* Posts by Lysenko

986 publicly visible posts • joined 23 Jan 2015

You can't find tech staff – wah, wah, wah. Start with your ridiculous job spec

Lysenko

Re: Not knowing how to look can make it hard to find

It's as if they still living in the Dark Ages, before statistical science was discovered.

A surprisingly large number of people still fall for Gambler's Fallacy, but even those who don't are often guilty of faulty reasoning and imprecision.

For example, given the premise:

"if I toss this coin five times and get 'heads' every time, what is the sixth toss likely to produce?"

... the 'enlightened' person may answer that the sixth toss has a 50:50 probability whereas the gambler's fallacy predicts that it is time for "tails". They're both wrong. There is only a 3.125% chance of tossing "heads" five times with a fair coin so the balance of probability is that the coin is loaded or some other trickery is at work. That means the likely outcome is either "heads" again or else the answer will dictate the outcome by causing the tosser (oooh, err) to alter his technique.

Lysenko

Re: Not knowing how to look can make it hard to find

Our interviews are very real-world, and of course still many people don't make the grade.

I'm sure I wouldn't. I have software deployed in C, C++, C#, Go, Python, SQL92, JavaScript, TypeScript and Kotlin on 4 different embedded ARM architectures plus Linux, Windows and FreeRTOS.

Can I remember the exact syntax, APIs, class libraries etc for all of them off the top of my head? Of course not. As with law (my other subject), you pull all the relevant information from archive into active memory to deal with the case you're handling, then flush most of it again when you context switch to something else.

If you ask me a programming question "cold" you'll get an answer in procedural Pascal. I can't remember map/reduce syntax in every damn language off the top of my head and I'm not going to try. As with law (again), the key to productivity isn't remembering every precedent verbatim it is knowing: that a precedent exists, where to find the details and how to apply it to the problem at hand.

Like Uber, for socialism: Chinese leader calls for more use of AI, big data and sharing economy

Lysenko

Re: State capitalism, not socialism

Socialism at its most fundamental level is about the (democratic) ownership/control of the means of production by the people.

socialism: "a political and economic theory of social organization which advocates that the means of production, distribution, and exchange should be owned or regulated by the community as a whole."

I don't see the word "democracy" there and, if it were universally seen as a fundamental characteristic of socialism, why would anyone ever feel the need to call themselves a "Democratic Socialist" or "Social Democrat"?

However this is just splitting hairs. Majoritarianism is perfectly democratic and the Chinese Communist Party would have no difficulty winning an election and then using its democratic mandate to pack everyone who voted "erroneously" off to re-education camps. Pairing collectivist priorities with democracy ("the dictatorship of the proletariat") is simply "mob rule". Stalin might not have been democratically elected, but one of his contemporaries was.

Lysenko

As for the "socialist" part...like most countries that call themselves "socialist", I'm not sure that we in the West truly understand the meaning and implications

The meaning is straightforward enough. It is a collectivist world view, much like fascism or the human body. You probably try to avoid injury and you have an immune system, but do you really agonize over all the individual cells in your body that expire every day? Would you hesitate (presuming you had the choice) to send millions of leukocytes on a suicide mission to repel an infection? Of course not. The integrity of your body as a whole is of paramount concern and the individual living cells that make up that collective exist only to further the interests of the whole. Even the neurons forming these thoughts are disinterested in their own survival per se - they are just better protected because they are functionally more valuable.

Fascism, Nationalism and Theocracy are essentially the same, they just use different metaphors. Whenever people embrace the concept of there being something "greater than themselves" there is a tendency to try to make individuals service that "greater good".

So the 'Year of Linux' never happened. When is it Chrome OS's turn?

Lysenko

Re: No ...

So, your point is that MS Office is crap because it messes up documents in good, solid ISO-standard format from LibreOffice?

No, my point is that de jure standards (ISO) don't interest people, they care about de facto standards (MS Office). Whether MS Office is "crap" or not is as irrelevant as whether Esperanto has a better verb structure than English. My customers don't speak Esperanto, they won't learn Esperanto and it doesn't matter if it has an ISO standard, UN recognition and an endorsement from the Vatican, it is not the standard they recognize and no amount ISO fiat or philosophical open source posturing is going to change that.

