Posts by I could be a dog really 1209 publicly visible posts • joined 14 Oct 2022
The only thing you've missed out of that is that the moment you start to market anything - you can be sure that the US tech industry will go all Standard Oil, IBM, and use every dirty trick to kill you off. All the big US tech players have shown that they are willing to break the law in search of profits - so expect any customer looking at your products/services to :
• get "nudged" into a realisation that their existing IT might suffer "an accident" if they put any business your way
• get offered exclusive deals, if we don't have 100% of your business, then it's cost you double on your licence fees - long term idiocy, but it'll frighten off pretty well every finance department
• hints that if they use your services then they'll get sued
• and a host of other tactics.
That last one will also be backed up by a barrage of lawfare suits against you - it doesn't matter if you manage to have each of them dismissed after decades in the courts, it'll burn up cash very fast. Think how Apple and Samsung battled in court for years over "rounded corners", over how things scroll, and the like. Every little detail of what you do will be scrutinised to see if it could, if you are blind and rely on a vary partial advisor, look like it infringes on any of your many patents, design patents, registered marks, look and feel, ...
My feeling is that the best way to hit the likes of Microsoft is to make them hurt themselves. AIUI they market their software as supporting open standards (specifically OOXML), but actually they only "sort of" in changing ways. I.e. they make their files incompatible with software correctly implementing OOXML. If "big customers" turn around and demand that there be a "open to inspection" validation suite, and that MS demonstrate full compliance, then that removes the biggest host MS has - that if you take Office documents and work on them in another program, they often get "slightly mangled" in some way. If you could hold their feet to the fire until they make their entire system actually standards compliant, then that removes their biggest stranglehold.
I know "not a lot" about the innards of Windows, AD, etc., but I've been trying to explain to people for quite a few years that it really doesn't matter where MS tell you they are storing your data, how many policies they have in place to protect it, and so on - there's this business of getting a login, which involves a global mesh of systems gaffer tapes together into which anyone with access (i.e. MS and many national government agencies) can insert a step to subvert it and give themselves full access.
MS don't, and cannot, provide a GDPR compliant service to anyone.
Having a degree demonstrates skills in taking exams and writing papers. It does not say anything about the intelligence of the person.
That depends al ot on the degree and the "quality" of the institution conferring it. There may well be degrees where all you really need is skill in taking exams and writing papers - but really, I find it hard to believe that even the best blagger could manage that without having learned at least some of the subject matter.
But for many degrees, mine is in engineering, if you haven't learned the concepts and how to apply them, then you don't stand a chance getting one. As an aside. one of the reasons engineering degrees open up opportunities in so many other fields is because (non-idiots) recognise that engineers have learned how to apply concepts/rules/calculations to new problems - and that makes them better than (say) someone who has an in-depth knowledge of how todays's finance systems work, but might have s.f.a. idea how to cope with next year's tax code changes.
Sorry, I know that sounds a bit "I've got a degree, respect me". It's not, it's pointing out that there are "real" degrees, that really do test your ability to understand and apply concepts.
I think the fundamental problem hinted at is not so much what you describe - that's indeed doable. It's that each vendor has it's own way of doing things, and with Cisco, an endlessly changing number of ways of doing things. So if you are trying to do an API that works across a mixed network then you end up with either a horrendously complex API to cover all the different ways of doing things; or you have an API that severely restricts the features you can use to some subset supported by all vendors/platforms; or more likely you end up with both, severely complex but still massively limited.
If all the indsutry vendors got together and co-operated, then things could be different. But such "co-operation" these days usually comes down to everyone agreeing that there should be one single standard ... as long as it's ours.
Oh yes, obligatory XKCD.
Preferably /not/ all administered via proprietary apps from a Windows instance on one of those very VMs, but yo do meet some people....
Like the people that would have their redundant AD servers on separate hosts - so there's redundancy right ? But then have the host OS dependent on AD, so if you have a power failure, there's no AD until you can boot one of the hosts and hence an AD VM - and you can't do that without having an AD about. I deliberately stayed out of doing Windows stuff, but even I knew that one.
It's where people can trip up - they assume certain things are going to be available at certain times, and then reality points out that their assumptions were wrong. In the situation described above, you risk being in a situation where your whole virtualised network system needs telling how to set itself up, but you've put the system that can tell it on a virtual machine that can't talk to anything before the network is up.
