Posts by Lee D

Re: Money

And I can't help but think that if you scrapped business class, shoved everything together, and evened out the legroom to something sensible, you could get MORE passengers on the same plane.

Sure, they wouldn't be paying a thousand pound for a fancy seat, but maybe more people would WANT to fly in them, and the businessmen wouldn't be too upset about having only decent-legroom rather than their own personal box.

P.S. I judge any company that flies their staff business class. If I were an airline, I'd make all my staff fly economy. But then, my economy would be adequate rather than prescribed torture barely above legal limits. Eat your own dog-food.

Re: What difference does it make what wifi router you put on it?

More bandwidth = more simultaneous clients.

There are probably at least 8-9 devices in the average family household connected to wifi. Phones for mum and dad, the kids tablet, dad's laptop, the Xbox, the Wii, the Chromecast, etc. etc. etc.

800Mbps is thus only about 80Mbps per device. That's only 8Mbytes/sec (roughly) for downloading their new 50Gb update to their XBox game, for instance.

Plus... if you buy kit that's capable of the higher speeds, you not only get lower bandwidth, but you're playing nicely with everything else that has those higher speeds (so instead of 10 devices all fighting and taking their percentage of overhead, they know to dial back and the one client needing more gets it). Latency is also a factor, plus that wireless is very noisy (so 1Gbps theoretical means you have more room for errors and retransmissions before you do start to affect the connection).

Don't forget, a 4K stream, sucks up 25Mbps per second (according to Netflix). Say ten houses share 10Gbps. Now you have 1Gbps. That house has 8-10 devices. Each get 100Mbps on average. Add in noise, retransmission, upstream-and-downstream latency, broadcast traffic, interference from neighbouring networks, etc. You're now into less than they'd have got from a local Ethernet cable 20 years ago. Given that we're all using cloud for everything from home photos to school work to streaming movies, it quickly disappears. And the kids downloading the new maps for GTA V could easily knock the connection for six for an hour or more (I've yet to see ANYONE but myself run a router capable of proper QoS and fair airtime sharing, and it's trivially easy to knock someone off their games just by doing a large torrent or similar).

And this is a protocol that you expect to be using 20+ years from now, too. Some hangers-on will still have whatever modems support this sitting in their living room long after everyone else has upgraded to terabit-internet, and they still need to do a half-decent job.

Like the infamous (and apocryphal) quotes... 10Gbps should be enough for anybody. And there's only a need for a handful of computer in the entire world, right?

To be honest, it all depends on your usage. I can run a school with 500 pupils off a 100Mbit leased line. But every member of staff probably has 75-100Mbps connections to their home, with an order of magnitude less devices dangling off it. And I guarantee you that their home connection is like treacle compared to the school one. I bet their iPhone gets a faster individual connection on 4G than on their Wifi too, once all's said and done.

By the time this hardware is out, everyone will be on 5G and, embarrassingly, it'll outperform their home broadband. If they want to be able to keep up, the DOCSIS cable standards, or the latest DSL standard, have to be able to deliver 5G speeds times by the average number of devices in a household (to combat interference from the neighbour's devices too, even if they are idle), after all the losses, conversions, interference, channel-congestion, etc. Or else people will literally stop bothering buying home broadband.

Case in point: I don't have a home broadband connection. I bought a Huawei 4G router, slapped a high-data-usage monthly recurring SIM into it, and I pay less than a BT line would cost, get all the data I need, at speeds twice those quoted for a BT line in my inside-the-M25 new house, with no commitment, no installation, and deals like "Netflix doesn't count towards your data allowance", etc. And I can pick it up and take it round a mate's house because it's about as big as the cheapest of smartphones and lasts 6 hours on battery.

This DOCSIS standard, if anything, doesn't go far enough.

Re: Lets all bash the British company

- Insisting on keeping copper isn't a problem. The problem is that the copper isn't backed by anything at the exchange. VDSL over copper can do amazing things, way more than they actually ever achieve, and nobody is dictating what happens past the cabinet.

- Wholesale cost BB? You mean more than just about every developed country while providing much less in return? BT have had DECADES of opportunity to get their costs down by doing the above.

