Posts by Ken Hagan 8168 publicly visible posts • joined 14 Jun 2007
"Huang said that the future Volta GPUs would have an aggregate of 1TB/sec (that's bytes, not bits) of bandwidth into and out of the stacked DRAM, and added that this would be enough to pump an entire Blu-Ray DVD through the memory in 1/50th of a second."
You were OK until you mentioned "an" entire Blu-Ray DVD. The big caveat to this massive bandwidth is that a single component only sees a tiny fraction of it. It would be more accurate to say that if you happened to have a DVD already divided into several thousand shards then you could stream each shard past a different (and isolated) processing unit in that fiftieth of a second. Obviously that's less impressive. Actually processing that data as a coherent whole will take longer.
"Leave the "old" .com, .org etc for true multinational organisations."
I'm afraid you've missed the point. I don't want there to *be* any true multinational organisations. When I access a server via DNS lookup, I want to know what laws apply to the server just by examining the TLD. If .com continues to mean "global", I can't do that.
"I was happy when all we had was .com .co.uk .org and .net"
Sadly that only works if everyone agrees to abide by US law, and most people don't have that option. Even within the .com domain, most sites probably aren't exposed to the ire of a US court to any extent beyond losing their domain registration and that only costs them a few dollars. ".com" might as well be ".anon" for legal purposes.
The current crop of gTLDs are, indeed, a complete waste of time. However, the country code TLDs could be turned into something useful if the rules were changed that you could only have foo.{cc} if you set up in {cc} as a legally liable entity with a cash pile appropriate to your parent organisation's size. (Ideally, we'd align IPv6 address ranges on the same lines.) End-users could then assume that anything on *.{cc} obeyed "local law" and that the local entity could be forced by a local court to pay out a sensible amount in damages if they lost a case. Politicians could pass laws applying to anyone with a {cc} domain and have a reasonable chance of enforcing them. End-users could then choose to filter their internet usage by {cc}.
This doesn't even need international agreement, since the registrars for the {cc} domains generally already are within the legal jurisdiction in question. Anyone who doesn't want to live by the new rules is free to "emigrate" by moving their operation to a new domain name.
This would leave .com, .org, etc. to those who want to be part of the US. I'm fine with that. They, in turn, would have to be "fine" with consumers in every other country on Earth preferring to deal with a site under the local {cc}. The big multinationals seem to have those registrations already, so I doubt it would be a problem.
Given that all of us are subject to some legal system, it is frankly amazing that the internet has become so pervasive without any serious effort to partition it into legally coherent sections. All we've had so far are (local) politicians queuing up to call for a (worldwide) ban on stuff they don't like. It's almost as though humanity at the end of the 20th century saw the possiblity of creating something that would by-pass all existing legal restrictions and everyone thought: "Yeah, let's move all our social and commercial lives over there. Who needs laws anyway?".
"Why does anyone care about the operating system used on a mobile telephone?"
No-one does, but the youngsters have started using the word "phone" to refer to a hand-held PC. Usually these have SIMs in them, so they aren't actually not phones, but calling them phones is a bit like calling your car a portable radio.
I, too, have an elderly phone that does what I need, so you're speaking my language. Sadly, we are no longer speaking the vernacular.
Cutting corners (and introducing security holes) saves developers (or their employers) money and gets them to market ahead of their competitors. The costs are borne by the customers. The customers can only move to a better supplier if the better supplier hasn't yet been eliminated from the market (owing to their higher development costs).
Any developer with half a clue can write shit in any language. Bugs have nothing to do with programmers, languages or education. It's all economics.
"companies are largely blackmailed that if they do not buy their brand terms someone else will"
Is this a problem? (*) If I search for an actual brand name, I'm looking for that brand (duh!) and will certainly not be spending money on some time-wasting lying swine who wastes my time with a deliberate red herring. If they were on my short-list (and I'm searching for each one in turn), they will be crossed off it.
(*) Who knows? Almost certainly no-one. What passes for "analysis" in the advertising world wouldn't pass for bog-roll in the hard sciences. It's cargo cult stuff, just a story told by marketing execs to whoever sets their budget.
"Don't MS keep telling us that all the code is new at each release?"