To be fair, I've been there myself, but in a much smaller way. Several work hats ago, some firewalls used DNS entries (who in their right mind would use addresses when there's DNS ?), but I could never get everything to quite work out when we had a cold start - power failures were rare, and the UPS was more IPS :( There was a bit of A depends on B, B depends on C, C depends on D, and D depends on A. So on the few occasions we had to do it, it could be worked around by simply starting hardware up in the right order, restarting a few services that hadn't started properly till we had working DNS, then restarting anything that fell foul of the problem. I'm sure I could have worked it out - but rare event, soon goes out of management's tunnel memory, and other more pressing stuff to deal with.
It's also neither new nor restricted to the computing world.
My late father was an electrical engineer, and used to tell the tale of how the UK had a nationwide blackout in (IIRC, sadly can't find anything about it online) 1948. I don't recall the cause, but for whatever reason, the entire National Grid collapsed. Then the engineers tasked with getting it back up and running found out that one of the assumptions in power station design was "incorrect" - you couldn't rely on a supply from the grid to run all the ancillaries needed in order to get a power station up and running. Doing a "black start" of the whole grid hadn't been considered a realistic scenario - easy to criticise in hindsight, but with this new fangled national grid, there's always be power from elsewhere isn't there ?
Essentially they had to isolate all loads from the grid at main substations, use whatever power was available to get a bit of the grid up, use that to start a power station, bring up more of the grid, bring up more power stations, and as supply got to exceed what was needed for starting power stations, they could then start to reconnect load (i.e. switch the power back on to users.) And to think, all this would have involved dispatching people to physically operate breakers etc., in the dark, and something many younger readers might not comprehend, no mobile phones. There might have been some limited local radio facility, but I guess it would have involved a lot of messages passed via telephone calls - a rudimentary form of packet based messaging :)
It was a massive wake up call, and resulted in a project to retrofit gas-turbine generators to a number of key power stations - just to provide a black start capability - these earned their keep by being excellent quick-start plants for peak lopping, especially that surge in demand when everyone puts the kettle on during (e.g.) the ad-break in Corrie - see TV Pickup which until I looked it up I didn't realise was still as big an issue as it used to be. Younger readers might like to get parents or grand parents to reminisce about the days when we had a choice of 2 TV channels and no recorders, so you watched stuff when it was transmitted or not at all. So a popular program could see maybe half the homes in the country tuned into the same thing, and hence all putting the kettle on at the same time.
P.s. I'm now one of those old farts who comes out with stuff only to get blank looks from the youngsters. Only the other day, I mentioned Blaster Bates while having a drink with a few colleagues - not one of them had even heard of him ! That did make me feel old :(
Well done on missing the point.
At no point did I suggest deleting any emails - quite the opposite as I think it's a s**t idea.
What I suggest is outright REJECTING emails that fail whatever tests you have in place. That's not the same as deleting them, because for any legitimate email (leaving aside mail servers configured by idiots), the sender will immediately get a bounce message back telling then a) that their message didn't arrive, and b) most likely given them some clue as to why. If it's important, and b) doesn't apply, then they can try sending the bounce message to the recipient who can then send it to you to look in the logs - but i.m.e. this is rarely necessary.
With quarantining, I get maybe 10-15 requests a month asking why an email didn't arrive
That's only the ones the recipient knows that they should be getting. They don't know about the messages they didn't know they were supposed to have got and so won't ask about them.
it's amazing how many tickets come up simply because users haven't been trained
On that I think we are 100% in agreement :)
If your users don't know how to use something that you implemented, that's entirely on you...get out of your basement, lair, whatever it is you hide in and start organising training sessions. Your stress levels will reduce, your workload will reduce and most importantly, your attack surface will reduce.
Agreed, but why would anyone build a system that's so obviously a pile of s**t by design ? I don't. This isn't about what I do, it's about the absolutely s**t by design mail service provided by just about every online mail provider - and especially the big ones. For a previous employer, I successfully ran a reliable mail service for clients which works just as I describe - it either delivers mail into the users inbox, or it REJECTS. It's not hard. I did have a "technical problem" once that resulted in the loss of some emails, but because this was the exception rather than the norm, I was able to easily identify them and notify the users concerned.
NO, no, and thrice no.