- "Not insisting VM and altnets also wholesale to encourage demand led installation" - given that almost all that consists of is NTL's bankrupt cable business, and Hull's sole ISP, that's not something you can just regulate. If you did, you'd be forcing OpenReach to take it all on, and providing new installs too. Pretty much the bit that's not happening is new installs, VM rarely cover any area at all that isn't already well covered by BT/OR cabling.

- "not enforcing standards based fibre installation to ensure we see the same standard of fibre installations across the nation permitting access level competition for all installs not just BT wholesale" VM is DOCSIS3. BT chose VDSL. Both are well-established standards. Everything else doesn't matter as it's IP transit... the beauty of IP is nobody cares what box is at either end so long as it talks IP. ISPs haven't had to have banks of modems, one per active customer, for decades.

"At least they pay uk taxes and are accountable to uk rules and regs, which is not the case with VM owned by US giant Liberty Media off the back of debt acquired by the smaller cable operators who pay(ed) little to no tax due to paying interest and repayments on previous loans which may now not be in the UK anymore thereby dodging taxes on profits the loan company (LM subsidiary or close related business) may make."

They bought up businesses with the only asset being a cable in the road and enormous debt. Pretty much penalising them because they're not English is about the worst argument you can make (where are all the British-operated companies putting in new connections or buying up the NTL stock? Oh, that's right, nowhere...)

I'm far from a nationalist (I think it's silly to blindly cheerlead for an entire nation just because you were an accident of its geography). But all I see is a shitty long-term, ex-nationalised UK company messing up the entire industry for everyone via monopolistic practices, and the only serious rival being foreign. I know what that tells me about my nation.

And I've dealt with more than enough "highly skilled" installation, civils and management people for BT/OR to know that I'd rather they were on the dole queue and having to prove their utility rather than the shitshow we get at the moment.

You want a fix? Nationalise Openreach. Make the infrastructure government-paid-for, force it to provide a line upon request within a year wherever the person is in the country, charge a standard rate to everyone to cover the long-term costs of everything. ISPs then provide the IP service at the other end. Like it used to be... get a telephone line from a government department, but who you pay for phone calls is up to you.

No need to standardise on one tech. No need to scrap copper. No need to steal the business of the only rival who has their own infrastructure, or pay off their debts for them. Unless, of course, they go somewhere you don't (vanishingly rare) in which case you might want to buy a deal with them rather than them have to buy a deal with you.

And every time you build / repair a road or pavement... put a tube full of fibres along it and a cabinet at every junction point.

Increased internet costs for everyone for 5 years. Then everyone with access to broadband. Then a highly-redundant and high-capacity network to back it all. Then no "separate line rental". Then no BT install farce. Then you can call it a utility like any other.

Apple are a designer brand.

Note that this has nothing to do with actual "design" (don't get me started down that path with Apple products).

People pay to have the Apple name. Same as Beats headphones, Nike trainers or Calvin Klein underwear.

I'm under no illusion that that is what's happening, and neither (I suspect) are most Apple buyers. But it's something that would drive me insane if a company that I actually used did so.

But, to be honest, I have never once cared about designer brands, chosen a brand "just because it's that brand" (though by chance I have a number of Samsung devices - printers, TVs and phones - that I'm very happy with, none of them big-name items or expensive smart things) or any celebrity endorsement, or even "what other people have".

Re: Rule One...

Nope.

Rule #1: START polite and reasonable.

When you then don't get a reasonable response, you escalate accordingly.

Each new person that it gets escalated to, you give the benefit of doubt, explain again, start over.

But, to be honest, sometimes there is no alternative but to get arsey with people when they can't fulfill a reasonable request.

Cases in point:

- IKEA: "No, we can't refund your faulty product and the remaining unopened products of the same type, just one day after you bought them from this store because we don't stock that model any more". Let me assure you, you will. I had to assemble an entire unit inside the complaints department, shoving customers out of the way to do so, to get the right attention for this one.

- Apple: "Sorry, we won't allow you to make more than 5 Apple ID's a day, despite you having 500 iPads on-site and us having this conversation for four years in a row and always allowing it before, no we won't put that in any form of writing (e.g. email), oh, by the way I'm the 'Head of written complaints'"

- Three: "We're going to charge you for a phone that never arrived, that you phoned up to block because it never arrived, that the contract for it was in the same parcel so you could never have signed, and then when you cancel the DD via the bank because we refused to do so, we'll threaten you with a lawsuit and then phone you 30 times in one day to chase you for the money" - Strangely, I've been threatened with no end of lawsuits and even offered to initiate a few on their behalfs, and never once appeared in court or lost the argument...

