car. Which is easily verifiable information, and to save you the effort: all of continental Europe is LHD.
Just require a cutoff switch before the non-standard input enters the house grid, like how a master breaker works.
PV installations tend to have one already, either separate or as part of the inverter. Problem is, that's not enough. A single decent panel can output 18V at 10A or more; a PV setup tends to have several panels in series ("strings"), with one or more strings feeding the inverter. If you interrupt the string(s) at the inverter (which may be deep inside the house, so there's another problem in actually finding and reaching it, especially when the house is on fire) there can still be 150 to 250V DC present between the leads of each string.
Breakers between every second or third panel and the next in a string might be a solution, but requires a lot of extra cabling to bring each of those points in the circuit down from the roof where the panels are to a point where the fire brigade can see and reach them.
Particularly since there was a point when the driver couldn't be approached due to the fire?
It wasn't the car that was on fire, several batteries thrown clear of the car from the crash were.
The problem concerning the car itself were the batteries still inside, and the potential hazards they might pose.
I wonder if EV safety could be improved if they were to run off a bag of spuds in the boot.
You might get an electric vehicle to move if it's not much heavier than that sack of potatoes. Better to turn them into alcohol and use that as fuel for an IC engine.
right, lithium being the second most chemically reactive substance in the known universe, and the most reactive of all of the metals.
It is not.
Reaction with water gets more violent with the periodic table row going up
Dunno. I personally think the actual recommendation should be to drive in a way that allows you to avoid perfectly stationary trees,
Given that Tesla now reports, via AP, that the driver was doing 155km/h* (nearly 100mph for youse non-metric types) prior to the crash he was either unaware (unlikely) of this recommendation, temporarily unable to heed this advice for any number of reasons, or he decided to throw it to the wind..
BTW, I do worry about trees next to the road, but rather in a way that makes that I'm aware of their presence and not having cause to meet them through other events.
* if correct, that's nearly twice the speed limit for that road.
I would have thought that not having a massive block of incompressible cast iron or aluminium in between you and the other object limits the amount of energy that can be absorbed by the crumple zones when compared to an empty void.
I think you have that the wrong way round. With that massive block of incompressible cast iron or aluminium, an impact that displaces the front of the engine will displace the back by more or less the same amount (give or take the deformability of mostly bolted on stuff). With a bonnet holding separate bits of gear such as boxes of electronics and the ventilation/aircon system for the passenger compartment, you can design the front crumple zone for optimum energy absorbtion; any incompressibility of the bits in it is way less relevant.
And a thermometer: you want to know whether any of the battery cells has developed internal problems, and is now near the point of thermal runaway.
Never mind that for this you need to poke around in whatever inaccessible place (design, and/or state of wreck) the batteries are.
"A car battery - which is what a UPS is, tied to an inverter - is 12v 400A or more. I guarantee you that you don't want that discharging through you in any way."
No. A car battery is 12V & 400AH. Amp-Hours is the unit. That means it is theoretically rated to supply 400amps for 1 hour or 200 for 2 hours etc.
400Ah is a honking big car battery; most are 40 to 70Ah for family cars, and up to 100Ah for the average van. I read it as 400A current capability, or in car battery parlance, Cold Cranking Amps, the current it can deliver to your start motor on a frosty morning with your engine's content resembling molasses.
A water short here is a friend, since it will simply discharge intact cells, taking them out of the equation.
Erm, quickly discharging one of those cells will release all the energy present in it, heating the cell and whatever it's shorting it. Now consider there are a bunch of those cells in a more or less enclosed space.
Not good.
The problem with applying water is that once you start you have to keep going until every last scrap of exposed lithium is reacted and gone,
As said elsewhere, it's not metallic lithium you're dealing with (holds for all types of Lithium rechargeables; only primary cells contain some metallic Li). It's a packet of energy that wants to go somewhere, and is capable of doing so vigorously which is, in essence, the problem.
A car battery is 12v 400A or more.
Correct.