You actually believe that? How many hundred million lines in a Windows release, and you seriously believe that they chuck the whole lot out and start again for the next one?
Earth calling Anonymous Coward: Microsoft's marketing department doesn't always tell the truth.
"I strongly suspect he'd be in for a shock if he ever caught one"
That's generally true. I've seen cats catch fairly large birds, only to regret it a few seconds later. I've heard that a cornered rat is equally unco-operative. I imagine that scaling the whole process up to the size of a deer is quite interesting.
I suspect the green solution to your little poser is simply to increase the amount of habitat available to the deer. After all, we've spent most of the last 10000 years deforesting and generally nicking land from the rest of the ecosystem, so it is not surprising to find that "wildlife" is short of space. Of course, allowing wildlife to increase its resource consumption isn't a sustainable policy, but having got all the people out of the way you can *then* release the bears and wolves.
Nature got along fine for a few billion years before we showed up. It's not actually an unworkable solution, but I suspect most of the human population will reckon that a cull is a better one.
There is no infinitive there to be split. Perhaps you, too, need to "seriously read a book" on grammar before reaching for that particular icon.
On the other hand, the most horrendous abuse of English was in the original article where it talked about the "vocabulary" of gestures in Windows 8. Where's a vomit icon when you need one?
With apologies to those who aren't Windows developers. You won't get the references, but you can probably guess.
Having recently used PROCMON.EXE, I may be a little biased (or bitter) but I suspect that reading in the EXE is a pifflingly small fraction (<1%) of the time spent annoying the end-user with an hourglass ion. Before you even get as far as WinMain(), you have loaded and run the DLL entry points of several dozen system DLLs. For every one of those executables, the kernel will have crawled over the registry to see if various app-compat hooks or debugging hooks are required. If your program ever shows a file dialog box, you'll pull in all the shell DLLs, which rejoice in trawling the registry and file system picking up the current user's preference for just about everything that you can configure in Control Panel. Each of those registry accesses checks per-user and per-machine hives. Depending on how old the app is, every single one of those registry and file system lookups might be virtualised, so each registry access also checks the Wow64Bollocks parallel universe in both hives. File system accesses are similarly virtualised because you can never have too many directories pretending to be System32.
And by the time you've done that and reached the very first instruction of the actual program the end-user wanted, your memory hierarchy is absolutely cache-busted, so everything runs like molasses.
"Remember that he's working exclusively in Linux, so the idea of switching to another keyboard and monitor just to work on one of his other machines is a completely foreign concept to him."
I see the point about X and remote access, but ... really? I don't care how good your laptop is, it will never come close to the quality of the keyboard and screens I have on my desktop. Portability is its only advantage.
I was about to make a snarky remark about this legislator, but it appears that all the El Reg commentards (so far) have missed the point as well, so I'll be kind.
CO2 emission is not in itself a bad thing. There is a natural cycle and as long as you are respiring/burning/carbon that is already part of the current natural environment, you aren't changing the amount of carbon in general circulation. Therefore you won't cause the atmospheric concentration to change. You could burn the whole cyclist and it wouldn't affect global warming.
The problem is in releasing fossilised carbon. That adds to the total carbon in circulation.
"and until it is patched everyone should disable Java in their browser."
The vast majority of users have no need to enable Java in their browser, ever. Any installer or update that re-enables the browser support without getting the user's permission first is IMHO performing an unauthorised modification and is therefore probably in breach of the law in several countries.
You do realise that it would be fitted out with cameras. ("Day 500 in the Big Brother Spaceship...") Also, given the deliberate simplicity of the mission, you'd hardly get any of the credit for the mission. Rather, you'd go down in history as being the most sex-crazed and shameless couple in human history.
I'll pass, thanks, but I might tune in for some of the later episodes.
OK, so I'll make an ungenerous assumption about why *she* cuts out all the nude girls, but that still leaves me wondering why *he* compiles the list of advertisers. (IT angle: given that he can't start compiling the list until the magazine has been made safe, the process is strictly serialised, so they can't even claim it is more efficient this way.)
"For some reason in the US, these people hold views which are taken seriously."
These people would certainly like you to believe that they are taken seriously, but they are probably responsible for the under-performance of the Republicans in recent years. I'm sure Barack Obama reckons Sarah Palin was a gift from God. Like terrorists, there are probably fewer of them than they'd like you to believe.
