Posts by Filippo

Excuse me, what just happened? Resilience is tough when your failure is due to a 'sequence of events that was almost impossible to foresee'

Re: Partial Failures

You're switching one catastrophically bad and very hard to diagnose thing for one that's catastrophically bad but which I assume leaves a log with the exact problem somewhere. That's a pretty good improvement.

The best time to plant a tree is 20 years ago. The best time to build a semiconductor foundry is 5 years ago

Great. It looks like we've moved past "privatise profits, socialise losses". Now we're socialising investments too. Why don't we just transfer taxes directly to corp shareholders and be done with it?

Re: So

The bad planning here was made by the market actors. This is a spectacular failure of market freedom.

Pakistan's Punjab province tells citizens to get jabbed or have their SIM card blocked

"According to research from academics in Spain and the UK, this reluctance has been fuelled partially by an incident where the CIA used a hepatitis B vaccine drive in its search for Osama bin Laden. The operation sought to obtain DNA samples from infants in order to find a genetic match with the fugitive terrorist leader."

I didn't know that. That's horrible. Undermining trust in medicine ought to be a war crime.

An anti-drone system that sneezes targets to death? Would that be a DARPA project? You betcha

Yeah, but it's not their drone. I'm not a very big fan of military occupations in general, but they can hardly be blamed for failing to safely disable someone else's flying bomb.

'Universal Processor' startup Tachyum unveils full-system Prodigy emulator ahead of sampling later this year

Exceptional claims demand exceptional evidence.

UK launches consultation on forcing landlords to allow gigabit broadband upgrades

At least part of it has to be this. I, for one, immediately hang up and blocklist anyone who calls from an unknown number and declares to represent an utility provider. They do not get a chance to explain. And it could be worse; I know several people who just don't answer to unknown numbers, period.

I am well aware that I might be missing on some relevant communication this way, but what else am I supposed to do? I get several such calls every day, and determining whether they are genuine takes a significant effort - and 99.99% of times they are either spam or scammers or both.

If my provider really needs to tell me something, they'll have to contact me via snail mail, or an email that I can recognize as genuine.

Or, even better, phone companies should take action to stop unsolicited calls for real, and then they may get to use the medium for its actual purpose again.

Apple settles with student after authorized repair workers leaked her naked pics to her Facebook page

Re: How?

I guess it depends on the nature of the fault. Ability to switch on may not be sufficient to validate the repair.

G7 nations aim for global 15 per cent tax on big tech and bin digital services taxes

Re: Too soft too weak

Not necessarily. It's true that corporations pass any increase in costs to the customers, but this makes them less competitive against companies that weren't dodging taxes to begin with (i.e. small/medium businesses). And more competition is just about the only thing that can actually lower prices for customers in reality.

Tiananmen Square Tank Man vanishes from Microsoft Bing, DuckDuckGo, other search engines – even in America

The point of duckduckgo is to prevent data from flowing from you to the big corporations. The other way around is fine.

The policy of truth: As ransomware claims rise, what's a cyber insurer to do?

IMHO, paying ransoms ought to be illegal.

NASA doubles down on Venus missions, asking what made the planet uninhabitable

Venus is substantially closer to the Sun than Earth. That probably accounts for a good bit of its temperature. However, Venus is a lot hotter than Earth - more than what can be attributed to its position alone. That's because its atmosphere is full of greenhouse gases that trap heat, or at least that's what we figure based on the data we have.

The point of the missions is to get more information about that. If there's some currently unknown planetary mechanism that can have a big effect on heat retention (in either direction), it would be really, really nice to know about it, given the current issues on Earth.

Fun fact: there's an altitude on Venus, high up in the sky, where both the temperature and pressure of the atmosphere are nice for humans (alas, the composition is still deadly, so no cloud cities).

Wyoming powers ahead with Bill Gates-backed sodium-cooled nuclear generation plant

Re: You know what

I'm sure there's enough researchers in the world for humanity to work on both better aircon and better power stations at the same time. I strongy doubt we'll fix the energy problem by only attacking it by one side.

Microsoft previews Hot Reload for .NET developers, sets date for .NET 6

I wonder if there's a way to "hot reload" outside the debugger? There are cases where I'd like to be able to upgrade a running program without stopping it.

Google employee helped UK government switch from disastrous COVID-19 strategy, according to Dominic Cummings

Re: Hang on

The fundamental issue is that exponential growths are one of those things that the human mind is really, really bad at modelling. Another one is very low probability events, e.g. getting severe side effects from a vaccine.

