Posts by Yes Me 1742 publicly visible posts • joined 11 Jan 2008
"former IBM employees sought to invalidate a portion of their separation agreements in which they waived the right to bring collective legal action against the company."
I really hope they succeed on appeal. Such clauses, and NDAs in separation agreements, seem to be a violation of very basic civil rights. I don't see a Trumpian Supreme Court agreeing, though.
They don't even need to do that. The EU, if you haven't noticed, has been 100% consistent throughout the negotiations, while the Tories have been changing from one set of absurd demands to another as fast as the seasons have changed each year. Only a day ago, this:
Britain hopes never to need to use proposed powers to break its Withdrawal Agreement with the European Union, Northern Ireland Office minister Robin Walker said [Reuters]In other words, we'll risk destroying what remains of our negotiating position and the peace in Northern Ireland for the sake of something we hope we never need. These people are crazy, so it's no good expecting any sane replies about Galileo. (He was only an f***ing Eytie, after all.)
"you can make all the same mistakes... in an entirely new way."
Not that this was ever known to happen before in the history of software, of course... Nothing like this was observed in the 1960s, for example. (Search for "second system syndrome")
What is definitely happening now is that China is working hard -- very hard -- to bring its semiconductor design and manufacturing supply chain up to standard. It may take them a few more years, but Trump has taught them a lesson and they have learnt it: don't rely on foreigners for your basic hi-tech. So in the end, the US and other "developed" countries will be the losers. Meanwhile, Huawei may have a few tricky years but they will come back stronger than ever. By then, Trump will be irrelevant except to historians.
The Internet runs on redundant paths. So yes, a political shutdown of a route via China would be a nuisance and would reduce global capacity. But no, it wouldn't break the Internet. That's what BGP4 is all about (even when it has a hissy fit like was reported yesterday, it doesn't break the Internet).
Surveillance and weak crypto is a concern. So is DDOS. But the fact that a path goes through China isn't really important even for that; spooks exist everywhere.
This whole "clean Internet" stuff is political grandstanding and technical nonsense. It's just another battlefront in the trade war that Trump has started.
And whatever American-style libertarians say, this problem is not the price of democracy. A democracy can perfectly well have, and enforce, laws against hate speech and the like. A democracy does not have to tolerate intolerance. Yes, it would cost Facebook real money to censor evil postings. It might even drive them out of business. But it would be perfectly fair and democratic, since it would apply under the rule of law to every publisher.
The thing about plain old ASCII email is that it is remarkably robust and (excuse shouty) IT LOOKS THE SAME ON EVERYBODY'S SCREEN (as long as they have the sense to use a fixed width font; I like Consolas). Yes, occasionally some italics or boldface might be handy but the advantage of universality trumps all.
Use a version control system for the code, by all means. Use Git if you must. Use an issue tracker. But for accurate communication between geeks at long distances, plain text wins every time.
They (Google) don't care about your users. They care about the mass market, which is where they collect the private information that makes advertisers super happy.
I was forced only the other day to fire up Chrome, by a video streaming site that simply told me that my other browser was no good. Also because using Chromecast except from Chrome is a bust. So they got a bit more of my private life into their machine learning system.
As we know very well, all problems in computer science can be solved by an extra level of indirection. Which means in this case that all the security crap you know and love (firewall, access control lists, certificates, crypto algorithms, switching from TCP to TLS1.3, from UDP to DTLS, need I go on?) -- all of it -- will have to be duplicated in the browser.
So isn't this just a way of helping along the plan for Chrome to take over the universe?
If the law says that, the law is a ass. People should not be allowed to post dangerous lies in the guise of "free speech". The main anti-vaccination lies are proven false and are intrinsically deadly dangerous. People die because of them. It should be a crime to post them, and a crime not to take them down.
Also they are only suing for $5M, it says. That's not even a gnat's bite for FB or Zuck. He won't lose sleep over this.
"technology-as-a-service" is the actual buzzword they're using. It's how IBM became a near-monopolist in the old days: you couldn't buy Hollerith card processing equipment or early computers like the IBM 650, because IBM refused to sell them. And guess what, the rental costs were enormous...
Until, that is, IBM was forced by US anti-trust law to sell stuff too. In 1956.
Memories are short, apparently.
Apparently, then, somebody in the Administration understands that like any country, the US needs imported hi-tech skills because they don't actually have enough home-growns. If H-1B holders are underpaid, that's really a separate matter from whether they're needed, and should be resolved separately.