"And a well planned switch takes time"
My switch from XP-in-a-VM to Win7-in-a-VM took no longer than the time to install the OS and apps, and then to swap over the virtual discs containing my data. I wouldn't call it well-planned -- how else could I have done it.
It was, of course, just one system and one that I was intimately familiar with. But then, that probably describes rather a lot of systems. If I were migrating my company from XP to Win7 then, yes, I would expect to do quite a lot of planning, but the people who currently face that problem have probably already started work and if they haven't then it is probably because they do intend to "run it in total isolation from the Internet".
I'm not holding out much hope for Windows XP Embedded. It is almost exactly the same product and so if MS kept the patches rolling then all XP users would be OK until 2016. However, I think it is far more likely that MS will take the view that embedded systems are not internet-facing and therefore don't need patches.
"the lowest number into which a specific group of numbers divide exactly"
You posted this at 2am GMT, so perhaps you need to get some sleep, but...
It works for me. You have a group of numbers (the factors) and you can consider multiples of each one. The lowest such multiple that is shared by the whole set is the lowest common multiple. The LCM is useful where the original factors are the denominators of a set of fractions, because it can be used as a denominator for all of them. For those of us who dislike (computational) complexity, choosing the minimalist solution (lowest common denominator) is then the clearest indication of excellent mathematical taste.
Any other choice just marks you out as a bit of a wally.
"...refers back to this meaning, not the mathematical one, and encapsulates the idea of a minimal set of features that work everywhere"
Even allowing for substitution of "minimal" where you meant "maximal", you've rather missed the point the author was trying to make.
I don't believe for one minute that the phrase does not have its origins in mathematics. The fact that popular culture got the semantics almost perfectly arse-backwards is not surprising. (Look at quantum leap for another example. Popular culture really *likes* the wrong end of the stick.) The bottom line (sic) is that any competent mathematician would have used "highest common factor", but popular culture rejected this, presumably because it contained the word "highest" which sounds like a good thing, and so they jumped instead for the other phrase they didn't understand from that same maths lesson.
That was sort of my thinking too. The vendor is pitching at institutional customers, not domestic, so we're talking about the sorts of building materials used in larger buildings, not houses, and a purely economic desire for insulation, not an ideological one. Hippies just don't come into it.
Perhaps a Google account is now worth more to its owner.
If you don't take care of your google account, you might compromise all the related services (most obviously docs) that you get from Google, so you make more of an effort. For hotmail and yahoo, there probably are no such related services, so no such consequences, so no such care.
It's still not quite the same as "I paid for this account, so I'll take care of it." but perhaps it is close enough. If so, the corollary is that this isn't going to change soon. Yahoo and Hotmail are not backed by a company willing to provide additional *free* services that might make those accounts equally valuable.
...he is presumably experienced enough to know the likely reaction and he has apparently let the matter rest almost immediately. It sounds a lot like Red Hat management assuming that Microsoft would prefer them to do it this way and therefore leaning on their kernel developers to at least try and persuade Linus. But no-one's heart was really in it. Even Linus seems to have bl**ped out his bad language.
I make no assumptions about the East Pondian experience. I already live in a country where medical insurance has been provided out of general taxation (and so independently of medical condition) since 1946. It ain't perfect, but it's pretty good evidence that you don't *need* to fine tune prices on a per-customer basis if you can spread the risk over enough customers.
I think the real question (given the abundance of different systems that have been tried on the East Side) is why would anyone on the West side of the pond reckon their healthcare system was representative of anything much.
"How can we be sure that this will not affect insurance claims or result in genetic discrimination?"
Oh that's easy. You simply make it illegal to set an insurance premium (or refuse insurance) on the basis of medical information about the applicant. Since we do not pick our genome, it is reasonable to legislate that no-one should be penalised for having a duff one. It's what they did for car insurance and sex, after all.
As a side effect, by reducing the financial value of traceable medical records, you then make it more likely that people will consent to sharing that information with researchers.
"Patenting genes is the first that should go."