Deliver it if you think it's ham, reject it (i.e. don't accept it) if you think it's spam. It's really that simple.
The message either arrives in the recipients inbox, or the sender gets a message from their server to say it couldn't be delivered.
Quarantining stuff is a f'in stupid idea. Users either never look in their junk/spam mailbox, or they just go and delete everything unread. Either way, it is part fo the problem - mail was sent, recipient didn't see it, sender has no way of knowing that it wasn't seen.
Microsoft are more than willing to do anything which is not outright illegal in order to ...
Would Microsoft also be willing to bend or break the law? Maybe history gets to be the judge...
Err, that's "a few decades" late. They've been found guilty more than once in courts of breaking the law - and the punishments were ... little more than a slap on the wrist and an instruction not to do it again.
Certainly back in the 80s, a common trick was to simply rip off stuff, and rely on the legal process being so expensive that anyone fighting back would just run out of money before MS had run out of legal options to delay. Like the in the current case about licence reselling, the playbook goes: lose in 1st court, appeal then lose in appeals court but you've dragged things out a year or two as well as adding £lots to the costs the plaintiff has to consider they might have to pay, appeal to the next court and drag things out some more, ... until you run out of courts to appeal to.
That can drag things out for a decade or more if they play it right, and no small business can survive with no income (MS have just killed their market remember) and having to pay all their own costs in the hope that they'll get it all back eventually.
I'd suggest that what's needed is even simpler.
If large customers, i.e. government scale, turned round and insisted that any standard used must not only be "open" (in a way that OOXML is not), but that there needs to be a (documented) validation suite against which any product can be tested. In producing such a validation suite, MS would be forced to admit to it's tricks and expose the "real" OOXML - or it's own products would fail and be locked out.
Just forcing MS to use OOXML in the open way it's claimed would level the playing field. TBH, I'd be happy is OOXML got nuked from orbit and ODF adopted - but I'd take a really open OOXML as second best (even though it is a pile of manure.)
So, the challenge is to get large customers to turn round and tell MS to demonstrate 100% compliance with the published OOXML ...
Mind you, the much, much, MUCH bigger sin committed by MS (and all the other megacorp manure heaps offered as email systems) is that they will accept mail, tell the sender they've accepted it for delivery - and then not deliver it. Thee is absolutely no excuse whatsoever for ever accepting an email and not delivering it - OK, I'll allow for bugs and technical failures, but by design, no f'in way.
If you aren't going to deliver it, don't accept it "<"period">"
Icon: What I'd like to do to all those megacorps that actively make the spam problem worse. Because unreliable email is worse than spam. Not getting an email and the sender thinking you've got it is worse than spam.
And it's not like there haven't been alternatives to SMTP proposed which would in practical terms kill much spam simply by changing the economics. True, massive upheaval in changing - but given the manure forced on the internet by the big megacorps, it could have been done and settled a decade ago had they actually wanted to improve things rather than use spam as a sales tool.
Online, BT OpenRetch have a section for developers. Included is information on providing your own cabling from CSP (customer splice point) to ONT (optical network terminator). See here, open "How to build a fibre network" and review pages 26 onwards and page 38.
I.e., you do not have to get OR to supply it, and you can install it yourself.
So work out what route will work for you to get from the back of the house to where you want the service terminating. The CSP can be inside, and they are quite small now. So you could have them run the fly wire into (say) the back bedroom and put the CSP in the corner where you hide it behind some drawers or something.
So basically, pull in the internal cable to go wherever you want, leave excess length where you want them to put the CSP, have the kettle on and biccies ready when the tech arrives.
And don't forget that your router does not have to be next to the ONT - you can use a Cat5e or Cat6 cable to take the ethernet connection anywhere in the house (length limit is (IIRC) 90m for Cat6 which should be adequate for even the biggest of houses ! My ONT is in the attic (fully boarded with fixed ladder) and the incoming service goes downstairs to my "den" as one VLAN on a trunked connection (I only had one ethernet cable available at the time.)
A trick with both internal and external corners is a double bend. One bend takes you vertical, then a second one goes back to horizontal at a different level.
Internally, you run along the wall, turn up/down so you run right in the corner, then turn again to run along the other wall. Externally it's a bit more tricky as you'd end up with a section of cable running on the outside fo the corner and potentially rather vulnerable.