Of course you sometimes have to print it. For example when your stupid employer decides it's essential to autoexpire and autodelete all email after 6 months. So anything of importance (e.g. employment related promises from your stupid employer) must be printed.
What are you on about? The International Postal Union has never had anything to do with IP addresses, or even telephone numbers if that's what you meant. The ITU (previously CCITT) defined international telephone prefixes like +1 and +44, but the rest of telephone numbering is arranged nationally. IP addresses are completely different because they are non-geographical, so are allocated by a handful of Regional Internet Registries. And of course, with IPv6 they are no longer a finite resource, in any realistic way.
What do you mean by "slackening of governance"? There's never been anything resembling governance of IP address space; there's only been allocation and registration.
And what do the authors of the study mean by "policies regarding the regulation of IPv4 markets"? There is no regulation of IP address space and never has been. So no policy, either. Again, it's just a matter of allocation and registration. They're only numbers, and unlike telephone numbers, they have no historical connection with geography.
Anyone who relies on IP address bits for any kind of intrinsic validation is asking for trouble.
Most governments are two-faced.
I don't believe that Israel ever sanctioned IBM for supplying the equipment that Nazi Germany used to tally the concentration camp victims. Guess what, they had 6 million moral incentives to do so, but one overwhelming economic incentive not to.
Also guess what, the USA mistakenly believes it has an economic motive to sabotage Huawei. How convenient that they could hide this behind a moral incentive.
That doesn't mean that I'm OK with the mistreatment of the Uighurs. It's awful. But let's not pretend that the US is whiter than the driven snow. We're talking about Trump and Pompeo, for heaven's sake.
"mainly due to his post-war preening"
WTF? Even Turing's mother didn't know what he did in the war until 20 years after he died, shortly before her own death. Most of the world didn't know until 30 years after he died, when the Hodges biography was published.
And others who served at Bletchley Park with him have made it quite clear that he was one of the leading lights; the other one who could be compared to him in mathematical importance was W.T. Tutte, but he didn't co-invent computability theory, artificial intelligence or mathematical models of morphogenesis; he is, however, well known in graph theory.
No question that Alonzo Church's work was important too, but it was Turing who built the bridge between theory and real-world machinery.
The assertion that the Z3 was Turing-complete is very dubious. Being programmable doesn't make a machine equivalent to Turing's U. "Program code was stored on punched film" according to Wikipedia. Even its successor the Z4 had no memory for the program; all it had was "a mechanical memory with 64 words." [1] That makes it programmable like a Jacquard loom was programmable in 1805, or like ENIAC was programmable in 1945. None of these was Turing-complete, because to emulate U, the machine must have a rewritable memory where its own program is stored. (Somebody editing Wikipedia seems to be very confused about this).
Actually nobody has ever built or will ever build a Turing-complete machine, because it needs an infinitely long rewritable memory tape. But the first approximation to one was only built in 1948, in Manchester, England. I'm typing this, and you're reading it, on a slightly updated version of the 1948 device.
Zuse was indeed a great inventor, but he knew nothing about Turing's work until well after the end of WW II.
[1] See Konrad Zuse's Z4: Architecture, Programming, and Modifications at the ETH Zurich, by Ambros P. Speiser, in The First Computers - History and Architectures, edited by Raúl Rojas and Ulf Hashagen, MIT Press, 2000.
Of course it's about trade, but sadly the logic is utterly flawed. What's happening is that the US (Dems as well as Trumpettes) has given China a strong message: develop your own semiconductor supply chain, right back to the rare ores mined in developing countries. So 10 or 15 years from now, China (including Huawei) will simply laugh at US sanctions and they will be making cheaper, better stuff just as they are today, with no dependency on Western industry. The current trade war is just a minor glitch for them.
They'll be fine. They have an enormous domestic market, many eager customers in the developing world where the Writ of Trump does not run, and just a temporary problem in some Western countries. Come back in ten years and you'll see that they will have the benefit of a major new semiconductor supply chain in China itself. The companies like Ericsson who may benefit from the US/China trade war today will be wiped out tomorrow.
British telcos will be running mainly Huawei kit ten years from now.
"ditching Huawei is about much more than following trump."