The LibreOffice proposition is equivalent to learning tourist phrasebook French and then demanding that the rest of La Francophonie constrain their vocabulary and grammar accordingly or learn Esperanto. Not going to happen.

Lysenko

Re: No ...

I use both MS Office and Libre Office -- the latter on numerous "other computers" that I own. I use MS Office only on my "main" machine.

Same here. I've got 9 PCs around the house and only three of them have Windows (and therefore MS Office) installed. The others are all Linux and if I'm writing anything on them (usually code documentation) then LibreOffice is fine. It writes basic .doc files perfectly well, however, once you switch on revision control and start passing the document through five different offices in three different countries/languages, all of whom are using MS Office, it is completely unworkable.

It's even worse if you're doing 3D design.Slightly dodgy formatting glitches in a Word document aren't usually fatal (though they could be if it's a contract), but if your 3D model is slightly glitchy vs. industry standards (which means SolidWorks) you can easily burn lots of time/money on useless 3D printing.

Lysenko

Re: No ...

Microsoft have not only released Visual Studio for Linux

True, technically, but you can't compare VSCode with full fat Visual Studio. VSCode competes with Atom and Sublime. It's a programming text editor, not a full IDE.

MS owns the desktop because of MS Office, not Windows. I've seen lots of cases where users had no problem with MATE or Cinnamon but couldn't cope with the incompatibilities and missing features of LibreOffice and Thunderbird or Evolution. I'm one of them. I require a Word Processor and a Spreadsheet to be precisely compatible with .docx and .xlsx formats (including revision tracking and multi-lingual capability) because that's what customers send to me. The same goes for (Adobe) Illustrator, PhotoShop and SolidWorks.

This isn't new. For years Macs survived (almost) solely on the basis that Windows didn't have Quark Xpress. Applications own the desktop: Windows just tags along for the ride because LibreOffice, Thunderbird, InkScape, GIMP, Eclipse etc. are uniformly inferior to Windows equivalents and/or not fully interoperable with de facto industry (Windows) standards.

PS: I'm a Linux and embedded software and hardware developer. I haven't targeted Windows in years, but I still use it because of the above and because Visual Studio (the real thing) is far better than Eclipse or IntelliJ.

Lysenko

Re: No ...

You aware of this I assume? Windows is essentially doing WINE in reverse, starting with an Ubuntu bash shell.

EU: No encryption backdoors but, eh, let's help each other crack that crypto, oui? Ja?

Lysenko

Inspired by the recent cryptominer ruckus...

... and also the various "@HOME" grid computing projects (Folding, Einstein, LHC etc) over the last couple of decades, they could leverage the Daily Mail readership and assorted other "think of the children!" merchants to assist.

Register with "paedos@home" or "terror@home" or to contribute your computer resources to the war against [bogeyman du jour], or include "[bogeyman]miner.js" on your web site to enlist your visitors from all over the world as well[1]!

[1] Yes, I know there isn't enough computing power on the planet to brute force AES in a sensible time frame, but like I said: I'm discussing Daily Mail readers and political venality/stupidity here.

Europol cops lean on phone networks, ISPs to dump CGNAT walls that 'hide' cyber-crooks

Lysenko

Including word of mouth? Then how do they communicate at all given they must assume all methods of communication are not only hostile but capable of being intercepted and decoded (not even one time pads are immune as plods can intercept the pads before they're used)?

You assume that any communication mechanism might be intercepted, which includes the arrest of messengers. Encryption is flawed on its own because even one time pads are susceptible to RIPA attacks so you need to conceal the communication end points and/or employ some form of steganography.

This isn't a new concept. Agatha Christie crims were aware of this and posted messages using plausibly deniable language in newspaper classified columns. The same technique works perfectly well with CraigsList or USENET or (ElReg comments). If you need to send specific instructions that can't be reduced to deniable language then you encrypt and steganographically encode it.

