* Posts by Adam 1

2545 publicly visible posts • joined 7 May 2012

Supreme Court to rule on whether US has right to data stored overseas

Adam 1

Re: @d3vy ... WTF?

> And the DOJ have had 3 years to get an Irish warrant, but continued this fight

What is their angle. I would get if it was some backwater country under a military junta every other week, or if it was an openly hostile Iran or North Korea or something, but Ireland? Just fax the form just to them and as long as there are reasonable standards met, the cd with the data will be in the post (metaphorically). You wouldn't tolerate the argument in reverse. That should tell you something.

WPA2 KRACK attack smacks Wi-Fi security: Fundamental crypto crapto

Adam 1

Re: Why does anyone care about wifi security?

Do you care if someone prints to your printer? Or starts sniffing about to see what might be protected by some ancient long broken Windows authentication scheme? Some of us live on oppressive regimes like Australia and the UK that collect "metadata" from every website you visit because, you know, terrorists. The sort of people who would want to exploit this are no doubt going to be doing things I don't want to go through my connection. Do you really want your media centre indexing the folders of media that they might have on their public share? Didn't think so.

And on your other question, if had need to buy dirty sugar water and use their internet, my VPN is definitely ON with local traffic denied.

Storms blow away 2017 Solar Challenge field

Adam 1

Re: hollow laughter

> Best Li battery - 234Wh/kg

> Petrol - almost 13,000 Wh/kg

That actually doesn't matter. Actually that's not true. It does matter, but it isn't the blocker issue for practical EVs. We get nowhere near 100% efficiency from an ICE even at the flywheel. Then we lose a whole bunch more through the transmission and drivetrain. Even HCCI is only 50% thermally efficient (don't get me wrong, that is brilliant) but those figures are starting with petrol, not crude, so it isn't true well to wheel figures which need to account for transport and refining. It is really hard to do apples Vs apples, certainly not using those two figures.

What currently hurts EVs for range is simply recharge time. If they can get the fast charges to the point where it takes 10 minutes and gives you another 300km, suddenly that plays nicely with fatigue management. I think that is definitely possible and not a ridiculous compromise in most use cases. Alternatively, if manufacturers can agree on a standard then we might even see swap and go battery packs (like we have for BBQ LPG cylinders). That will need machinery to disconnect, swap, then fit the new one. Whilst it'll never be as light as an ICE (all other things equal), having more flexibility over where the weight is put can really help with the dynamics to counter some of the weight penalty. The glass isn't always half empty.

Adam 1

> Sorry, but solar cars are not practical, and will never be practical, unless you can repeal physics.

Ah, but physics is just applied mathematics and the commendable laws of mathematics do not apply down here.

Adam 1

it's not all or nothing

Perhaps the trickle down feature to real world cars will be some highly efficient motor or regen energy storage or aerodynamic profile or low rolling resistance tyres or some magic coating that stops dust from sticking or a hundred other things.

I could make the same arguments about motorsports. What good is a vehicle that needs to be torn down after a few hundred km, that needs special equipment just to start if it stalls etc. But we readily recognise that little breakthroughs while pushing the envelopes can be filtered down into mass produced cars and bikes.

Assange thanks USA for forcing him to invest in booming Bitcoin

Adam 1

Re: Paper profits only

Why can't they cash in their bitcoins? As far as I know the visa status et al ban (although didn't MasterCard break ranks?) only stops credit card payments. It wouldn't stop any other exchange for goods or cash.

'Open sesame'... Subaru key fobs vulnerable, says engineer

Adam 1

Re: This won't be addressed

But a decade before these models the Rex would have been close to the most stolen vehicle on sale. It's a bit hard to judge on popularity amongst car theives because cars are broken into for different reasons (joy ride/use in a crime/rebirthing/theft of contents/stereo etc)

The first two are much harder now due to immobilisers. The third is harder because of datadot. The fifth is less of an issue in modern cars, especially those where the stock head unit supports Bluetooth and Android auto/Apple car play (there isn't much of a second hand back of the truck market for stock head units). Smash and grab is still an issue but frankly a secure rolling code isn't going to help you avoid that.

WPA2 security in trouble as KRACK Belgian boffins tease key reinstallation bug

Adam 1

Re: So in theory

> You can't really trust TLS if you're not really sure if the hostname resolved to the right address.