I guarantee you that you don't want that discharging through you in any way.
The average person's resistance between their fingertips is several tens of kOhms. Now please apply Ohm's Law to calculate the current through a 12kOhm* resistor, with a 12V, or even 24V source.
1mA with 12V, or 2mA in case of a 24V battery.
Which will take even a small 40Ah battery 4000 hours to discharge. I couldn't handle that, but rather because of boredom and getting hungry, not because of that one milliamp
Any actual danger from an UPS is the AC output and other mains circuitry, same as conventional mains.
There has been, AFAIK, one fatality from a 9V battery, when a sailor rammed the probes of his Simpson VOM through the skin on the fingers of his left and right hand, with the current now finding a much lower resistance path through his body's internals, including his heart.
* makes for easy calculation, and is way below the average. My own is several 100kOhm, even when sweaty.
I would guess they'd have made sure they were stood on something insulating and started cutting anyway.
That will only help against the case where a hazardous voltage (i.e. 50V or more) exists between the car and the ground. With a crashed car not standing on its four wheels only that's not likely to be the case. The problem, as I expect it was, is that it's very likely that one side of the battery pack is now connected to the car body through damaged isolation, and any wire should be treated as now carrying the full voltage of the pack against the body. With over 300V present you don't want to accidentally touch a frayed wire, and the amperage available will make cutting stuff not the sanest of propositions.
Tesla investigating accident is like a butcher certifying his own meat
The gist of the article is that he thinks Tesla should not be involved (I can see both pro and con that argument), but there's one particular fail: "Er rijden al zeker zes jaar elektrische auto's rond. Van Nissan en Renault weet ik dat de accu bij een crash wordt ontladen, zodat er geen spanning meer op de auto staat". Translated: Electric cars have been around for six years already. I know that with Nissan and Renault the batteries are discharged, so there's no voltage present on the car.
Discharging the batteries would release whatever amount of energy was in them, either causing or exacerbating a fire. I suppose he meant "disconnected", and while it might be a transcription error, it appears to be a direct quote.
Anyway, with a car (and its battery packs) that damaged the fire brigade may well be hesitant in accepting that battery safety measures did work correctly, and the car is now inert.
If you replace your keyboard, for example, and you must enter a password,
Hot-plugging while locked only for HID, anything else can wait until the user is at the console again. There'll probably be flaws in this approach too and other ways to extract credentials from a locked machine, but at least this attack will be foiled.
If the laptop has an ethernet port already, you can simply plug into that to get the same effect.
For an autonomous device able to collect the data the same way as described here that would require a short length of cable, a port (or at least the magnetics thereof), something that speaks IP, and a power source. Doesn't need to be much bigger than the USB dongle, but not all laptops have Ethernet ports anymore, and if they do they're often not quite as accessible as its USB ports. Desktop PCs have their network port somewhere on the rear and nearly always in use anyway, but a free USB port on the front would be there in a lot of cases.
And the USB device gets provided with power by the target.
If the idea is to leave the dongle connected to a LAN for remote access then it needs to plugged in permanently.
Have you read the article? It doesn't. It collects the credentials the PC thinks it's sending to the domain controller (because that new interface is now the default gateway), then it gets pulled out again by the cleaner/janitor/inconspucuous visitor. Needs 13 seconds for that.
Afterwards the collected data gets used to access the target in some other way.
North Korea tends to announce theirs, so that their capability to utterly destroy the imperialist capitalist scum won't be swept under the carpet. It's therefore comparatively easy for observers to view the brilliant achievements of the North Korean scientists and engineers, even if they then need to try to belittle those achievements by claiming there were only three successful launches where actually there were four, with two so perfectly alike and synchronised that they appeared to the ignorant as being photoshopped.
That's not a mirror, that's a backup of sorts.
The HP SAN gear I know is easy enough to run as a real remote mirror set. And if you don't want that because you want to be able to cut the link in case some data at the primary site gets corrupted and you want to keep the secondary site from joining in the fun, you have to have some kind of cutout switch you can activate, which obviously requires you (or some monitoring stuff) to detect the problem within the replication delay.