For such concepts, you cannot just trust your instinct. You have to run the numbers, at least in a crude way. Problem is, the vast majority of people do not run the numbers on anything, ever, for any reason. They trust what their gut tells them. That includes politicians, and how could it be any different? The whole system is supposed to make them represent majorities, after all.

In late February, even first days of March 2020, I wasn't too worried about the virus.

Then I fit the numbers into an exponential model, and I still wasn't worried, because I just thought that it could not possibly be that bad, and surely I was just using an oversimplified model, and I am not an epidemiologist after all.

I only crapped myself when I saw that, 10 days later, the actual numbers matched my prediction to within a 1% margin. If I had been in charge of anything, that would have been 10 critical days wasted.

And that's me, with my STEM training and all. What chance does a regular Joe have, when half the media tells him to panic and the other half tells him to chill?

Surprise! Developers' days ruined by interruptions and meetings, GitHub finds

I wonder whether they looked at the quality of meetings. Three people attempting to nail down a tricky requisite definition is a thing. Twelve people from five different departments, discussing ten points, each of which is only relevant to three or four participants... is another thing.

Activist millionaires protest outside Jeff Bezos' homes to support tax rises for the rich

Re: A useful little test

That "test" is testing for the wrong parameter. The problem being denounced is not "we are not raising enough taxes" or "we don't have enough money for the poor"; it's "the current taxation scheme is unfair". You can fix "we need money" with voluntary payments, but you cannot fix "unfair" that way. You just can't; it would literally be making the problem worse.

You can, of course, argue that the current system is not unfair at all, and that would be an opinion worth debating.

But people who don't share that opinion don't need to give to charity and/or voluntarily pay more taxes in order to be coherent. Because those actions don't do zilch to make the system more fair, and arguably work against that objective.

Re: Tax avoidance costs

That doesn't work. You can't base a system around voluntary taxation, because it gives a strong competitive advantage to non-volunteers. Such a system is inherently unstable. Taxation must be as evenly spread as possible (for some vaguely reasonable definition of "even"); a voluntary basis is the opposite of that.

These people who protest, I'd wager the reason they protest is because they understand the above perfectly. So, no, volunteering money would be against the ideals they're pushing.

Man paralyzed from neck down uses AI brain implants to write out text messages

Re: Why no delete?

Well, they could implement escape sequences...

Elon Musk hits the brakes on taking Bitcoin for Tesla purchases

Re: What is the point

The original philosophy was extremely hostile to anything centralized. However, a currency that anyone can arbitrarily make more of doesn't work. Hence, decentralized but difficult to make.

I hear that other cryptocurrencies consume less power, but I don't know how that works in practice. It seems to me that any cryptocurrency must either be centralized at some point, or consume massive amounts of power as it becomes more successful. I don't understand how a decentralized-but-not-power-hungry scenario could be possible, and I'd welcome pointers in that direction.

"Sure electric vehicles may ultimately burn fossil fuels as well but they burn less than a directly fossil fuelled vehicle does because the generating apparatus can (usually) be run at peak efficiency whereas a fossil fuel powered car seldom does. The electric car therefore (as far as I can see) does less harm than those vehicles that it replaces."

Also, an EV can in principle become cleaner as the grid becomes cleaner, while an ICE would still be burning fossils even if we had a 100% green grid. In theory you can make clean synthetic fuel from green energy, but in practice the efficiency is horrible.

Preliminary report on Texas Tesla crash finds Autosteer was 'not available' along road where both passengers died

I was worried about this when I got a car with adaptive cruise control and lane-keeping assist. It's not enough automation to allow you to just do something else, but it's enough to give your brain significantly less to do. I was concerned that driving would fall under the threshold of too boring to pay attention to.

In practice, it turns out that I use the spare bandwidth to pay a bit more attention to other objects besides the car directly in front and the lanes, e.g. cars on other lanes that look like they may switch without signaling, cars further ahead whose behavior may suggest problems ten seconds from now, things like that.

It's when you get enough automation that you start feeling like you could check your emails - that's where the real danger begins.

Re: As a motorcyclist....