Re: I'm missing...

If someone can figure out how to turn it into a viable attack before Microsoft can fix it, it doesn't really matter whether they privately disclose it first or not. Someone, somewhere will have found it and be attacking it, or will be paying VERY close attention to the CVE entries etc. the second they are registered and poking around anything they think related.

A whitehat holding onto a critical exploit is no better than a blackhat doing the same, the risk is just the same, and there's no way to tell which they are (or indeed whether they are both... taking the glory for discovering/fixing it while secretly being the guy selling it to the baddies and profiting both ways).

Fact is, you need admin rights anyway. Already game over.

Re: Maybe

Agreed on the net neutrality thing but on a mobile network, I think you'll find it's always been there. For a long time, anything like Skype etc. was dialled down compared to other access, presumably because it's competitive but you could reasonably argue that the real-time requirement isn't suited to what should be "data sent in the gaps" between their own real-time data (i.e. phone calls).

Nowadays, not so much of that goes on, but if I was unhappy I could move to any other provider in seconds. Grab a SIM, slap it in. However, data usage limits are growing all the time (even if you exclude the "unlimited" offers) so it's not really a huge issue at the moment, I feel. Three do a package where only TVPlayer and Netflix are excluded - I can quite easily guess which company most people would go with, whichever one bundles the most. Nobody is going to just subscribe to a service purely on the basis that their ISP doesn't count it against them... it's the other way round, especially on mobile. They'll just move their number to a provider that doesn't penalise them for preferring a particular service.

Re: dodgy parking companies

For all the fuss people make about parking, I have literally never had a problem.

Mainly because, if they have stupid rules or payment systems, I just don't park there. I'll literally park miles away and walk instead, or just not bother.

I'm sure that this attitude achieves only one of two things a) the landowner gets what they intended and I don't park there, b) the landowner loses out on potential revenue because they're overbearing and I don't park there.

Neither of which I really want to change.

For instance, in any of the towns I go to, I have a nominated car park that I always park in because they are the one that don't have stupid junk. Ticket-in, ticket-out ones are normally the best, because then I have a timed, dated, stamped, paid receipt (and now that the TITO ones also auto-recognise your number plate you get to keep the receipt more often than not). Pay-and-display are open to interpretation, abuse, etc. but with a TITO one you can't argue about when I arrived, when I left, and you can't go around sniffing every car to see if you can pull it for being a second over because the ticket isn't even on the car anyway. There's almost no point patrolling a TITO car park, except to find abandoned cars (which will flag immediately on your database anyway after 24 hours, so you know exactly what you're looking for).

Plus, I have a dashcam for this purpose too. I even like to park near the signs if I'm forced to use a pay-and-display so that I can get those on camera too.

To be honest, though I don't doubt there are unscrupulous people, those are easily combated by the vaguest use of a personal record of any kind. Everything else, like speeding fines, is an idiot tax on people who park where they know they aren't supposed to. I still chuckle every time I see someone arguing with a traffic warden despite clearly being in the bus lane / on the double-yellows / etc.

Re: C64 Joysticks

Daley Thompson's Decathlon cost me two Sinclair Interface Two's, two edge connector repairs and no-end of joysticks.

Damn game.

And only because it was the closest I ever got to playing Hyper Sports, which I used to coo over because it looked so fun (hint: still is, but the weightlifting event is just RIDICULOUSLY difficult).

I've worked in plenty of places were people were let go for being off-sick long-term.

If you're off-sick it means you're not fit for work.

If you're not fit-for-work long-term, then it means you shouldn't be doing that kind of work and they shouldn't be insisting you do.

Just because people run scared of actually acting doesn't mean you don't have an easy and obvious legal recourse to say bye to them. You just have to make sure you're not doing it out of spite and that you give them reasonable recourse in case everything they have ever told you about why they were off happens to be true, and maybe even that there's more that they don't want to share with you for personal reasons.