Secure criminal comms isn't so much a matter of strong encryption as evasion of detection and plausible deniability of intended recipient and content. A direct PGP email or WhatsApp message is vulnerable to RIPA so cryptographic strength isn't helpful. Encode the same message in the high order bits of a photo posted to alt.fan.cats and it is impossible to prove that the message even exists, and even if you do, it is impossible to prove who the intended recipient is, thus neutering RIPA. To cite Agatha again: "When no-one suspects you, murder is easy".

Lysenko

Unless of course they are morons and use the phone to call their mum or access their personal bank accounts etc.

99% of law enforcement consists of detecting and apprehending morons. Any crim with reasonable OpSec awareness isn't going to be even slightly inconvenienced by this because they'll already be obeying the golden rule of assuming that all communication mechanisms are compromised/hostile.

NYC cops say they can't reveal figures on cash seized from people – the database is too shoddy

Lysenko
Pirate

I smell rampant mendacity...

... but setting that aside, the fact that they thought they needed a z10 to record the (evidently) colossal volumes of loot they were planning to blag is scary enough in itself. That is the sort of computing power the IRS use to handle asset seizure on a National scale.

Either there is vast corruption in the IT procurement department and they never needed such equipment in the first place, or there is vast corruption on the streets and the NYPD is the largest gang of armed robbers in history.

Windows Fall Creators Update is here: What do you want first – bad news or good news?

Lysenko

Re: Have they fixed the decades old bug in File Explorer ?

What he is talking about is some programs will auto generate an super long file name for a folder

The most obvious example being old versions of npm creating undeletable node_modules directories.

Watch out for Microsoft Word DDE nasties: Now Freddie Mac menaced

Lysenko

Re: Bye

It's not that unusual to encounter road signs when driving, but some people are just in a hurry to get on with their day and pay little attention to them.

So, if there's a clearly signposted side road and some lazy twit pays no attention and causes an accident, the problem is the side road? It should be removed, severely inconveniencing the residents of the small village it leads to, because some clueless morons can't be bothered to read road signs?

I think not. What you do is prosecute the idiots and revoke their driving licenses. In a case like this, that means firing people. Not because it will rectify this specific instance of the problem rather, like many such sanctions regimes, pour encourager les autres.

Raspberry Pi burning up? Microsoft's recipe can save it and AI

Lysenko

Re: Why?

Nano Pi A64 is better (and cheaper) for that job in my experience, but if you're doing motion detection (or something similar) then the software you use has more of an impact. Using a multi-core CPU effectively is nontrivial and it is very easy to burn a huge number of cycles (and therefore heat) in busy waits and context switches.

Lysenko

You're right that an A20 will deliver more MIPS per watt of excess heat, but if you run flat out (as here) it will get just as toasty (the H3 in other models is even more prone to odd thermal effects). That's not a criticism of course. SBCs aren't laptops. Cases, fans, power supplies and other peripherals are in the hands of the user, by design.

Ex-TalkTalk chief grilled by MPs on suitability to chair NHS Improvement

Lysenko

Someone has been bribed or blackmailed...

This person is an icon of imbecility. She was directly responsible for the under investment in security (and over investment in junk mail) that made the TalkTalk fiasco possible. Not even MPs could fail to see that. She couldn't possibly be eligible for consideration based on her record, so there must be something else in play.

No, the FCC can't shut down TV stations just because Donald Trump is mad at the news

Lysenko

Re: Actually

Hitler references are a bit old don't you think?

Rabble rousing populist with ludicrous hair, unexpectedly elected on a vague platform of making Germany America great again, overturning the political establishment, deeply suspicion of an Abrahamic religion, demonizing foreigners in general, berating the media, choosing lackeys on the basis of loyalty rather than competence and floating the idea of subverting the constitution by an Enabling Act Executive Order?

I see your point. No resemblance whatsoever.

Capgemini: We love our 'flexible, flowing' spade

Lysenko

Looks like they really wanted to go with Comic Sans...

... but didn't quite have the cojones, so they settled for Comic Cursive.

Another golden opportunity to use Papyrus missed <sigh>.