Interestingly enough I have zero reason to doubt your story - in fact, it is no surprise at all.
In our case, block of 4 purpose built flats, 2 stories, internal routing for cables. I attempted to contact the MDU team at least 4 times - only to be ignored or "actively ignored" (i.e. they responded, but only actively not engage.) At the time I was acting on behalf of the freeholder/director of the management company - I'm one of the flat leaseholders.
And lo, some shitty washing lines appear on the front of the building, without a wayleave, without permission. I'm now director of the management company, having taken over from the original developers. So far I'm up to 4 different departments who've developed sloped shoulders and basically just told me "not our department, try ..."
I''ve put very clear notices in the pit and the CSP that's got a spare way in it making it very clear that the apparatus does not have a wayleave and is to be removed - I'm rather hoping that if the leasehodler of the 4th flat orders a service, the engineer will do as the notice states and contact me.
I'm also getting really tempted to send a recorded delivery letter to OR's registered company address giving then 28 days to arrange an agreeable way forward - or I'll just remove the unauthorised apparatus and bill them for repairs.
Bloody Terrible OpenRetch
You can provide a battery backup and make the service reliable. Either a UPS into which you plug the supplied power adapters for NTE and router, or if you're a bit technically savvy, supply 12V direct from a battery and use a float charger of sufficient capacity to do the AC-DC conversion all the time.
But a basic small UPS doesn't cost much to buy.
why a business would put every bit of its critical infrastructure in one basket
Because it's convenient, and the salesperson won't have discussed any of the risks, will probably have assured them that it's all legal and GDPR compliant because "we'll specify an EU data centre", and ... most small businesses are run by people who won't know to ask awkward questions. For larger businesses with people who do know, then it usually comes down to "everyone we need to deal with uses M365, so we need to use it or none of our documents will stay formatted properly".
Indeed, the best way to make sure a project doesn't succeed is to put off starting it because it will take a long time/take a lot of effort/need new skills/whatever. And the longer you leave it, the harder it becomes to escape - that's by design, MS have spent a couple of decades steadily making it a bigger and bigger challenge to escape.
I suppose one way to get things started would be to require all documents to be in a truly open format - specifying that "open" means fully documented, available to all on a FRAND basis, and with a test/validation suite available (that should stop OOXML in it's tracks !)
And with the sort of numbers being bandied about, if all of us (i.e. all the countries in Europe and round about) chipped in a fraction of that in supporting/sponsoring alternatives, that would quickly raise the level of what's available. At present, by design, the M365 ecosystem covers a lot of (apparently) integrated functions - some of them without open alternatives, most without any form of integration resembling what M365 users take for granted.
and going "UK seller" which they probably aren't.
And that's something that really pees me off about certain online marketplaces - I usually try and buy from a UK based seller, but do they have a filter for that ? Do they f***. And you see one with "UK SELLER" emblazoned across their images - then drill down and find they are based in China. And you want to report that as fraud - but can you find any means to do that ? Can you f***.
As others have said, these marketplaces d.g.a.s. about things like safety - they are only interested in maximising their profits.
It may well be compliant with the SIO standard - the standard explicitly includes undocumented BLOBs (such as "Document in Word97 format".) I recall at the time reading the BSI review of the proposed standard - it was really scathing about how rubbish it was. So it got passed by the simple process of putting enough shills into all the national standards bodies' committees to vote it through - and then they never turned up again, leaving those committees unable to function for ages.
There is a story that in the legal shenanigans around SMB, where Microsoft were basically forced by the court to provide the specification for SMB to anyone on a FRAND basis, they had to go to the SAMBA project (who they'd been trying to shut down) because the SAMBA project had better API documentation than MS did. For MS, it was a case of "the code is the documentation, what the code does is the standard"; while the SAMBA project had been forced to reverse engineer the protocols/APIs.
I can well believe that much of the Windows ecosystem is poorly documented, and relies heavily of "what the ancient code does".
My recollection could be a bit hazy, but I believe the story went along these lines :
IBM was big in business - all those mainframes and terminals, etc., etc. Desktop PCs were starting to appear, and businesses (especially smaller ones) were starting to see their value. But even for the big sites where "you never got fired for buying IBM", they were getting asked about desktop computers - and they had no answer.