Sure you can. The fake server cannot fake a signed public key (let's exclude Symantec et al giving out fake ones for the minute) so even if they send you fake content then you are going to get warnings about self signed certificates or mismatched URIs from your browser. They can definitely do a downgrade attack (eg POODLE) when intercepting your handshake. They can definitely block your service. But they cannot trick you*

*I guess another exception is the fruity browser that did a few too many goto fails and thereby passed over that check.

Magic hash maths: Dedupe does not have to mean high compute. Wait, what?

Adam 1

Re: unique mathematical hash

It is irrational to fear an accidental collision in even a broken hash like md5. A deliberate pre image attack is a possibility in other contexts like password storage but such attacks in the context of dedup are irrelevant because to take advantage of it would require you to predict something I haven't stored yet but will need to store.

Why irrational? Because even if we used every computer ever made and dedicated their processing time to try to randomly create documents for billions of years, the probability shows it still be unlikely to occur. It is far more likely to suffer corruption due to bits being induced by a solar flare during a write operation. Russian roulette is not a good analogy in this case. In that game your odds are much worse and there isn't a reward. The reward in dedup is a *lot* less write I/O than you would need to do a subsequent comparison.

Adam 1

Re: unique mathematical hash

> And yes there are programmers out there dumb enough to assume the hashes are effectively unique because the probability is so insanely low of them having a collision and then treat them as if they were actually unique.

So what is your threat model here? That some 733T haxors precompute a collision with a document that you might one day they think that you may wish to store, then store it so that when you do eventually try to save then later retrieve that file, you get nonsense?

Remember if the real document is stored first then the colliding fake will not be stored.

More and more websites are mining crypto-coins in your browser to pay their bills, line pockets

Adam 1

Re: race to the bottom

Rethink that assumption in context of other people paying the electricity bill. If this takes off, the person paying the electricity isn't the one who decides whether a piece of JavaScript gets invoked on some site. Even in the case of traditional bitcoin mining you could have botnets stealing CPU/GPU cycles from hundreds of thousands of PCs/IoT tat. Website owners already believe (generally) that they are entitled to use bandwidth and screen real estate and sound for content they know I don't want (called ads for short). Many don't feel the smallest bit embarrassed about downloading multiple megabytes of JavaScript Frameworks. They believe the visitor owes this cost and inconvenience for the privilege of seeing their wares.

Adam 1

race to the bottom

Crypto currencies (like all I guess) are not intrinsically valuable. The value that they are ascribed is a function of supply and demand. Digital currencies increase computational complexity to maintain scarcety, so if these approaches to website funding are ever popularised, the complexity will shoot up or the value generated will plummet. The pie will simply be divided more ways.

Adam 1

Re: Advertisers won't be happy.

> How difficult would it be to trial it on The Register? Offer readers running an adblocker the choice of paying their way with cryptomining, and see how that works. I'm up fot it.

You guys clearly missed the memo.

OnePlus privacy shock: So, the cool Chinese smartphones slurp an alarming amount of data

Adam 1

Re: To be honest

By a Tesla then what!? Oh man don't leave us hanging like that.

Look! Over there! Intel's cooked a 17-qubit chip quantum package

Adam 1

Re: Mandatory Discworld quote

I both see and fail to see what you both did and did not do there or somewhere not there.

Adam 1

it's a bit hard to tell

With stupid units like millikelvin. Who uses such nonsense?

Why is it so difficult to just say -29.313 Hilton?

Adam 1

Re: Intel YouTubed an unboxing video,

As it transpires, Schroedinger had it backwards. The cat must be thought of as both not alive and not dead at the same time.

Adam 1

Ah, but it'll shuffle the cards better.

Swiss banking software has Swiss cheese security, says Rapid7

Adam 1

Re: How is SQL Injection Still a Thing?

The problem with stored procedures is that they are vendor specific where as plain 'ol SQL is largely cross RDBMS compatible.

Sure you can use an ORM wrapper but these come with their own challenges.

In agree though that if we had a clean slate to redesign SQL, there is no way we would interleave the instruction and data elements or allow inline constant expressions that drive people to try and build up the command with string concatenation.

Adam 1

Re: How is SQL Injection Still a Thing?

I know right. Anyone would think it is still the undisputed number one vulnerability in the owasp top 10.