Release the KRACKen patches: The good, the bad, and the ugly on this WPA2 Wi-Fi drama

Lysenko

Re: Has to be within range

If you (as an attacker) are going to procure special equipment (as this attack requires) and physically locate yourself in the vicinity of the target then you could also physically tap the ADSL lines[1] which has the added advantage of not showing up in any of the target's logs. More prosaically, since you're physically in the vicinity, you could just look for open windows and burgle the target.

Physical proximity is a big deal in practice. Most attacks that I detect originate with skiddies operating via CN addresses so anything involving visa rules and airfares eliminates the vast majority of potential miscreants at a stroke. In fact (now that I think of it), I can't remember the last time anything suspect resolved to a local (and I mean country, not neighbourhood) address.

[1] TraceSpan and Broadframe make kit for this, but as with the KRACK technique you could build your own.

Russia tweaks Telegram with tiny fine for decryption denial

Lysenko

It also entrenches the principle that the Russian legal system isn't competent. ...[snip]....

Why? Because western courts acknowledge the difference between choosing not to do something and not being able to do something.

Have you read the relevant legislation? I thought not. This ruling makes perfect sense (which doesn't mean I approve of it) because the legislation imposes fines *for being unable* to provide the requested data. The protocols, encryption mechanisms and other characteristics of Telegram operation are entirely under Telegram control so if they choose to implement and operate something that renders them unable to comply with the law then that is their problem.

A UK court would make exactly the same ruling under these circumstances. You cannot use a force majeure defence if the force in question is entirely of your own making. Telegram have a defence in the case of secret chats (peer to peer) conducted via software released prior to this legislation, but they have no defence for continuing to allow encrypted communication via their central servers. They have had ample time since the legislation was enacted to switch to fully plaintext messaging with perpetual logging.

That's as far a this court is supposed to go. The Judiciary are not there to set policy or amend Legislation, they are there to rule on the law as it stands. There may well be a conflict between the legislation and the constitution, but that is a question for the Constitutional Court not the District Court.

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

Lysenko

Re: I've kept wondering why we haven't seen a WPA3 yet

What would be the point? If you don't know where the problem is then simply meddling with the protocol isn't necessarily going to help and will probably introduce new vulnerabilities. Attacks like these don't break AES (which is what WPA2 essentially is) so mindlessly increasing key lengths wouldn't make any difference. If anything, causing protocol proliferation would just fragment research/auditing efforts and risk increasing the attack surface as the lesser used protocol(s) come under reduced scrutiny.

The integrity of cryptography rests on a published mechanism resisting sustained attack by cryptographers over a period of time lengthy enough to confer credibility. That makes it a problem domain where you absolutely should not set about trying to "fix" things until you know exactly where and how they are broken.

Lysenko

...the attacker would have to be on the same base station as the victim...

If that's the case then it is somewhat analogous to a locksmith demonstrating the ability to come round to your house and pick the front door lock. The fact that such an attack exists doesn't instantly obsolete every domestic door lock on the planet.

"Security" isn't a two state condition, it's a sliding scale calibrated to threat level. If you want to keep out someone with locksmith skills then typical domestic door furniture isn't adequate, however that doesn't concern most people because they aren't trying to defend against that sort of threat.

Twitter to be 'aggressive' enforcer of new, stronger rules

Lysenko

Re: Violent Groups

Good point. Looks like some pretty unequivocal "glorification of violence" going here. You can draw a distinction based on fiction, jokes or sport but these guys are entirely serious about actually killing people and taking pride in what they do.

Culture, schmulture. DevOps, agile need to be software-first again

Lysenko

Re: I know my career hasn't been typical...

but, in fact you're breaking the normalisation rules because it is, really, a single entity

That's using an RDBMS as a transaction/version controlled document management system. You can do it with a two column table (PK|BLOB), but this is actually a case where Mongo or Redis (or just the file system) might be more effective. Caching web pages is a similar problem domain with well researched solutions.

If I end up using PostgreSQL to manage two column PK|BLOB tables then I'm usually doing the equivalent of writing a letter in Excel. An application I work on periodically harvests SNMP data which arrives as JSON objects. Those have vendor specific structures and encodings so I shove them in Mongo and then normalise them into PostgreSQL in the background.