The IBM mindset was "everything big", programmer performance was measured on KLOCs (kilo lines of code), and their internal teams just couldn't create a desktop computer that was a viable proposition.
Meanwhile, National Semiconductor had come out with this nifty little 8088 chip - and had a tech note on how you might build a system around it. Some small startup basically took that, and made it into a real system. IBM saw this, realised what it meant, and bought the whole company - and the desktop computer hacked together just by implementing the tech note got badged as the "Personal Computer". Of course, because it had the IBM badge on, it sold like hot-cakes into IBM using customers.
Of course, to make it actually usable, it needed an OS. The story about how it could have been DRDOS is legend, but a young chap called Bill Gates showed amazing business ingenuity by selling IBM non-exclusive rights to a specific version of MS-DOS (which was itself just a tarted up version of QDOS which he'd bought the right to for a fixed fee, and which IBM badged PC-DOS). Bill then went and sold the very same MS-DOS to other companies building desktop computers - each one customised to the hardware, this was before that became standardised. Better still, Bill developed MS-DOS and sold the newer versions to other manufacturers, and IBM found out the hard way that they'd been outsmarted and had to negotiate on Bill's terms to get the newer versions.
Back then, MS actually made good tools (prior to the PC/MS-DOS era, Bill had been writing compilers) that fitted an actual need - and thrived based on those tools being reasonably good. But like many others (I don't plan to ever own another Mac), over time they've borged from "sell better tools to do a job people need to do" to "stiff people for all you can get by offering them superficially shiny stuff".
We don't know exactly when this was, but even these days, many networks still do not have that wort of logging. I'd go so far as to say that even today, the majority of small/medium business networks won't have it.
Yes, a network admin with a few brain cells turned on could use the port tap function built into most managed switches to look at the traffic and see what's going on, but other than that, all they'd see would be "lots of traffic".
As an aside, I find the most powerful network diagnostics tool is ... the blinking lights on the network switches. Step into the server room/network closet/random purloined space and look at the lights - often you can narrow down the source of the problem just from that. If nothing else, it narrows down where to start with the technical tools.
Many "IT workers" are not very competent
Hmm, most of us fitted that category at one time, and some of us have not improved all that much [sfx: commentard whistling innocently into the air]
I've killed a few networks in my time, I've also fixed a few networks where others have killed them. I don't consider myself the "expert"* many current/ex colleagues do, but I'd agree that far too many "network engineers" just don't have any diagnostics skills. To be fair, if no-one has ever shown you how to diagnose some of these faults, they can be "interesting" to diagnose and fix :
* "Helpful" user found a loose network cable end, so helpfully plugged it into a spare port on the switch on the desktop. The small switch was there because the building didn't have enough ports, and the loose cable end was there for a laptop user (not common in those days). I got called when the "network expert" couldn't diagnose why nothing worked in the presence of a line-rate broadcast storm.
* A colleague had been experimenting with the mobile hotspot on his phone. I think there must have been a bug, because he'd turned if off, but it was now connected to the office WiFi using the same IP address as the router - so the flapping IP-MAC table entry made network performance flap between terrible and non-existent.
* And then there's manually configuring a server with the same IP address as a production server that nearly everyone in the company relies on (oops, whistles again).
* - and then finding that Windows is utter s**t and even when you've corrected the mistake, won't find the correct device without a reboot. This was back in the Win 3.11 & 95 era.
* And one of my favourites, re-purposing an old router as a network switch - but forgetting to disable the DHCP in it. Thing is, the network will continue to work, for days, weeks, even months - but one day when the on-site server is down, all the clients then realise there's another DHCP server they can get an address from and the network is broken and no-one remembers the "quick fix" for not enough ports.
* And I think this comes into the "there are 10 types of network admin/IT slave" category: you arrive at a site, and someone decided to move stuff around, so they just unplugged everything, you have no documentaiton on how it was previously set up (or supposed to be set up - not always the same), there's at least 2 routers, and you have no passowrds to anything, ...
And I'd guess that finding and reporting that would get someone a slap on the wrist - whereas what should happen is they get a thanks (for finding it before someone with bad intent does) and the relevant team can fix the problem.
you have to ask yourself what that data is worth and who would go through the herculean effort
Exactly, some people do have a rather heightened idea of what they think their data is worth to others.