Can't talk about everyone else's, but lane-keeping and adaptive cruise control in my car generally behave very nicely towards cyclists or motorcyclists, in that they are considered the same thing as a car, i.e. the bike occupies the lane. By comparison, most human drivers will attempt to overtake a bike within the same lane, or only partially switching lane. I suspect that auto-pilot cars may easily turn out to be safer for bikes than human-driven cars.

Train operator phlunks phishing test by teasing employees with non-existent COVID bonus

Re: But isn't this what (real) criminals would do?

Yeah... there isn't any reliable way to tell a phishing attempt by writing style alone. You also can't trust the sender address; it can be spoofed.

My main rule is to ask someone I trust for verification, before visiting any URL with a public-usage domain or a domain I don't know (remember to look at the actual link, and not at what the text says), or answering to an address from any such domain (remember to look at the reply-to field, and not at the from field). That should cover most cases.

Also, I assume anyone who calls me on the phone and tells me he's from my (or any, really) bank/utility company/insurance provider/whatever is a scammer until he can prove otherwise. No, knowing my name, date of birth, or other easily obtainable information, is not proof.

Microsoft has gone to great lengths to push its tech, but survey suggests many devs slipped through the .NET

Re: Hell I'm still using MFC

I'm about to go replace a DOS-based system that has been running for nearly 30 years. The new version has a WPF GUI. I fully expect the system to last 10-20 years. I don't give a carp that WPF won't be receiving any more updates. It's a framework that works, that has been around long enough that its bugs are either fixed or well-known, and if there's one thing I can trust Microsoft for, it's to keep legacy stuff running as long as it's reasonably possible.

Philanthropist and ex-Microsoft manager Melinda Gates and her husband Bill split after 27 years of marriage

Although the etymology of "philantropist" is "friend of humans", in current usage (i.e. the last several centuries at least) the word does not quite mean literally that.

Also, https://xkcd.com/871/

Appeals court nixes online blueprint sharing ban on 3D-printed 'ghost guns'

Re: Why bother with 3D printing

Roughly two thirds are suicides. A very small amount are accidents. So, if you only count homicides, the number is "only" about 40 times bigger than in the UK. And yes, that's adjusted for population.

Honestly, if you want to attribute that to "public attitude", be my guest, but I would then argue that if the populace has *that* attitude, then they *most definitely* should not have firearms.

Re: Why bother with 3D printing

Switzerland and Israel are small countries. They also both have extremely unique historical and cultural traits that are not really seen anywhere else. Basically, you've cherry-picked the most outlying outliers there are.

I'm not saying you should ignore them, but you can't just pick them while ignoring Australia, UK, Japan and almost all of western Europe. If you put all the experiences with gun regulation in these countries together, even with the outliers, it's pretty clear that strict gun control makes the average citizen safer.

Penguin takeover: We tried running some GUI Linux apps on Windows the official way – and nothing exploded

Re: Comparisons please? Benefits please?

Chances are the resources required are greater than either, and I don't see how it could be otherwise given that the described setup is basically running both. I really don't think anyone, anywhere, has ever attempted to claim that WSL would be more efficient that straight Linux.

Obviously, you would not do this just to run gedit. But the point of the article is to demonstrate how WSL works, not how some big and complex piece of Linux software works.

Also, the ability to run simple things is a necessary stepping stone on the way to the ability to run complex things. If, every time a system got to that stage, people said "it only runs simple things, so it's pointless", we would not have systems that run complex things.

Watchdog 'enables Tesla Autopilot' with string, some weight, a seat belt ... and no actual human at the wheel

Re: Genuine Question

That's something that always struck me as weird. It would seem to make more sense for the car to pull over and stop, rather than disable autopilot. I can't think of many realistic situation where driving in a straight line while slowing down is the best solution.

What next for Visual Studio? Microsoft's monster IDE can't please everyone and 64-bit will not solve legacy problems

Why does this need to be a problem? Who said that it is desirable to have one IDE to rule them all? I've used both VS and VS Code, each for different purposes.

Yes, it would be nice to have an IDE that has all the features of VS but at the same is as fast as Code, and so wonderfully designed that the gazillion of features don't feel like bloat. And I understand that's really difficult to do in practice.

But, frankly, I can just use Code for some stuff, VS for some other, and be happy that way.

Truth and consequences for enterprise AI as EU know who goes legal: GDPR of everything from chatbots to machine learning

They can't. That's (part of) why there will be much wailing and gnashing of teeth from the AI evangelists. There's this neat statistical analysis thing we call "AI" (but we probably shouldn't), and it has some fundamental hard-to-fix problems. One of the biggest is that, once you train it, it's a black box.