I could just dump them straight into a GUID|JSON Postgres table and process everything internally, but that would mean slower performance and less resilience to network outages. I could also leave the data in Mongo and analyse it in situ (through an ORM!?), but that would just mean I'm an idiot ;)

Lysenko

Re: Nice article.

As someone who's seen it actually working in practice, I can can tell you that Agile / Scrum / DevOps *can* work...

... and I can tell you that PRINCE2 and SSADM *can* work. Smart people were/are involved in those too. Does that mean they should be evangelised as generally applicable innovations that will transform the the face of software engineering? No. Let me, correct that: Hell No!.

Building Bezos' Bookshop using PRINCE would be as insane as building a rail signalling control system with SCRUM ... the (worrying) difference is that only PRINCE seems to realise that.

NB: Retired, no, age, wrong, mainframes, never touched one.

Lysenko

Re: I know my career hasn't been typical...

Ah, but that's not "Agile". To be agile you need to be implementing both the front and back ends at once, which means the front end people mock up data as JSON objects based on their (non-existent) understanding of robust data models. In phase two you encounter the total impedance mismatch between that and 3NF and then have to try to stitch it all together with ORM spaghetti.

In phase three you realise that you've buried all the business logic in the ORM so you're not really using the RDBMS (triggers, stored procedures, custom data types, views) effectively and have unnecessary raw SQL being concatenated (shudder) and injected client side thereby undermining security and subverting the DB's ability to optimise and pre-compile queries. Then, after a while, you realise that the festering mess of your middleware is so complex and deeply embedded that you can't actually change it without stopping the business entirely, so you're forced to tackle your performance issues with hardware or bizarre workarounds like HHVM.

As for your second point, "generalising the schema" is de facto denormalization and logically ends with something like Redis. That's fine if ACID and a stable data model don't matter - but (as you obviously know, I'm just venting) FarceSnapTwitGram+ isn't representative of what keeps the economy running. One lost Firebase entry is likely an irrelevance, but one lost (or even "eventually consistent") flight control record could cause chaos with knock on effects across a continent.

Far too much of DevOps/Agile/SCRUM is essentially borrowed from sewage engineering. You keep the pipeline (of crap) flowing at all costs and try not to think about how bad the contents smell.

Lysenko

I know my career hasn't been typical...

... but I've read countless versions of this sort of thing over the years and they universally fail to address three major use cases. The first is obvious: safety critical systems. You cannot have "blame free postmortems" when actual postmortems are involved and the people doing the blaming will be the Crown Prosecution Service. This isn't just about aircraft flight control, it applies to far more mundane things like factory conveyor belt systems. Which brings me to the second category:

Systems where the "user" is simply a wetware robot who exists solely to carry out the instructions issued by the software. An early example of this (for me) was a factory making flavours for ice cream: the ingredients had to be mixed in precise quantities with precise timing and moved along a conveyor arrangement. The process could have been fully automated but (back in the '90s) that was far too expensive and retrieving some ingredients from storage wouldn't be possible with a robot even today (without rebuilding the factory).

In such cases "user feedback" usually doesn't matter. The TIM study ("Time in Motion" - remember those?) tells you how long each process takes and the potential for accidents/mistakes is obvious and can be accounted for. Feedback regarding the process itself (how about we add X before Y?) is worthless because the "users" have no conception of what they are actually doing (neither did the programmer - a guy with two PhDs dealt with that). This sort of system gets its feedback from sensors, not users, which brings me to:

Huge numbers of "computing" devices cannot be patched. The software they run when they ship is there for life. You can't (realistically) reflash a microwave or a washing machine and this has spilled over into the sexy new IoT field with the (catastrophic) addition of Agile/SCRUM/"No blame" people let loose on the code. You cannot "iterate" this sort of software. You get it right first time or you're looking at a product recall and/or market vilification.

All these analyses and methodologies seem to start with the assumption that "software" consists of FaceBook and Enterprise CRM systems. It doesn't. In fact such systems are a tiny minority. There are about 50 billion ARM chips out there and about 6 billion smartphones. What are the rest of them doing?

IT at sea makes data too easy to see: Ships are basically big floating security nightmares

Lysenko

Re: Die Hard: Offshore

They already released that movie this year, and the sequel.