Of course, this leaves some scope for abuse, but you can do precisely nothing about that. In the same way that you don't know if someone returning a product actually genuinely couldn't get it working or whether they just want their money back because they overspent at Christmas. It's all part of the cost of doing business.

More worrying is that, if one person is off-sick, it means your office/department comes to a grinding halt. That's not a well-planned business.

Hint: I basically never have a day off work. I think my last sick-leave was about 8-10 years ago. Before that, it would be another 8-10 years. So if anyone was going to be upset about people malingering with illness, it should be me.

But as the guy says - not my business, so I don't care. If someone gets more money than me, more holiday than me, more responsibility than me, a better job title than me, more days off than me, or whatever... good luck to them. If I think I'd getting a duff deal overall, I'll be sure to speak up, don't worry. But if they've negotiated a better deal, kudos to them. It doesn't mean I MUST be given exactly the same, as I have no knowledge of the differences in situations.

This is my biggest problem with unionisation, by the way, the concept that everyone is equal which means they all get top wages for doing the lowest-common-denominator of work. I've never been a member of a union, as I negotiate my own salary and conditions. Strangely, that means I'm often under onerous NDA's with my employers for getting a better deal than anyone else (and as part of my job I see, manage and have to manipulate salary etc. data so I know I'm not being fibbed to but equally know that even if I found out something that way, it's none of my business as I was only processing that data and couldn't act upon it).

I think the childish jealousy of "he gets paid more", "she works less hours", "why doesn't HE have to do X", etc. is what tears apart a team much more than any disparity. You know what? The people I see who get the biggest raises etc. more often than not work harder for it, suck up more for it, try harder to get it, and even ASK for it where the other people don't. If you don't ask, you don't get. Feeling hard done by? Ask for a raise. The people I see who get stuck on a wage structure for decades are the ones who "never want to make a fuss".

Re: Endurance != Reliability?

That's my point...

Increasing speed is pointless.

Increasing durability is pointless.

We need them to do no more than produce a standard, small module, in massive quantities, and bring the price down. They can sell that module singly in a small drive, or thousands together in a humungous one. But until the cost per module comes down, they aren't practical and wasting money on FASTER controllers/chips isn't helping.

I'd be more than happy with whatever chip is in that several-year-old SSD, multiplied up to fill the box, and sold at a decent price. A 1TB Samsung cost me £300 several years ago. By the same token, a 15Tb one should cost £4500 max. In actual fact, it costs £7000 ($10k).

If they can't get the modularity, mass production, and scale correct, we are never going to get affordable SSDs. And I'd be happy for them to abandon HDD production to do so.

Re: The Nasty Little Truth About Deep Learning

Machine learning of a trained network tends to have a logarithmic pattern - it learns quickly at first, then quickly plateaus and it takes a lot to "untrain" it onto something else.

This is why most of these "AI" things peak with basic functionality, because after 100 trainings it might get the idea, but between 100,000 and 1,000,000 trainings it improves very little indeed. And it also becomes MUCH harder at that point to change what it was trained on... because it may have reinforced the wrong parameters a million times and you can only feed it a handful of corrections.

This is why Google doesn't have just one massive AI that they use to do all their AI jobs (e.g. "Viki"). They start fresh each time and retrain only on what they want it to know. Because when the plateau strikes, it's no longer any fun to beat your head against a brick wall. Their Go robot loses at poker, their poker robot loses at Go.

It's also the reason that PhD students in the area can operate - train a model, get it to do something interesting, realise that you can't make it do any more, write up paper, flee for some high-paid job.

Anything sold to you as "AI" today is lying. It's not even close. It's just a huge statistical model with heuristics to tune it to what you want it to do. It's not intelligent in any way, it's just seeking statistical similarities with its training material. The more training material, the slower, harder and less reliable a particular result will be (e.g. train it to see bananas and apples and it will start to classify things in the wrong group, as opposed to just training it to see bananas and saying yes/no). And the best bit - being "AI" you have absolutely no idea what criteria it's judging on. You train it, sure, but is it just looking for "image is mostly yellow" or "image has mostly yellow in the middle" or "image has a curve" or what? You have no idea the hidden criteria it's associating with the image of the bananas you're training it on. Which means you have no idea how it will react to any one image, that you have to counter-train it (i.e. give it lots of things that are not bananas), and you will also find it very difficult to modify its behaviour later on if it turns out to not be looking for what you think.