'There has never been a right to absolute privacy' – US Deputy AG slams 'warrant-proof' crypto

Adam 1

Re: Gravity pulls things down.

> but locally, yes, it pulls things down.

Noʇ poʍu ɥǝɹǝ ᴉʇ poǝsu,ʇ

Adam 1

Re: and key recovery when a user forgets the password to decrypt a laptop.

The user is not obliged to keep their recovery key. They are free to destroy it / never write it down. If they take that approach, then key recovery is not possible. If he is offering a TLA recovery key for each message that we can opt to not record, I think that is a compromise that we can live with.

Adam 1

> Gravity pulls things down.

Nope. In the immortal words of my second favourite accordian player

"My pancreas attracts every other

Pancreas in the universe

With a force proportional

To the product of their masses

And inversely proportional

To the distance between them"

-wise words to live by

Adam 1

Re: "unlike public key encryption, which has easier key distribution but is less secure "

That is a bit of a meaningless statement though. A one time pad (correctly generated) is perfectly secure because it can leak no information about the message. Each bit is equally likely to be flipped as it is to remain constant. It is 0 probability of being deciphered if the key is never discovered and is truly random.

RSA is considered to be very secure because we don't know of any way to factor the product of very large prime numbers. Not because it cannot be done, but because even if we dedicate the world's GDP to trying all possibilities from now till the heat death of the universe we still couldn't crack it. It is "practically secure" rather then "mathematically secure".

What you are saying is that

0 < 1 * 10 ^-someginourmousnumber

Sure. It is. But both are good enough (at least with classic computers)

'Israel hacked Kaspersky and caught Russian spies using AV tool to harvest NSA exploits'

Adam 1

> Where's the popcorn icon?

Or falafel and hummus would be great too.

Samsung rings death knell for disk, gears up for QLC flash production

Adam 1

Re: Death knell for disk?

Nah. What interest would a consumer tat company have in a product that needs replacing soon after the warranty period expires?

Apple's iOS password prompts prime punters for phishing: Too easy now for apps to swipe secrets, dev warns

Adam 1

elegant solution

You've got to hand it to the LaserWriter manufacturer. They introduced a

feature into their password hint that achieves the holy grail of both really helping legitimate users to recall their password and allowing the user to provably tell if the prompt is legitimate or fake.

Leaky-by-design location services show outsourced security won't ever work

Adam 1

solution seems easy enough

Separate out the permissions to view photo from read geotag from image. Any app that lacks the applicable permission gets a modified version of the jpg where the EXIF tag for it has been nulled out. Full permission apps can see the regular file contents. That way your photos app still shows where the picture was taken and the backup app still backs up the full geotag, but some little time wasting app can't without being noteable in the permissions requested.

VPN logs helped unmask alleged 'net stalker, say feds

Adam 1

Re: Not sure why they outed PureVPN

One interesting angle is that China has been putting their foot on VPN provider's throats of recent. I wonder aloud whether one of the licensing conditions required to operate has directly or indirectly exposed their users.

Adam 1

Re: Interesting, very interesting

> If they bill you for usage, they HAVE to keep logs.

If by usage, you mean GB/month then it may be very difficult to not indirectly profile you based upon the size of the traffic between specified periods of time. I wish I remember where I read about it, but there was a fascinating study of profiling a user's traversal of a HTTPS delivered news website by datamatching the size and concurrency of the connections to that server.

If you mean X concurrent users per account, it is very doable. It is hard in the sense that leaking information is very easy to accidentally do, but definitely achievable without substantial cost overhead.

Adam 1

Re: PureVPN has some explaining to do

This isn't about whether they covered their backsides legally. Their capital is trust. They claim to protect your privacy so well that not even they can figure out your identity. They broke that promise if not in letter then definitely in spirit. This will cost them customers. Not that they cooperated (they should) but the mere fact that they had access to that information in the first place means that they are either lying or clueless (or the TLA is lying).

Adam 1

Re: PureVPN has some explaining to do

I think you are missing the point. They claim not to log. They are of course obliged to share the logs with any TLA. They must hand over everything that they collect that relates to the warrant they have been served, which should sum to an empty file.