Dear America, you can't steal a personality: GDPR godfather talks privacy with El Reg

Lysenko

That's why I queried the original comment. It seemed to mean either:

a) nothing .. or..

b) that there is some sort of "security" that definitively stops CAN being externally hacked.

The former is a pointless comment and the latter is hogwash. Take the (TI) AM3352 or DRA746, for example. Both are automotive SOCs with dual gigabit ethernet and CAN interfaces. The only thing that keeps the CAN away from the ethernet is whatever firmware the chipset is running (and you can guess why they've got dual ethernet interfaces).

I mentioned OBDII simply because it's a supported/non-destructive way to patch into the CANBus that anyone can play with. Getting to CAN by hacking a Spotify downlink is hacker territory, but there's nothing to stop it besides (hopefully) flawless software/firmware which is exactly the same protection you have against remote activation of your USB webcam.

Lysenko

I remind him that the auto industry is adamant that there's no internet access to the CAN bus

What's that supposed to mean? As it stands, it is equivalent to saying there's no internet access to USB, which is true only insofar as USB signalling isn't a native internet protocol, but that doesn't stop you streaming a webcam over the internet. CANBus is essentially RS485 with packet collision detection so (just like ModBus) it can be thunked over TCP/IP as well.

Any vehicle with an OBD-II port sold in the last decade is surfacing CAN to the outside world, so with a few days hacking (or blog reading) you can graft a cellular modem onto an OBD dongle and route CAN traffic anywhere on the planet. Exactly what you can do with that obviously depends on what features the CAN interface surfaces (intentionally or otherwise).

Neglected Pure Connect speaker app silenced in iOS 11's war on 32-bit

Lysenko

Re: they just need to...

I wasn't doubting that the IDE or the APIs could handle it (that's much the same in VisualStudio), I was referring to application code. I can see it being transparent in an interpreted language or a VM that simply doesn't allow you to mess with pointers, but I didn't realise ObjC was restricted like that.

Lysenko

Re: they just need to...

I'm not contradicting you (because I don't do iOS), but are you saying that shifting from int/pointer/long all being 4 bytes to pointer/long being 8 bytes (but int is still 4) won't break anything? Doesn't the compiler support memcpy, bitwise operations and so on?

Personally I've found 32->64 (on other platforms) second only to ASCII->UNICODE for obscure gotchas in forgotten areas of pointer arithmetic and the like.

Lysenko

Re: they just need to...

Or, if they can't do that (which seems highly likely given the nature of the fiasco), they can just open source whatever they've got an wash their hands of the problem permanently after a few months. If they don't want to be in the software business (defined as not having developers on the payroll) then they should get out of it entirely and stop trying to subcontract/project manage things they don't understand.

Lysenko

Re: They can't get the code for their own app?

Ever tried to exhume orphaned code from escrow? It's often not as easy as just doing a pull from GitHub[1], and even when it is, if you outsourced the development in the first place it's very likely you don't have the skills to do anything useful with it and didn't have the skills to audit it before it was deposited (minified/obfuscated code still compiles just fine and that's as far as people like NCC ever go).

The real issue in almost all these cases is companies outsourcing for skills instead of time. Outsourcing is safe only when you have all the necessary skills on the payroll, but insufficient time/staff to do all the work yourself. When you outsource entire competencies you have no way of effectively managing the project/risks and sooner or later end up playing Russian Roulette.

[1] Why are the developers out of business? If there is a receiver involved and there is any question that payment shenanigans were a factor then the escrow can be locked up for years.

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

Lysenko

Re: Advertisers won't be happy.

Midori is a WebKit wrapper and that makes it unreliable with ad blockers (particularly uBlock) in my experience, and places it behind the curve on JS and HTML standards (just like Safari). I've got it on some Raspberry Pi installs. It seems fine for what it is, but my main PCs are development platforms so only Blink or Gecko will do. I actually have Lynx on one machine. I keep it there just so I can say "I've got Lynx on one machine" ;)

Lysenko

Re: Advertisers won't be happy.