It may have retained some personally identifiable information, and it may not spit it out in millions of tests, and then spit it out in production for no discernable reason. It happened before, and nobody can prove their shiny new model won't do it again.

In order to be consistant with the spirit of GDPR, this type of model should either be demonstrably not trained on PII at all, or should be considered as itself containing PII and therefore not be released to the public. I suspect the new legislation will think along these lines.

Certain applications, while interesting and superficially harmless, will simply not be possible under such restrictions. Hence the wailing and gnashing of teeth.

Seeing a robot dog tagging along with NYPD officers after an arrest stuns New Yorkers

Re: Facial Recognition Error

It's not that; but until it's way more accurate, it shouldn't be, by itself, sufficient probable cause for arrest.

State of Maine says Workday has shown 'no accountability' for farcical $56.4m HR upgrade

Depends on the consultant. My personal philosophy is that I should never charge the customer for my own mistakes - whether in analysis or development or whatever. If it turns out that something I thought would take a week takes two, I do not charge an extra week. If bugs are found years after installation, I fix them for free.

Sometimes I lose money doing this, but I charge fairly large fees for my sector, so it evens out - however, my customers don't feel cheated by fees accruing for reasons they can't even understand, and I have a strong incentive to make tough-to-break and easy-to-use programs, because my margin lies in getting as few support calls as possible.

However, I work in industrial automation. I don't think my model could work in the fluffier sectors.

CERN boffins zap antimatter with ultraviolet lasers in the hope of revealing the secret symmetry of the universe

Re: Another test of General Relativity

There's nothing particularly "ludicrous" about two opposing forces canceling each other out at some stable point, and the opposite is far from "obvious". Also, using emotional language doesn't lend weight to scientific arguments.

Mysterious case of Arizona state senators skipping a vote on tackling Apple and Google's app commissions

Re: Obvious Explanation

That would mean Apple and Google are effectively too big to be regulated. I find the prospect scarier than plain old bribery.

What happens when your massive text-generating neural net starts spitting out people's phone numbers? If you're OpenAI, you create a filter

My thoughts exactly. I really don't think there's any gray area here. As soon as someone gets his phone number leaked this way and sues, OpenAI is going to be in serious trouble.

Where did the water go on Mars? Maybe it's right under our noses: Up to 99% may still be in planet's crust

Re: Not really surprising

The trick is in the time scale. A Mars atmosphere will eventually escape, true, but you don't need it to last forever any more than you need to be able to build a house that lasts forever. It just needs to be reasonably easy to maintain.

Most claims that Mars had an atmosphere and subsequently lost it to space seem to imply that the process took many, many millions of years.

This means that, if you have the ability to terraform Mars in any time frame that makes sense for a human civilization, then maintaining the atmosphere afterwards would be an utterly trivial effort by comparison.

I can't run the actual math, but I wouldn't be surprised if you just needed to fling an ice-rich asteroid at it every hundred years or so. Any civilization capable of terraforming Mars would find such maintenance to be an irrelevant amount of effort.

Space station dumps 2.9-ton battery pack to burn up in Earth's atmosphere after hardware upgrade

Re: Rare Earths

The cost of such an operation would be greater than the value of any recovered components, probably by several orders of magnitude.

Out of this world: Listen to Perseverance rover fire its laser at Mars rocks as the wind whips around it

Re: Nothing there...

And yet, you felt compelled to comment. Makes you wonder.

A Code War has replaced The Cold War. And right now we’re losing it

The problem is not technical, it's philosophical. If you had an engineer who consistently made buildings that collapsed, you would not fix the issue by providing him with stronger structural materials; you would fix it by sacking him and getting another engineer. Or, even better, setting up a legal framework where people who can't build a stable house don't get to build any house at all, and if they do, they are personally liable.

Similarly, the fundamental issue cannot be fixed with better languages or better compilers or better OSes or better CPUs. Rather, the entire field needs to be rethought. But I don't see any way of doing that without making professional development so onerous that 90%+ of software companies fold, and the rest start charging 10x current prices.

https://xkcd.com/2030/

Surprise: Automated driving biz finds automated driving safer than letting you get behind the wheel

Re: But... but... we are driving because we like it, right?

Eh, it's just tech. It'll start out costly, but if it can be made to work, eventually it'll be standard on all cars.

