Posts by AndyS 943 publicly visible posts • joined 23 Jun 2009
Re Cash Float - isn't a company credit card the usual way to approach this?
I work for a large multi-national. I generally have no out-of-pocket expenses while travelling. All is paid for on the card, and the expenses sorted before the card is paid.
In a previous system I then paid the card myself (once the money was in my account). In the current system the company pays the card directly (once I've submitted receipts against each item).
In both systems, I'm never out of pocket.
Glad to see some of the idiocy of the Trump administration reversed and an order of sanity restored.
My comment on the article announcing this ban:
"if they've any sense (which they've proved they do - when compared to the toddlers who've been running the US), they'll simply wait out until the idiots leave the Whitehouse, and start negotiating with adults instead."
Seems that's exactly what they did. Good on them.
Also, not as sexist as it may seem! Yvette Amos, for example, did an exemplary job for Wales recently.
US gun laws are a tad odd, but hard to discuss without getting dragged into the cesspit of their politics (which are also... a bit odd).
Fact is though, in this situation, the fellow with the gun acted entirely reasonably and correctly.
The wider discussion about gun laws is largely irrelevant to this situation.
“Why Is @BarackObama constantly issuing executive orders that are major power grabs of authority?”
Trump, on executive orders, in 2012.
He's since issued more than any other president, of course.
I'd imaging there will be quite a lot of tidying up to do when the adults enter the room later in the month.
Well to be fair, you did elect a racist twit. Who has incited race riots, stripped oversight of police violence, put brown children in cages in the desert, refused to address Covid19 appropriately...
So yeah, it's a bit of a shit-show. Turns out the country is, in fact, still full of racists.
Nobody is fine with it, stop being obtuse. If the software was developed in SA, and didn't recognise white students, it would be exactly the same problem.
The word "racist" carries overtones of decisions and intent. This is simply a faulty bit of software - it is clearly not working well, and clearly not acceptable. You could make an argument that there is a problem within the software development community that repeatedly leads to problems like that - you could even argue that this is a symptom of systemic racism within that community.
But to say the software is "racist" is, well, kind of silly. Although, that said, everyone knows exactly what he meant, so, big deal? The issue isn't his choice of words, but the software causing him problems.
Have Boeing given up maintaining that they are "making an already safe plane even safer?"
That was a proper tone-deaf new-speak bit of lawyer approved verbiage if ever I've seen one. I haven't yet seen them actually step up and accept any responsibility, presumably because they're still hoping to keep slipping the right envelopes to the right officials and avoid any legal responsibility (remember Trump's comments about needing Einstein to fly planes... whatever that was meant to mean?)
> That still seems the big question to me. If MCAS can be safely made so much less powerful now, how come Boeing felt they had to beef it up so much in the original test flying program? Because they've still go to fix the problem of the plane climbing due to the aerofoil effects of the bigger engines.
I read that the original design, and safety assessments, were based on an "open loop" design, where MCAS would make a single correction, but that it was implemented as "closed loop", so the system would continually poll the AOA sensor and apply further corrections every few seconds. In an "open loop" system with a failed sensor, the annoyance of a single unexpected downward trim is easily correctable. So the failure was probably classed as "minor", and a single sensor input was allowed. In a "closed loop" system, the plane will continually over-rule the pilots until it hits the ground. Clearly in "catastrophic" territory, as demonstrated by Lion and Ethiopian.
Classifying it as one in the safety assessment, but then erroneously implementing it as the other, is what should land people in jail for several hundred counts of manslaughter (never going to happen though - there was good profit to be made, after all). It's also a text-book example of why self-certification is a corrupt disaster. The true American way.
On the positive side for Boeing, fixing it is as "simple" as implementing it the way it was originally designed. Plus a few other tweaks to improve it, since even the way it was originally designed was a bit crappy, and everyone is now looking (only using one sensor, when two were available).
Didn't downvote your original comment, but how about this:
> I, for one, stopped clicking years ago.