Agreed. Everyone knows they are there, everyone knows they are worth researching, and everyone in the field knows the methods for doing that. There is no novelty or inventive step. Only a truly broken patent system would allow patents on such mundane work. Patents should protect the 1% inspiration, not the 99% perspiration. Otherwise it is just an intellectual land-grab with the spoils going to those who are already rich and powerful enough to dominate that sort of process. There is no "benefit to society in return" from granting such people a monopoly.
"that my now-ex girlfriend Silent-Treatmented me into buying"
Is that a euphemism for having her mouth full, or for refusing to speak to you? I'm sure others will want clarification on this point, too. It makes quite a difference to the interpretation of the story.
The majority of Windows licences are either OEM or some sort of volume licence deal. The licence terms for the former prohibit use in a VM because it is not the hardware it was pre-installed on. The licence terms for the latter are typically lax enough that "using up" two licences wouldn't actually be a problem.
Typically, if you move a VM to a new PC, the guest sees the change of processor. Quite possibly it seems a change of motherboard, too. These are the changes that Microsoft are least likely to write off as "typical hardware upgrades". You VM will, however, have the same network card and hard disc (and possibly also graphics adapter). These are the changes the Microsoft are most likely to ignore for licensing purposes.
"Yes, VMs can, but you have to have a license..."
Not only that, but most VM hosts are optimised for the performance of the guest rather than the concealment that there is a VM. Windows is perfectly well aware when I am running it in a VM. If people start habitually using VMs to get around licence restrictions, Microsoft will simply add code to prohibit this.
It will probably turn up as an "update". Leave your machine running one Tuesday and wake up the following morning to discover none of your documents are readable anymore.
"People think voice is easy as even the stupidest person can use voice. The reason it works is because you have a brain doing an immense amount of tricky processing getting it to work."
For an estimate of just how much processing is involved, consider the fate of the novice language student who can understand the teacher (talking slowly and deliberately) but is flummoxed when confronted with a native speaker. Also consider that most experts reckon it actually becomes harder to learn a second language if you leave it too late because most of that tricky processing is actually burned into hardware and if you wait until your teens or later then this option isn't available to you.
And, as you point out, that's just the easy problem of parsing the stream of sound. Actually understanding enough context to resolve ambiguities is Hard(tm).
In fairness, there's nothing wrong with being a crank, as long as you appreciate that you are almost certainly *not* an undiscovered genius. (Equally, if you aren't a crank, you still almost certainly aren't an undiscovered genius. In fact, it's probably even less likely.)
If it isn't "obscure", it isn't on the leading edge. As you say, the "obscure mathematics" of the 1920s is now (after substantial cleaning up and selective editing) entirely mainstream and dished out to several million students per year. (Meanwhile, a lot of analytical mechanics that was totally mainstream in the early part of the century is now only taught in specialist courses. I doubt very many physics graduates have actually used the Hamilton-Jacobi equation in anger. I certainly haven't.) Most of today's obscure mathematics will remain obscure, but eventually some crank will churn out the right runes and six months later it will appear on a T-shirt.
In short, you get a Nobel prize for being a crank and then everyone else copies you.
"Whether or not the story was in error, mere deletion rather than a correction alongside the original would seem more sensible than offering the world a chance to scour up copies of the original story to see what the fuss was all about."
Sorry, but I don't understand that at all. In science, retractions should make at least as much of a splash as the original. Nature have presumably taken the view that no-one at all viewed the original article and therefore there is no mis-apprehension to correct. The fact that El Reg is able to dredge up the original story shows that this isn't the case and a proper retraction should have been posted.
"Although the story doesn't give much detail on his disability, it's quite possible that the car's (presumably automatic) controls were specially adapted for him and this is some way compounded the issue of not being able to stop."
Having read the rest of the comments, it appears that far too few people have picked up on this point. We are talking about a fly-by-wire car that has been modified in ways unanticipated by the original designers and is now being driven by someone who (at best) will only have skimmed the part of the instruction manual that tells you how to kill the whole car in the event of a catastrophic malfunction.
"by some estimates, Reader is on 90 per cent of PCs in the Western world."
Odd. Anyone running a Windows PC without an AV package is regarded as recklessly insane, and yet:
None of these packages flag up Reader, despite it being the biggest single attack vector there is.
People running this attack vector are *not* regarded as recklessly insane.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