Upgrade from .NET Framework to .NET 5 can be hard. New official tool may help... slightly

Re: How, then, does one port a legacy application?

Depends on what you mean with "DLL hell". I've always understood "DLL hell" to be the phenomenon where poor versioning of shared libraries can break an existing app in hard-to-diagnose ways. For example, an app could randomly decide to upgrade or uninstall a DLL, and break some other unrelated app that relied on it.

This is not an instance of that - .NET versioning is much better, so that existing .NET Framework apps will continue working just fine alongside new-style .NET Core apps.

This is more like an instance of, a new major version of your favorite framework is out, and it's not backwards compatible: you can keep using the old version and miss on the new features, or you can get the new features by investing significant work.

This situation happens all the time, not just in .NET - actually, .NET so far has been reasonably good at being backwards compatible compared to e.g. web development, where it seems to me that, if you want to stay up-to-date, you generally have to do major upgrade work every few months, and redo the app from scratch every couple of years.

Mobile World Congress to run this year's Barcelona event in June with 50,000 attendees. We're speechless

Re: Under 42 and heathly

Fatality rate is not the whole story. I hope you get better soon, but you should be aware that you have a substantial chance of needing multiple months to recover. A substantial number people still have symptoms after six months. Some are bad enough that they have difficulty working. I'm also young enough that death would be unlikely, but I really, really don't want to suffer fatigue or confusion for months and months.

Also, your mother is not attending MWC, but the man in the grocery line just ahead of her might have, the week earlier. Or the man's wife, two weeks earlier. Or the man's wife's boss, three weeks earlier. Or the man's wife's boss's husband, the previous month. Or... well, you get the point, I hope. Exponentials are nasty that way.

Europe considers making it law that your boss can’t bug you outside of office hours

Good

I'm sure people will find ways around this, but it's important to at least send the message that attempting to get people to do work out-of-contract is *bad*. I've got friends who are frankly being abused this way. Meetings getting scheduled for 9 PM, getting flak for not answering emails on Sunday, that sort of crap. It has to stop. You want this kind of availability, then pay consultant rates.

I was targeted by North Korean 0-day hackers using a Visual Studio project, vuln hunter tells El Reg

Re: Opening some Visual Studio projects can cause code to execute

Visual Studio's project file format includes events, and that includes an event on project open. It's used to do things like update dependencies or whatnot, stuff that you may want to be done before build so that e.g. IntelliSense can work properly. With web development being the unholy mess that it is, people do all kinds of things with it.

I don't know when it was introduced, but it was probably there since they introduced XML project files. Which would be in, I dunno, 2005 or something like that?, definitely before 2010.

Anyway, I get being angry that you can get pwned by a spreadsheet or a fancy text file - but a vsproj? I mean, anyone who opens such a file is by definition a techie. They (should) know it contains scripts.

Want to let an AI-powered doctor loose on patients? Try slapping a food-label-like sticker on it, says Uncle Sam

Re: I'm sorry Dave, I can't do that.......

In fairness, the same label could be sticked to the doctor's forehead, and it would be mostly accurate.

Pirate Bay co-founder criticises Parler for its lack of resilience

Re: Geeks versus Politicians

"History has shown us already what these kinds of tactics can eventually lead too, and it's not looking very sunny."

I agree wholeheartedly! Delegitimating the electoral process, openly discussing plans to murder opposition members, and breaking into parliaments to chase out regularly elected officials, has been shown time and again to lead to very dark places indeed.

O wait, that wasn't what you meant?

UK Space ponders going nuclear with Rolls-Royce: Hopes are to slice the time it takes for space travel

Re: would rather

Putting those two in opposition is an extreme logical fallacy and research on renewables is fairly well-funded, but nevermind that. Exactly why would you shift resources from space exploration towards renewable energy, instead of shifting them from e.g. the military, or subsidies on entertainment (or on fossil fuels), or any of a long, long list of public monies that don't help anything that could conceivably be called "progress", and in some cases are larger than space exploration expenditure by entire orders of magnitude?

Yes, Microsoft Access was a recalcitrant beast, but the first step is to turn the computer on

open source, no-cost, ready to use

I still get a request like that every now and then, and it never ceases to astonish me. You've been customizing a bit of software for 20 years, how can you possibly think to just randomly find a bit of free software somewhere that does exactly the same stuff? Or that re-applying the same 20 years of customizations to a bit of free software is somehow going to also be free?