I can imagine a few alternative scenarios. In one, they already had a warrant for the person who was otherwise of interest, traced his user account through say credit car or email to VPN account link, then asked them to log that user and got confirmation. In another scenario, they could have worked with one of the service providers to deliver him an iframe with some DRM callback to unmask his IP. (Some VPN providers don't necessarily handle ip6 so well or leak via DNS). Or you know, people make mistakes. Maybe he was disconnected one time.

It is the same argument that I make against CAs that are prepared to issue a fake cert of some site to their favourite TLA. You have only one job. If it's easier, replace the good* guy FBI with FSB or MSS and "abusive stalker" with "political dissident" and ask whether the PureVPN lived up to their claims.

*degrees of good

Adam 1

PureVPN has some explaining to do

Whilst it is a good thing to bring an (alleged) piece of work to justice, I'm not such a fan of "the ends justify the means" logic at play here.

"We do NOT keep any logs that can identify or help in monitoring a user’s activity."

Says the massive bannerhere. I look forward to observing some incredible gymnastics of the English language to reconcile that.

SCARY SPICE: Pumpkin air freshener sparks school evacuation

Adam 1

Stop! These puns are too much. I wannabe ill. Goodbye!

Avast urges devs to secure toolchains after hacked build box led to CCleaner disaster

Adam 1

Re: Mycroft Holmes would be so proud...

It is UTC+8, not like it could be Perth or Indonesia or Malaysia or Philippines or some other country working on night shift to make it look like China.

Australia approves national database of everyone's mugshots

Adam 1

Re: Clearly essential...

No this wasn't a dig at across the ditch.

"On 14 July 2017, Ludlam resigned from the Senate after it was brought to his attention by barrister John Cameron that he held dual Australian and New Zealand citizenship, rendering him ineligible to hold elected office in the Federal Parliament under section 44 of the Australian Constitution"

-everyone's favourite reference website

This was his contribution to the stupid waste of taxpayer money* meta data retention bill. I hold different political opinions to him on many issues, but on this issue he was bang on and unlike many others on both sides of the isle, he actually had a grasp of the issue at stake.

*According to my metadata log at some large ISP, I am visiting the IP address of my VPN provider and nowhere else.

Adam 1

outrageous

Just because something can be done technically, does not mean that it should. And just because there are some benefits in an action does not mean that the action is beneficial. Governments do have a responsibility to take measures to keep the public safe. But they need to balance this with their responsibility to restrain the excesses of future governments.

Some people will take a mile when you give an inch. What do you think they will do if you give a mile? Will try say 'great we're happy' or will they push for 2, or 10 miles.

It'll start with crowds and terrorism suspects on a TLA watchlist. Who can argue right? Then we'll add pedophiles hanging around playgrounds. No arguments right? Then bikies, known drug dealers and of course, with this mob, Centrelink recipients. At some point you are going to be sitting on the bench in the park minding your own business when some G4S / Wilson guard crash tackles you, serving an infringement notice because your dog shat on the grass and you have an overdue copy of 1984 from the local library.

Hey if you have such fancy AI, why not use them to highlight unusual packages being carried into the stadium/train/area of risk. This proposal has real consequences for freedom of association. It also has consequences for journalists when their source can be identified. Turnbull of all people should get what that means *cough* spycatcher *cough*.

So yes, there are some positive outcomes to safety but overall this would seem to me a dangerous idea. Let's not throw the baby with the bathwater. If you want to save some lives, maybe try taxing sugar drinks or banning fast food advertising during children's programs or phasing out diesel vehicles in cities or doing a buyback for the old diesel trucks that emit tens or hundreds of times more dangerous NOx particles than their modern equivalents, or coal, or say that all cars sold from 2019 must have active collision avoidance and at least 6 airbags fitted. Any of those will save an order of magnitude more lives.

Adam 1

Re: Clearly essential...

> Are there any sane politicians out there, speaking out against this?

Sadly they turned out to be kiwis.

Adam 1

Re: Public Theatre

My irony meter is looking forward to the 'mandate the burqa' brigade.

Adam 1

Re: Hacker honeypot

Oh sorry, you must have missed the memo. You can rest easy knowing that nothing was hacked. The information was simply retrieved by a legitimate authorised access point by persons unknown.

Google touts Babel Fish-esque in-ear real-time translators. And the usual computer stuff

Adam 1

It looks like you are trying to interchange two words of totally different meaning but similar spelling.