Pretty much, there's not much difference between what people are pushing as "AI" and a Bayesian spam filter. Sure, they're useful. But they are far from reliable or predictable. And at the end of the day it takes a human to feed it enough data (not just emails but "This was spam", "This wasn't spam") to actually get close to useful, and then it can be easily undone by anything it's not encountered before.

That's a worrying facet for a machine that's driving your car in the real world. Pretty much if a UFO were to park itself on the M25, people would still recognise it as a hazard and know how to stop their cars safely. "AI" like this won't necessarily, and you have absolutely no way to tell what it'll do until the day it happens.

They know that already.

They all have iPads or smartphones. By comparison a Raspberry Pi is a toy. P.S. you think they care that they can use a machine, if it can't play GTA V and Overwatch?

I work in IT in schools, always have. RPi are a waste of time. MicroBits are even worse. Pretty much everything they want to do, they know they can do on an iPad, Android phone, or portable PC. They already don't care about OS or architecture.

What they do know, though, is that you need a decent PC if you want to do anything serious, especially 3D or video editing, and everything else is so damn powerful that a RPi 3 is just a joke. They might like plugging it in, that's about it. It's not a wonder-toy to a kid that has even the most basic Samsung Galaxy (which will be smaller, faster, have integrated 3D, run on Android, and probably is in their pocket). And yet, you still need monitors to plug into, power supplies, kits of parts, keyboards, etc. And a space to do it in. Gosh, if only every school had, like, a suite in which they had all that kit and workspace already? You have to shove the PC out of the way to play with the RPi or add in so many other parts you could buy them all an iPad / Chromebook each anyway.

Guess what app they all try to install on their iPads? Word. We don't even give them Office 365 but they all expect to open things in Word by default even on non-MS platforms. They are completely platform-agnostic, but they "know" they need Word.

So, sorry, but the RPi's aren't all that interesting. And staff don't know how to utilise them. And apart from a little computing lesson where you learn to plug it all in, everything they do is better off done on a real computer.

If you go to BETT, they were pushing MicroBits one year. Literally no products - just pushing what they could do but you couldn't buy them. And a few years before it was RPi. No lesson plans, no teacher assistance, just boxes of gadgets. Both are now almost invisible, like the 3D printers before them, visualisers before that, etc.

This stuff isn't for education, which has entirely different priorities to you and I (and I work in private education with kids who go to Eton... they literally lose all interest in RPi etc. within minutes but will happily build their own drone aircraft). Coding? Yeah, they "did that". Whether it was Scratch or Python it consisted precisely zip of their lessons over even their primary education and then that was done. Why? Teachers who can't code and who have huge curricula which includes a lot more than coding. Or playing about with little gadgets.

To be honest, you could run a school without visible Windows. I have seen one or two attempt it, reverted one (reluctantly being a massive Linux fan, open-source programmer, etc. myself but they really messed it up), and I tell you that you could run the kid-side easily on anything you liked. Google GSuite for Education, a handful of third-party website subscriptions, any decent browser and that's 99% of what you need for the kids to get through all they ever need to - including testing and assessment. Guess how many go that route? Very, very, very few. Why? Teachers ask for Windows and Office. Why? 10 years ago, those teachers were the same kids with the teacher who was baffled by Windows 7 and who would have stayed on XP forever. Trace it back enough and you still see things like "Word is the word-processor, Excel is the spreadsheet" as if nothing else exists.

Sorry, but education is a market based on teaching things that you're told to teach, when you were never taught them yourself, so they stick exactly to what they know and can pick up quickly themselves. Just try explaining app vs website to most teachers and you could be there for hours. Especially when you then demonstrate that "the iPad app we must have" won't run on Android / Chromebook / Windows.

Education doesn't care about your RPi's precisely because - as I warned at the time - RPi doesn't care about the teachers. Look around and only OTHER TEACHERS provide resources to use for them linked with the UK national curriculum in any way. And that curriculum changed smack-bang as RPi came out. Teacher won't touch that without paying £1000 for the "pack" of RPi's with massive book of lesson plans and a 3 hour course for their entire staff with a support line the other end. That, pretty much, doesn't exist.