And of course, there's always the obligatory XKCD
It's been pointed out many times that a) only a very small number of people would have been assessing the tenders, meaning that it's likely that none of them are currently affected; and b) there are quite strict rules on tendering which makes it very difficult to reject a bidder because "we know they have a track record of being crap". Depart from the strict rules, and the bidder can then take legal action for being unfairly excluded.
Also, the number of businesses who can actually take on projects of this size is quite small - so for all we know, they might have been the "least crap" of a very small list of bidders.
Note, I am not even remotely connected to this other than having the luxury of waiting for them to sort it all out - hopefully it will all be working in 3 years time !
Thinking about it from a safety PoV, it should have been a case of power failure causing the gates to unlock - that's the only safe failure state. There's not enough detail to know if that was the case, if it wasn't, and they needed power to unlock, then I'd have put in a RIDDOR.
It's not like there's no precedence for locked gates to cause (or excacerbate) a problem. Take a look at this list of disasters for some oldies :
23 June 1968, River Plate v Boca Juniors
Seventy four people are killed and more than 150 are injured when spectators mistakenly headed towards a closed exit after a first division match in Buenos Aires.
The fans at the front of the melee are crushed to death against the doors by other fans unaware of the closed passageway at the back.
12 March 1988, Kathmandu, Nepal
At least 93 people were killed and 100 more were injured when fans attempted to flee from a hailstorm inside the stadium.
But they could not escape because the stadium doors are locked, causing a fatal crush at the front.
in less than four minutes, with the windy conditions, the fire had engulfed the whole stand. In the panic that ensued, fleeing crowds escaped on to the pitch but others at the back of the stand tried to break down locked exit doors to escape. Many were burnt to death at the turnstile gates, which had been locked after the match had begun
It's events like those that lead to rules about emergency exits not being locked, or locked with easy to open options like push bars. So yes, IMO if the gates needed power to be unlocked, then they were inherently unsafe - and that would make it illegal to stage events there.
We forget that at the time MS were only just moving away from an "anything goes" free for all to something with at least some veneer of trying to apply security & stability. Of course, had they created standards for how stuff should work, and actively enforced them, then things might have worked better. But this was still the days when every developer had their own ideas of how stuff ought to work ...
Ah yes, that artificial accounting division between capex (capital expenditure) and opex (operating expenditure).
If you spend a chunk of capex on new servers, licences, etc., then that comes off the books in one hit and depreciates over whatever period is used (for IT it's generally quite short, probably 3 years.) If you spend eveery month on opex, then a) that comes form a different pot of money even if it's out of the same bank account, and b) it's a "small" bit every month. So you can spend more on rental that you would have done on purchase - but the beancounters think that's better.
The other factor is that these days "nobody cares"* - the market for almost all these ARM based gadgets is people who "open the box, switch it on, use it as the manufacturer intends". There isn't the demand* for open standards so tinkerers can have some form of standard boot environment to help them use stuff as they want. Back then systems just weren't all that integrated - there wasn't the big abstraction layer that meant no program ever has to talk to hardware, and the standard APIs are king. No program running on a modern system has to know about interrupt vectors, I/O addresses, and all that.
* Yes, in here we're a self selected group who are different. But we are barely a rounding error in sales numbers and the manufacturers generally just don't care. In fact, the manufacturers generally don't want people loading different software and some of them go to great lengths to prevent us doing that.
Apart from when you want to run "up to date" software.
At some point, Apple will drop support for whatever you have and no new OS versions. At some point, some software you rely on won't support the old OS you are running. At some point, the pain from that gets to be more than the pain of upgrading, and then you find ...
In my case, Apple make nothing newer than the old (2015) model laptop I'm currently using that is upgradable at all - I have the last model with upgradable storage (which was upgraded, massively), upgradable RAM went some years before that. I've never owned a computer that hasn't been upgraded, and I don't intend to change that.
For my next upgrade, well the pain is going to be more - it won't be Apple, so significant changes to workflows and finding/learning new software, and migrating "many years" of music and photo libraries etc.
So if you are running older Apple hardware, don't run Mac OS on it unless you enjoy running ancient software. Not long ago I had a job to do, and found that my old laptop (the one before my current one, and hence I tend to take around as it's less valuable if it gets damaged) couldn't load the web pages of the system I was trying to configure - the web browsers were just too old.