The primary konking out because of a power failure would have been taken care of with a real mirror.
What sort of barbarian keeps Chianti in the /fridge/?!??
A USB dongle that I can plug into a PVR (or other box) that will appear to the box to be a standard USB drive, but in reality connects wirelessly to wherever your actual storage resides.
There's the Eye-fi SD card, adding wireless connectivity to certain digicams (not all of them, and I doubt that it'd work in a SD card reader which you'd need to plug one into any random USB host).
One retro-project that I did attempt was to put an 8 GB (8 because the computer won't see any more) IDE SSD in a circa-1994 Thinkpad 701C (butterfly keyboard), for sexy and silent (no fan) computing--but I didn't make it work.
As I have a couple of 701's, one of them rather broken electronically, a project I have in mind is a Butterfleee: an Asus EEE in a Butterfly casing. Requires several round tuits, though, which haven't turned up yet. And it's probably going to morph into a ButterPi once those tuits do materialise.
The Libretto 110 with a 16G PATA SSD works like a charm. It's got W98SE on it; but I do have Linux distros from around that time ...
but not the more manufacturer specific ones like variable speed Sirius computer floppies, CBM PETs, Apple ][ etc.
I don't know about the other ones, but Apple ][ drives didn't use a hardware stepper controller in the drive itself for the head positioner; the controller generated the required signals in software, driving the stepper motors directly from the controller through 4 transistors. Cheaper initially because the drives contained less electronics, but soon the balance tipped as standard drives were produced in ever-greater numbers. Several ][ clones offered floppy controllers that had a standard Shugart interface.
You need to add a level converter to the Pi anyway (Pi GPIO's are 3.3V logic, floppy drive is 5V logic), a pin header in the right place, couple of Blinken LEDs, and design a chassis to mount it all on anyway. Once you're dealing with all that, you might as well design the entire circuit board using any parts you're familiar with that are fit for the task.
Building one around a Pi is probably fine when you need just a few and you don't care much about a cobbled-together look. But if you're into runs of a hundred or more and they're going to be used in wildly different gear such as CNC machines, synthesisers and traffic light controllers spending some time on engineering can be quite worthwhile.
Plenty of things used to plug into a floppy connector and NOT work without an OS driver (Zip Drives, Tape drives, MO drives etc)
Unless you can build a device that acts as a floppy as seen by the floppy controller and the BIOS, which would mean something that acts as if it has one or two heads (drive controller limitation), 1024 cylinders and 63 sectors (BIOS limitation), AND reads/writes data in a format compatible with the floppy disk controller, you'll need a driver or TSR to make it appear to the OS as a storage device. And you still had to tell the system, again through some driver or TSR, that there was another 'floppy' drive with an unusual number of tracks and sectors, where the BIOS setup only allowed 40/80 tracks and 8/9/15/18 sectors.
BTW, I've only seen tape drives connected to the floppy controller. ZIP were IDE, hooked up to the parallel port, or (later) USB. MO were predominantly SCSI. LS120 were IDE, and I think those 21MB flopticals were too, but I never saw one of those. In all those cases you wanted a faster interface because of the larger data capacity; with those QIC80 tapes it didn't really matter because the medium was slow
FWIW, it connects to the host via a (parallel) floppy drive connector,
Picking a nit: floppy cables may have 34 conductors (half of them ground), but parallel they're not. There's one serial data signal, plus a bunch of drive control and status signals such as step, direction, drive select, track 0 and write protect.
What it will destroy is your mile-high revenue generated by the difference between slave wages and the obscene prices you sell your stuff at.
Destroy? It's not an insignificant sum, but it's just a quarter of APPL's annual PROFIT.
Cook, stop moaning. Neither you, nor anyone employed by Apple* is going to be reduced to eating stale bread and wearing second-hand clothes.
* explicitely excluding supplier's employees, if that wasn't obvious.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