Until then, every school you see will be paying for Microsoft licensing annually (but we pay for one copy of Windows/Office per full time teaching employee only anyway), still buying iPads AND Chromebooks AND PCs AND other stuff, and then still complaining Word can't run on a RPi. And your kid will come out thinking half-a-dozen copy-pasted lines of Python is "programming".

P.S. I was also an early tester for RPi 1.0... including diagnosing the Ethernet/SD card USB bus saturation issues with Broadcom directly. My RPi is gathering dust in the attic, as are the school ones at every school I've worked at since they came out.

Re: Threatened the ICO ?

"or could reasonably be expected to be aware of it "

Such as, for instance, ensuring that you run industry-standard software to stop unauthorised devices on the authorised machines.

"Reasonable" in terms of data protection has included - in case law - things such as reasonable preventative measures to ensure compliance with your verbal "don't do that"'s. Saying "I told them they should have a password" just doesn't pass muster any more. You have to show that you've enforced that and are aware of those exceptions. To not do so is negligent in your data handling duties.

It's also been a factor that you can say "we don't allow that" until the cows come home - but the courts only consider it reasonable if you're also CHECKING that it's not possible, and that people aren't doing it. You can only do that by putting in, for example, device and data control systems. Courts deem that to be the "reasonable" measure, not "Oh, well, it's Sheila, we did tell her".

The fact is - this is all a consequence of DPA case law, where the definition of reasonable has been decided by a judge but not written back into law. GDPR is an attempt to codify that case-law back into actual words.

Hint: An NHS trust was fined for NOT BEING ABLE to prove that a lost disk had been encrypted before it left the building. Not that it WASN'T encrypted. Not that it wouldn't have been expected to be encrypted. But that they couldn't definitively prove that it WAS encrypted BEFORE it was posted and then lost. Case law is not on the side of liberal interpretations of "reasonable" here. Even *potential* for someone *unauthorised* (i.e. not necessary for their job) to see any amount of personal data that they don't need to see as part of their job, can be interpreted as a breach. i.e. that there was even a brief window of opportunity for Fred Bloggs who works for the company to have BEEN ABLE to log into something that might have given him more info than was strictly required for his job? Fineable offence, including personal liability of whoever facilitated that.

You can scream "but nobody ever would prosecute for something so minor" until you're blue in the face, because that's not how the courts are interpreting it.

Take an example: Some minimum wage phone operator sells on your customer list to a rival before they leave. It's STILL a breach of the DPA, even if you told them not to do that, even if their doing that was a breach of everything in question, and even if they were authorised access to those records as part of their job. You will still be fined, as a company, for a) it happening, b) allowing it to happen without a reasonable safeguard against it. It really doesn't matter what THEY do, which is the essence of the whole problem. They just shouldn't have access to anything they could do that with, or be able to splat that information about willy-nilly and you need to show reasonable attempts to control that data (which doesn't wash if you just say "Oh, well, they had an Excel of every email address"... the next question the court asks is "Why?" and "How did they get that?")

It seems incredibly like imposing further conditions on the distribution, which is prohibited under the GPLv2: "You may not impose any further restrictions on the recipients' exercise of the rights granted herein."

It seems quite clear to me that making people pay for the software, and then denying them future updates in perpetuity should they EVER exercise any of their distribution rights under the GPLv2, is quite a bit more than "imposing further restrictions".

That's pretty moot, however, because you'd have to be an idiot to want to do business with this guy at all anyway.

Re: I always thought Microsoft would be the one to get them...

I used to work for a school that was taken over to become an academy.

In the process, they wanted to merge sites, ditch half the IT staff, etc. But not before they'd forced us into IT service agreements that would benefit the "superhead" and his golf-chums into perpetuity, by selling us everything from cabling and networking to software and hardware.

One of the products they wanted to push was LightApp (I believe it's dead now). They exhibited us at BETT using it, but actually we already vetoed it and refused to touch it. It was a thin-client solution based on pushing X-Windows sessions into thin-clients, and then replacing everything on the backend (i.e. the IT team) with a remote server managed by the company in some god-forsaken third-world country. We vetoed it on many grounds, everything from "no local support" to "we don't have an internet connection reliable enough" to "data protection issues" to "security issues". Bear in mind they wanted hundreds of students to use those thin-client / remote-sessions for EVERYTHING they did, plus all the school admin, etc. It was just laughable.