Some of us remember when 'PC Compatible' PCs were the same - each manufacturer did their own thing, and it shipped with a version of MS DOS which was customised to suit. As I recall, this was accepted practice until someone at one of the PC mags came up with the idea of "if you can grab the install disk for $[game] off the shelf and it'll boot and run off it then it is actually 'PC Compatible'" Only after that did a de-facto standard take hold, which wasn't really designed, more just a case of "what the designers of the original IBM PC happened to do".
Absolutely.
I had an Organiser (the II I think). At the time we were only just getting desktop computers in the office, but I had this one in my pocket. As the article says, revolutionary removable solid state storage (2716 2kbytes EPROM IIRC) and a file system that worked well on the "write once" storage that could only be erased in bulk. I recall writing BASIC programs on it, and of course, using it to remember "stuff".
And it all ran for "ages" off a single PP3 battery.
Cutting edge stuff.
If you are using such parental controls, then you simply restrict access to other sources in your settings. Just the same as if you give your kids free reign to do what they want on an iDevice, they would be free to install anything (regardless of suitability for their age) from the Apple store right now.
Agreed.
It could be "interesting" to see what a world looked like with a genuine split in the root servers - the "fooked by the US" version and the "rest of us" version. I suspect what would happen is that very quickly, people* would hard-code their systems to only use the still working ones. And instructions on how to configure all popular DNS systems would quickly spread.
* People who like myself run their own DNS, ISPs who g.a.s. about their customers, businesses that care about being able to continue working, ditto governments, ...
DNS is controlled by ICANN. If the US ahem "leadership" went mental and did something nutty with DNS overnight, then you lose your national "broadcast" infrastructure and also your ability to update end units (e.g. with hard-coded IPs, etc).
It's not that simple. ICANN might control the root zone, but they don't control the root servers. There's nominally (IIRC) 13 root servers, but I believe at least some of them will be behind an anycast setup so there's more physical servers than 13. Only a few of those are where US authorities can direct them to do something, the rest are well out of touch and the people who run them would simply refuse to load a new (doctored) root zone.
and the OS would only run on one specific model of Dell laptop
That and many other issues (such as "only runs in Win 95") effectively make such software automatically time expire sooner or later. Of course, the manufacturers have no problem with this - by the time anything expires, they no longer have to support it and it's not their problem.
Good description, but there's a detail - and it's about commercial viability.
For a body panel, it would be absolutely possible for someone to take an undamaged vehicle apart and 3D scan the panels - and then produce press tools to make replacements. That's because it's something physical that can't hide hidden functionality like software can. So it's possible to make replacement panels - the question (to which we know the answer) is "is it commercially viable".
Unfortunately, body panels are "almost certainly" outside the realm of DIY.
BUt with the software based systems in modern cars, while you could (probably, mostly) copy and replicate the hardware (if it doesn't use custom chips) - the software is harder. And over the years, manufacturers have deliberately made it harder - and like in other areas, the excuses are "safety" and "security". At one time it was (AIUI) relatively easy to remap ECUs, but over time the manufacturers have put effort into stopping that. Given the relatively small market for "DIY software", and the technical difficulty, it's no surprise that there isn't the vibrant DIY cottage industry in automotive software that there is in other fields.
Why should you or anyone be forced to absorb running costs or disclose IP still used in other products just because one product has been discontinued, is no longer supported?
That's the interpretation the vendors would like you to take - that users are wanting something for nothing.
What users are wanting is fairness - that they don't spend what can be a considerable sum of money for a game, only to find it stops working at some arbitrary time when the vendors decides they don't want to permit it to run any more. As hinted at already, what really needs to happen is that there's a clear statement as to how long it will run for - then people can make a judgement as to whether it's worth paying out for something with that lifetime. Actually, I suspect that there's scope for legal action under existing consumer protection laws if someone had the resources to take that up (it won't be cheap.)
This is exactly the sort of enshittification Cory has warned about - vendors now reserve the right to decide how long you can use something for, regardless of what terms you "bought" it under. And they enforce that by making stuff dependent on some form of licensing or activation service even for stuff that doesn't need it.
One of the problems is that there are few businesses who can take on this scale of project - very very few. And let's face it, none of them have a particularly good track record. It's a bit like trying to choose a security consultant when your only options are Reggie or Ronnie.