Right. So after reading the article and writing a 4,000 word essay explaining everything wrong with the world, you end by saying you don't even bother clicking through to articles like this any more.
Nobody likes a whiner.
> Let's say I'm reading/writing code with the occasional line which is 200 characters long. And I'm not using a pretty-printing editor which will dynamically reformat code for me.
Isn't that the exact crux of Linus's argument? That you should get a better editor, which can handle this?
Ooh, can we do Jake too?
When I was running Tescos from the USS Nimitz during 911, I forewent my entire paycheck to keep the kids in school. Of course it wasn't easy, as we were limited to using IBM 5100s, but I wrote the assembly code myself to lash things together, while under fire from the Japanese.
What about trying to link employment and home address to car insurance premiums? Or lifestyle choices to predicted medical issues?
There are plenty of legitimate uses too. Or was it wrong for the early HIV campaigns to target the gay community? Because that's the sort of information that a model like this could produce, with massively increased accuracy.
Here are some names you should have used:
Arheddis Varkenjaab and Aywellbe Fayed
Arhevbin Fayed and Bybeiev Rhibodie
Aynayda Pizaqvick and Malexa Kriest
Awul Dasfilshabeda and Nowaynayda Zheet
Makollig Jezvahted and Levdaroum DeBahzted
Steelaygot Maowenbach and Tuka Piziniztee
> is it still identifiable private info?
This was my question. To take an example, Bob is gay, and has HIV. That's pretty sensitive information.
However, the relevant trained healthcare model will simply weigh the sexuality as a factor against HIV infection. In simplistic terms, the inclusion of Bob's data may push the model's link between those two facts from 20.1 to 20.2. Bob's statistics affect the model's behaviour slightly, but there is no way you could ask the model if Bob is gay - it simply doesn't know or care.
Obviously if the training dataset is retained, that's a completely different story - but I can't see how personally identifiable information could be gleaned about a single subject from a true "black box" model.
This style of response upsets me.
Basically what you are saying is that, in the event of this system failing and trying to nose-dive the aircraft into the ground and kill everyone on board, if nothing else is going on and the cockpit is quiet, there is a procedure to handle the situation.
Now assume you are climbing towards mountains, and your airspeed indicators are bollocksed. You think you are going too fast, so you withdraw the flaps. Your angle of attack rapidly increases, then the stick-shaker activates, there are voices shouting "DON'T SINK! DON'T SINK!", "TERRAIN! TERRAIN! TERRAIN!", "STALL! STALL! STALL!" or similar. The noise of the stall warning horn is blaring out, the yolk is shaking aggressively. Meanwhile the nose of the plane keeps being forced down. Your airspeed is increasing, but you don't know how fast you're going. You disable auto-trim, and there's a temporary lapse where you get the nose back up, bu all the stall warnings continue, so your reflex is not to pull up hard. The nose drops again. You are pulling back on the stick with around 50kg of force, which should also disable all auto-trim and autopilot (as it did on previous 737s) but by that point you are only 1000ft above ground level, your instruments are still reading nonsense, you have no fucking clue why the plane is nosediving and you can't pull the stick back any harder. There is a quiet clickity click noise from the trim wheel, which you already disabled but for some god-forsaken reason it's going again. You're in a nose-dive, 40 degrees down, 500ft above the rapidly approaching mountains.
That all happens in about 2 minutes, in the case of Ethiopia.
Yes, it might be possible to (temporarily) disable the system. No doubt an exceptional crew might handle the situation better. But you know what? Blaming the pilots here is not how you prevent this shitstorm happening again.
I knew a guy who had a student doing some all-night runs in a laminar flow wind-tunnel.
The tests would start in the evening and run smoothly for a few hours, after which (at around 10pm) the flow would unexpectedly transition to turbulent, ruining the rest of the night's results.
The supervisor eventually sat with the student to see what was going on. The tests were set up, the tunnel started, all good. A few hours past, dinner was eaten, and boredom set in. That's when the student turned on the (loud, heavy-rock) stereo system...