They allowed me to trial it as a pupil so that I could voice concerns and they could answer them. So I logged in via their thin-client, got full root access in a matter of seconds (no security at all, they just assumed you'd never look in their chmod 777'd folders for all the admin users), and left a document on their desktop detailing my objections.

One* of those was: They sucked out the icons from MS Office and used them as icons for OpenOffice/Libreoffice (I think it was OO at the time, I can't remember), with Word, Excel etc. as the names. Prima facie trademark infringement.

Needless to say, at the time it was the least of my worries, and the least of theirs trying to sell us such a junk piece of system, and they never saw a penny of it. I left soon after and I've never heard of them since.

(*) Best one, though, was that they promised us it would "run any Windows program". I don't think they knew that I was a Linux programmer and so could understand what garbage that was - at the time, WINE was barely viable for an old version of Office, let alone anything else, and virtual machines weren't heard of in Windows circles.

As part of this, we had "Ranger Suite" (since bought up by RM, so that's dead too), which is a Windows GPO deployment / user control program that shuts down rogue processes, forces the desktop settings, reports violations, allows screen-based remote control, etc. etc. and creates and manages users in AD. It was basically THE front-end security on a Windows machine. They said it would run under WINE and do everything it always did. I nearly died with laughter at the suggestion, and the salesman ran from the room and ran crying to the head saying I was being unprofessional. My boss then countered saying that the salesman is the one talking rubbish and didn't have an answer when proven that it would NEVER work (I doubt you could run that software now under WINE, it's so heavily AD/GPO/Registry/task-hook based), so nothing happened and we never saw him again.

Two weeks later, the guys in charge of trying to move us to this setup offered me £600 a day to go around their other schools and help them sell it, on the basis of "he's smart, but lots of money should be enough to let us use that smartness against our other clients", I think. As my boss correctly predicted I would tell them at the time, and how I re-iterated when asked, "there wasn't enough money in the world that would make me lie and con schools out of money for a living".

But it's funny that 15+ years later, people are still pulling the same tricks with no knowledge of how to do business.

Re: Shame they pay no tax elsewhere

America has always double-taxed things. They basically don't care what the outside world taxed you, they will tax you too.

Many dual-citizenship people find this out - pay US taxes while living abroad, or give up the US status. It's a very common dilemma. Even if you use certain credits to not have to pay the US tax, you still have to fill out all the US tax forms to claim that even if you haven't lived there for decades. I don't know of another country that does that.

If is was "just a matter of feeding in more data", Google would have the world's best AI running across their datacenter already.

Sadly, it's not that simple. "AI" as you know it at the moment is just the same as it ever was... progress in the field is limited and has been allowed mainly because of commodity hardware but what they've found is that - though they can throw much more parallel, much faster, much more powerful, much more prevelant, much cheaper hardware at it - it doesn't change the fundamental nature of what it is: A statistical model.

Statistical models are not "intelligent", they don't "learn" as you expect. Quite often 99% of the gain is in the first 10% of the training and then very little else changes and it takes much longer to "untrain" it in order to show it exceptions that it had never seen before. And, at the end of the day, nobody is quite sure what it's trained itself to at all. It might be statistically correct most of the time, but it's not trained.

If it were just a case of throwing more hardware and time at it (time being much more important, I would posit, literally just training it 24/7 for decades), then we would have a Bitcoin-like economy where companies were fighting to throw as much time and power at a basic AI as they could to be the one with the most well-trained AI. Amazon and Google would lead the entire scientific field. Places like CERN would exist just to train AI en-masse.

But that's not how it works. Or how the technology works. Or how we even believe it could work. All the "AI" you know isn't... it's closer to a heuristically-determined expert system. We've had those since the 1960's, and though computing power has increased by factors of BILLIONS in some circumstances since then, not to mention that's just a single computer and the ability to scale the AI to billions of computers exists, it hasn't really got much better at all.