No, really, it isn't. You don't cite any evidence, but the fact is that all this started with Western vendors paying lobbyists in Washington DC to whip up Congressional hysteria against Huawei, because Huawei kit is better & cheaper & threatens their profits. This plays well in Washington, with appeal to both Democrats & Republicans, but it plays best of all with Trump and his infantile misunderstanding of international trade. And for whatever reason, the Tory right wing has dug itself into a deep hole that makes them slaves of Trumpery.
Assuming Trump is duly kicked out in November, all this will start to unravel by next February or March, so I don't expect that any UK telco will actually need to remove a single item of Chinese kit.
So it's OK to discriminate against China to the benefit of the US?
Are you sure that no citizens of the US feel themselves to be trampled on by the authorities?
It's OK to sell weapons to Saudi Arabia, but not to buy telephone equipment from China?
Are you sure that citizens of Yemen would agree to that?
The world is a dirty place. It's very, very hard to have clean hands.
By about a million % it benefits the UK more than China. And it's in the Cambridgesphere because that is the UK's high tech corridor, of course. It's where the skilled staff are. (It's also one of the strongest Remain voting areas, which shows that it's a much better location for an international company than any brownfield site in the Brexit heartlands. Don't imagine that Brexit is over...)
How about ISOC's work in developing countries (https://www.internetsociety.org/issues/access/) or their work in favour of encryption (https://www.internetsociety.org/issues/encryption/) or supporting Internet exchange points (https://www.internetsociety.org/issues/ixps/), etc. etc.
All funded by income from PIR, in the public interest.
Please look at the facts.
The California AG stuck their nose in for no good reason, but since they succeeded in scaring ICANN off, nothing else will happen. It's over. In fact it never began.
As for PIR's existing contract with ICANN, as far as I know they didn't violate it and it will run its course. I forget when it comes up for renewal, but that's public information.
The person making the comment about conflict should post the names of the people who were on both sides. There has always been some overlap between the ISOC Board and the PIR Board, because ISOC owns PIR. No conflict there. So who are you accusing of conflict? I'm not aware of anybody from the ISOC/PIR side who had anything to do with Ethos Capital.
Secondly, I think that any laws about this would only apply in the case of companies with publicly traded stock, which is not the case. In any case, both ISOC and PIR have conflict of interest policies and as far as I can tell are not afraid of review.
"there’s nothing in [the] recent announcement that addresses the unilateral decision-making that led to the Internet Society (ISOC) board of directors deciding in secrecy over a matter of weeks to sell the .org domain,”
Um, no, why should there be? The Internet Scociety (which anybody can join -- anybody) has a Board (selected by a well documented community process) and the Board gets to decide on things. And from the first announcement, they made it clear that the confidentiality was not their choice; the company offering to buy insisted on confidentiality, which is pretty normal in any $B deal that doesn't involve stock trading.
Where you detect hypocrisy is beyond me. ISOC was transparent from day 1 about why they wanted the deal and what they would do with the proceeds.
Apparently the piffle about the sacredness of the .org registry continues. It's a list of names, for heaven's sake. It's nothing other than a list of names. To have control of a name on that list, you have to pay a modest annual fee. That's all there is to it. There's no vetting, and never has been since 1998, whether the "organisation" using a name is good, bad, or indifferent, or even whether it is an organisation at all rather than an individual person or just a robot. The only qualification is the ability to pay the fee. If the EFF forgets to pay its fee one year, eff.org might pop up the next month as a porn site. There's no magic in .org. (www.magic.org shows this quite neatly.)
"The Internet Engineering Task Force (IEFT) points out that 'master-slave is an oppressive metaphor that will and should never become fully detached from history' as well as 'In addition to being inappropriate and arcane, the master-slave metaphor is both technically and historically inaccurate,'" he wrote... This comes from an IETF draft document published in 2018Factually speaking, the IETF did not point out anything of the kind. That quote is from a draft that expired in September 2019 and in no way is a statement by the IETF.
It is also a fact the one of the authors of that draft has made a point of keeping an eye on other drafts in order to sus out terms like master/slave, blacklist/whitelist and even balkanization, which some people from the Balkans don't much like. And some people think it isn't OK to call other people snowflakes, even if they're snowflakes.
Turns out this is a research topic: see the Human Rights Protocol Considerations research group.
Exactly! If the British had installed a democratic system in HK, say during the 1960s at the height of decolonisation but before the opening of China, the one-country-two-systems deal negotiated in 1997 would have been very different, and mouthpieces like Carrie Lam would never have been installed.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