You know many of your readers run script and adblockers

... and most of them will be running a Chromium derivative or FireFox, rather than (shonky old junk like) IE or Safari and that means webworkers are in play. You can therefore offload the mining to the background and use setTimeout/postmessage to ensure that the UI isn't significantly impacted (JS does Win16 era multi-processing -yay!)

I'm up for this too. I'm one of the subset of refuseniks who run uBlock/AdGuard but don't use NoScript because I build web stuff myself and (being a programmer rather a DTP operator), my own "websites" are giant .JS files with render functions rather than HTML pages.

Dear America, best not share that password with your pals. Lots of love, the US Supremes

Lysenko

Re: Oh really.

Oh great, I can stop paying my mortgage and it isn't illegal. Sweet! Of course I shall use you as my legal reference for this.

Correct. You can stop paying your mortgage and you will not be prosecuted for that. You are only acting illegally after you are sued by the lender, the Courts give permission to repossess the property and you refuse to comply with that Court Order, thus placing yourself in Contempt of Court (which is illegal).

For something to be illegal it has to be against the law, not just against the terms of a contract, agreement, promise or any other ad hoc arrangement. That would only apply to a mortgage non-payment scenario if it could be proven that you never had any honest intention of making the payments (i.e. Fraud).

Lysenko

Re: Interesting opinion to have...

under the terms and conditions, it is illegal to hand over the passwords.

No it isn't. T/C are contracts and breach of contract isn't illegal. Also, as a general principle, any contract that contains illegal terms is liable to be voided entirely so you can't escape a charge of (for instance) harbouring a felon just because you to signed a tenancy agreement with a confidentiality clause.

coming from the EU, they would need the written permission ... before they can hand over the information, otherwise ... liable to prosecution back in the EU.

Possibly so, but also not relevant. Countries aren't obliged to take any account of other countries laws unless they sign specific treaties (e.g. WTO) to that effect. Austrian law requiring removal of face veils directly contradicts Saudi law which demands them and a female Saudi national returning home after visiting Austria could theoretically be prosecuted for indecency, particularly if the incident occurred "air side" and therefore not technically within Austria. Caveat viator.

Lysenko

Re: Why the upset?

You're mixing up contract and criminal law. An EULA is a contract; it can only apply to someone who intentionally agrees to be bound by it. That means that you could be sued for giving your spouse the password, but your spouse is in the clear (in that regard) because they never agreed to the Netflix EULA in the first place.

Criminality only enters the equation if your spouse then steals goods or services from Netflix. This is an offence irrespective of how your spouse gained access to their network. Handing over your password for the purposes of facilitating a crime may well make you an accessory, but that isn't exactly a groundbreaking precedent. Pretty much any action intentionally taken with the foreknowledge that it will be used to facilitate crime is potentially illegal.

"Sharing passwords" is a red herring. It is exactly the same as sharing any other keys, combinations, or access codes with someone who you know intends to use them to steal something from a third party.

Lysenko

Re: Why the upset?

As was explained in the article, it seems that he was jailed for using a shared password to access a computer system, not for stealing data.

Unlikely. The Act in question does not criminalise unauthorised access per se except in the case of Government computers or others having specific national security implications. Otherwise the prosecution also needs to establish either:

(s4) "knowingly and with intent to defraud..."

(s5) "...as a result of such conduct, causes damage and loss."

That's easy enough to do in this case but not in the case of simply accessing your wife's FB account, with her permission. Of course if you use that access to start trolling or scamming then that's a different matter.

What is more interesting is the Netflix scenario. You are not a party to the Netflix T/C so you're not liable for any breaches in contract/civil law (that's your wife's problem), but you probably are guilty of crime. Personally I don't find this surprising. If my wife gives me the key to the office photocopier then that's an issue between her and her employer. However if I use that key to obtain photocopying services for myself then that's a matter for the police.

Alibaba's Jack Ma says: Relax, we're too wise for robots to take our jobs

Lysenko

Re: Famous last words.

The Mayan civilization was also pretty clever (and very scientific, too)

No, they were very religious and consequently invested far too many of their limited resources on useless bullshit like pyramids which are economically valuable only to tourism companies of later centuries. You can point many of the same criticisms at us of course, but that doesn't make the Maya a "scientific" civilization. On the contrary, they were remarkably North American, as is illustrated by the strong correlation between climate change denial[1] with belief that angels are real and prayer works.