IT's like saying that the way to train a child is to throw as many books as possible in its direction. Literally bury the poor sod under literature and expect him to be an expert in everything from Shakespeare to quantum mechanics. Kid not smart enough? MORE BOOKS! Kid can't read yet? MORE BOOKS! Kid gets something wrong? MORE BOOKS! Kid biases towards a certain answer? MORE BOOKS!

That's not how it works with real intelligence, and it's certainly not how it works with what passes for AI.

Everything "AI" you ever seen, from Alexa and Siri to artwork-creating robots, Google image detection, whatever you've seen at CES or any other show: It's the same thing. A statistical model, trained on a data stream that, after a very short period of time, has increasingly poor gains for the time/effort/resource/training it requires to add on criteria or more data. Literally they plateau very quickly after becoming vaguely useful, and then progress drops to nothing.

And without the human-led training, they are even worse. I can knock up some Java code - like many of my peers from CS courses did in the 90's - to show you neural nets, genetic algorithms, all kinds of stuff that will demonstrate "learning" behaviour. Right up to the point where you need it to do something slightly complicated. At which point the returns diminish to nothing.

There's a reason that most of the AI in the field lasts precisely the length of a PhD research project and then dies a death - do it, get results, realise that's the best you're ever going to get, write a paper, run away from the entire field.

Re: when in tandem...

Draytek Vigor routers are fabulous. I have the 2860 - firmware updates all the time (with new features as well as bug fixes), certified compatible with BT fibre offerings (and ADSL2, and Ethernet, and 4G). Failover, IPv6, all kinds of internal options including web filters and DNS filters and LDAP authentication and AP isolation and dual-frequency radios (including handover between frequencies for compatible hardware), proper QoS, SIP handling (including analog ports that run over SIP on board), VLANs, and every option under the Sun.

I run my house and a work site off them, they are just solid, fast and so featureful you'll spend your life reading the manual and going "I didn't realise I could do that!"

Re: Confirmation checkbox needed

But... like the infamous "four minute warning", it's 100% completely useless anyway.

By the time that alert was cancelled, any missile would have hit and done whatever damage it couldand basically no-one would have had a chance to do anything about it. Hell, that's assuming they could even get the alert out in time, let alone people actually receiving it, reading it and running immediately for shelter (where?).

A second person isn't exactly difficult to come by if you're working on a missile alert system. I presume that second person's function of late has been to slap the first person who fat-fingered it..

Re: Obvious ploy but...

Strange, then, that they haven't done that for 5 years when it would solve the problem overnight, isn't it?

Article 9 of exactly the convention you state, look at my highlights:

1.The receiving State may at any time and without having to explain its decision, notify the sending State that the head of the mission or any member of the diplomatic staff of the mission is persona non grata or that any other member of the staff of the mission is not acceptable. In any such case, the sending State shall, as appropriate, either recall the person concerned or terminate his functions with the mission. ****A person may be declared non grata or not acceptable before arriving in the territory of the receiving State****.

2.****If the sending State refuses or fails within a reasonable period to carry out its obligations under paragraph 1 of this article, the receiving State may refuse to recognize the person concerned as a member of the mission.****

So long as we declared him persona non grata at some point between then and now he is not, cannot and never will be able to be classed as a diplomatic member who enjoys those rights. No matter what Ecuador says.

It is a bug.

It just might not be a security-critical one.

There's no point having a dialog asking for a password that literally doesn't care what password you put in, whoever you are. Either the dialog shouldn't be appearing, or it should be refusing bad passwords.

This is not "a problem" in this particular context. But it's incredibly telling of the laxity of testing and the code-paths in the secure sections of code that Apple uses - not unlike the bug a few months ago that allowed anyone to get admin by.... doing exactly this... typing in any nonsense twice into a password dialog would let them log in.

What's wrong here is the process... quite what is popping up that password prompt and why does it accept the wrong password WITHOUT showing an error at all? And how many other places / weird combinations allow the same. If this was the only bug, sure, you could chalk it up to some form of coding accident. But this is only another in a worrying trend of "You must authenticate" "Gah, just have admin rights anyway" issues that MacOS has had.

Think not about what the bug is, but what it represents. Somewhere there's a piece of code that literally says "Even if that password fails, carry on regardless, using the admin rights, and don't tell the user". That's not a situation that you want to propogate throughout your OS code.