Would you like kelp?

Adam 1

It looks like you're squeezing your phone.

Would you like help?

Developers' timezone fail woke half of New Zealand

Adam 1

Re: Aukward moment

African or European?

Adam 1

Re: Aukward moment

Sorry, but anyone who thinks timezones are simple doesn't know what they are talking about. Did you know that not all timezones are on the hour. Oh hi Adelaide. In fact not all are even on the half hour (hi Kathmandu). Not all timezones are between -12 and +12 (hi Kiribati with your cheeky +14). Then you have daylight savings transitions on top. Congratulations to our Brazilian contingent for not having midnight on Sunday week. Yes folks, it goes straight from 23:59.9999 to 01:00 and causing problems for any software that gets the current date by truncating the current local time.

Why attribute malice where incompetence is far more likely. In many languages, the datetime class has a property to indicate whether it is UTC or local. Most of the time (ha!) it won't make a difference but if you send that via any form of serialisation framework it will get auto converted to UTC and back at the other end. If it thinks the time is already UTC, but the recipient is expecting to convert back to local then in the case of NZ you are going to add 12 hours (or subtract if the same problem in reverse). Either way, lunchtime becomes middle of the night.

Now that's off my chest, I'm going to go and open a new medical practice, providing psychological services to those poor souls who have the misfortune of working with timezones.

Dropbox thinks outside the … we can't go there, not when a box becomes a 'collection of surfaces'

Adam 1

was that a dare?

No no. Not the logo, that looks perfectly fine if you want to refresh your branding. I mean the mountain of dribble. Visual metaphors for collaboration? Open platform? Place for collaborating? Or are we playing buzzword bingo?

Java EE 8 takes final bow under Oracle's wing: Here's what's new

Adam 1

Re: The horse is but a distant memory...

Um. I'm making a joke, a humorous misunderstanding given the two possible meanings of the phrase "Closure will surely follow [the improvements that Oracle are making to Java]". In the intended understanding, the term closure is to express a change in open state of the stable door, from the saying "to close the door after the horse has bolted". In simple terms, too little, too late. The second, unintended meaning is the language feature of closures as present in many languages from lisp to c# but notably missing in Java. The comment about real closures is because many people confuse lambdas with closures. For some use cases they can substitute but not all.

I am not sure why you bring up JavaScript.

Adam 1

Re: The horse is but a distant memory...

Nope. Java doesn't support (real) closures.

Brit prosecutors fling almost a million quid at anti-drone'n'phone ideas

Adam 1

Re: Trained Pigeons

> Build moats round all the prisons

You guys really need to stop sending your politicians such mixed message! I thought you emphatically told them NO MOATS a few years back.

Adam 1

Re: Trained Pigeons

> Build moats round all the prisons then get some duck billed platypus wearing monocles, top hats and walking sticks trained to confuse anyone trying to smuggle contraband using interpretive dance routines.

> I see no problem with this plan.

Did it slip your mind that all* animals from Australia are venomous?

*Ok, I exaggerate. Some of the sheep are safe.

Java security plagued by crappy docs, complex APIs, bad advice

Adam 1

Re: if only

Sorry but that argument is pretty flaky and would equally apply to Wikipedia / Reddit / El Reg commentards.

It solves a different problem to documentation, which tends to focus on classes, constructor overloads, methods and properties. SO does a reasonably good job at pointing back to the official documentation where applicable.

Two examples from recent experience. I was trying to figure out how to write

if (old!=null XOR new!=null)

Obviously for bitwise xor there is ^ but there is nothing for a condition. SO reminded me that XOR is the same thing as !=

In another example, I had two Hashsets and wanted to know if there were any differences between their contents. The top answer would have worked but it was frankly going to be inefficient. Sure enough the first comment below the answer states that it works but it is going to be O(n^2) whereas the other answer (which was slightly less voted would be O(n). Sure enough, the second answer pointed out the method name and a link to the documentation for it.

The other thing that SO does much better than the documentation is explain why something was/wasn't done, like why can't youb in c# yield return inside an anonymous method? Is there a fundamental ambiguity around what you are trying to do or is it just not high enough on the backlog because people want other features? The documentation just tells you you can't do that. Maybe that's all some people care for, but I need to know reasons. For the curious, the answer is too much effort and not enough demand especially now local functions can do this.