[1] The Maya still exist of course, but the classic Maya civilization fell long before Columbus arrived because of drought, over-exploitation of agricultural resources and the belief that pouring ever more resources into Gods and Religion could fix things. The Romans, on the other hand, might read some auspices and pray to Jupiter - but then they would go and build an aqueduct or dig a canal.

Night out in London tonight: Beer, Reg and platform wars

Lysenko

Re: Pah!

Pah^2 ... that's just an argument about what sort of net curtains to hang in your cookie cutter x86 front room. Now, Z80 vs 6502? That was a real ideological debate !! ;)

Swiss banking software has Swiss cheese security, says Rapid7

Lysenko

Re: How is SQL Injection Still a Thing?

I mean, this is a language that is roughly the equivalent of COBOL. It's been invented almost forty years ago, on the mistaken assumption that programming languages should look like English. Its syntax completely obfuscates the execution logic.

As usual with knee jerk ageism, you miss the point. The new hotness in web monkey land is GraphQL which makes everything look like JSON, conveys even less about execution strategy than SQL (at least you can infer something from JOIN) and only obstructs nefarious injection by virtue of the fact that no RDBMS understands it natively (yet).

The problem isn't the query language, it is the vast attack surface you open up when you allow a front end to send commands either directly to an RDBMS or via middleware built by people who only understand ORMs and therefore can't code triggers and stored procedures effectively.

UK Treasury Committee chairman calls on Equifax to answer for breach omnishambles

Lysenko

Why are the details of 700,000 non-US "customers" included in along with test data?

Quite possibly they were the test data. For a predominantly anglophone country like the USA, the logical countries to source customer name lists for testing are UK, Canada and Australia. You can't use random strings in place of names[1] (cardinality is all wrong, so index optimisation would be too), you don't want data from "foreign language" countries and you certainly don't want to be using actual Americans (because: Tort Lawyers).

[1] Yes, I know there are ways around that but they require a degree of competence and attention to detail that these people clearly didn't see the need for on the payroll.

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

Lysenko

I won't pretend to understand buckets vs. blooms...

... but otherwise it makes sense. We had an analagous data storage issue and had to compare the aggregate efficiency of computing a fully unique hash for every item vs. a simple CRC32 and brute forcing the occasional collisions. The latter was clearly the most efficient overall.

The same sort of thing happens with many RDBMS indexing strategies. You accept that there will probably be hundreds of "John Smiths" and probably a few dozen born on the same day, so you use indexing to get you 99% of the way to your target and then table scan the rest.

Qualcomm offers concessions to secure NXP Semi takeover

Lysenko

It's also worth mentioning that NXP only just acquired FreeScale, so this is really a three way acquisition/merger.

Hitting 3 nanometers to cost chipmaker TSMC at least US$20 billion

Lysenko

Re: Desirable

Why the UK? We don't even have a military capable of defending the UK,

UK homeland defence has been based on the "Moscow Criterion" for half a century - and you can substitute any other capital city for Moscow. Carriers and so on are for force projection and interference in the affairs of other parts of the world. Actual defence depends on submarines and the same logic applies to Taiwan, so that makes the UK and France the only viable allies if the USA keeps going nativist.

Frustrated Britons struggle to locate their packages: Royal Mail tracker smacked

Lysenko

I like it. Once they've got AI embedded in such systems, unionize them and ask them to vote on strike action. Then we'll see if they've been coded to optimize long term customer satisfaction (which will inevitably be the claim) or short term C suite bonus packages (which will be the reality).

Ghost in Musk's machines: Software bugs' autonomous joy ride

Lysenko

First person?

Joshua Brown became the first confirmed person to die in a crash where the car was, at least in part, driving itself.

I think not. Cars have been "in part, driving themselves" since cruise control and automatic transmission were invented and there have been fatal accidents attributable to such systems. In this case the cruise control might have been able to disengage itself but failed to do so - that's a big step up from incapable of disengaging itself and ignoring manual override or simply autonomously accelerating.