As I posted above, it's interesting how similar the logos of some of the worlds main space agencies are. Don't get too excited.
Ignoring your dribble about how anyone could do this thing that nobody has done before, here is a good response to your accusation of unoriginality in the CNSA logo. TLDR: the US, Russia, China and Star Trek all have strikingly similar logos.
> At what age?
About 2.
My 2 year old saw a dear for the first time ever, while we were all looking the other way, and immediately said "goat!"
Which, considering she'd only ever seen goats once, about 3 months before, was a pretty good extrapolation, and one that no computer model I'm aware of could currently match.
So I guess "baby" is an exaggeration, but "infant" or "toddler" would be more accurate.
It seems to me that there are two main reasons to brew your own:
1. It is vastly cheaper - a pint of homebrew costs about 20p, and the equipment only costs about £50.
2. You can make the recipe up yourself, tweak things, add stuff in or take it out.
For most homebrewers, it's a combination of the two.
Both of which an expensive, all-in-one, web-connected, "smart," auto sanitising, pod-relying machine will negate.
So what is the point of it, and who is the target audience?
> ... the hidden deep-state
OK, I was following you up to that point, but then... What is this, an Alt-Right US rag? Let me guess, Hillary runs this deep-state from a pizza parlour?
Conspiracy theories don't do anyone any good. Adopting the same language as the nuts across the pond will do you no favours, any more than shouting "AM I BEING DETAINED" will get the police to take your civil liberties seriously.
"Create more velocity in our revenue attainment" = "make more money"
"Jump off a cliff onto new ground" - a fine mixed metaphor indeed. My boss is keen on these, things like "we need to stop climbing the ladder to take a step back and see how far we've come."
"Boosting the quality of service by professionalising people and technology" - what does that even mean?
I ordered an XPS13 with Ubuntu on it. It was the oddest experience.
The computer took so long to ship that my card, linked to the paypal account I payed with, had expired. Instead of holding the shipment when the payment failed, they shipped anyway, and some Indian team were then in charge of trying to get me to pay.
I asked them several times to send me a simple, correct invoice that I could pay via paypal. Eventually, a couple of weeks after the laptop arrived, I got an invoice, without VAT.
By that point, VAT was their problem, and I couldn't be bothered helping them any more, so I paid it.
Some months later, I got another email, from a UK team, asking me to pay the full amount (which was apparently still outstanding on my account). I ignored that, and have never heard back from them.
I've no idea how much they think I owe them, or what will happen if I ever try to order another machine from them...
> Can a third party prove that there's no backdoors in the executable even if the code they give is clean of backdoors?
Assuming the software is fully open-source, it should be possible to recompile from code, and install the locally compiled binary. Assuming your compiler isn't also a DJI product, this should give you near 100% certainty.
This is how hoby-level drones currently work - Betaflight is one of the primary bits of software used for racing drones, and it is fully open source. It's trivial to compile it from source (and many people do, to make it run on unusual hardware or to disable / enable different bits of it). Although DJI's offering is more complex, it could work the same way.
>It should be easy enough to specify a deployable paraglider or similar controlled-descent device, with independent control system and say 10 mins power supply, for use when the main flight power fails.
A full-craft paraglider, with separate power supply to keep it flying for 10 minutes?
Yeah, I suppose it would be easy to specify that. I can specify all sorts of things - like a spaceship with capacity to take 100 people to mars, and return them, with a transit time of of less than 6 months. Oh, and if something goes wrong, it will automatically return safely to Earth.
Doesn't mean there is any link to physical reality, sadly.
@Len, I agree 100%.
Despite the shit-storm that is the US political landscape, at least there are other branches trying to sort out the mess. It almost seems like, in the long run, they may make it harder to repeat the current mess.
The UK, meanwhile, is clearly being just as badly attacked, but I'm yet to see anything suggesting we've really attempted to take it seriously.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2023
