So far, no evidence the alleged ISP net tech is either working for an ISP or a network engineer. Still, plenty of time. The questions are carefully selected to examine the range of understanding in the specific area under dispute as well as the history of network engineering.
I'll add a few more. If there's no response by the claimant, then I might turn these into a drinking game.
7. Name the congestion control schemes relating to UDP that are also names of colours? (This is fair. If you're into drinking games, you'll know why.)
8. Name the author of the Multicast section in LARTC.
9. Name the predecessor to VIC.
10. What is the command to enable ECN on Cisco routers?
11. What is PIM-BiDi?
12. What are the current RFCs for IGMPv3 and MLD?
These aren't intended to be derogatory or insulting, I genuinely want a clear picture of who is debating this. An unquestionable demonstration of the level of understanding they have. How they interpret this is their problem, I'm interested and I follow the SOP of any geek who is interested, I ask questions.
There are no trick questions, there are no traps, these are simple, honest questions.
Trump despises Bezos and cares nothing for the unemployed.
Cutting from the low-earners won't change anything, since that increases the tax they pay.
Cutting a percentage point or two from all those earning $10 million a year or more to and diverting that to paying the low earners will eliminate the tax bill and will cost less than the tax would.
That would raise the profits of the company and share value. It's the bonuses that are the big source of income for big earners, so bigger bonuses will cut into, or eliminate, the loss from regular wages.
If society is so decrepit that OAPs have to work in hostile, abusive environments to buy food, then society doesn't get to complain when the environments are made less hostile and abusive.
The problem needs to be fixed. If you don't fix it, the worst possible people to get involved will. So next time fix things before they think it's a good idea. You get to choose your reaction times here.
And when all employers you can work for set unlivable wages, who do you work for?
Your proposal is like sending telephone sanitisers into space on a B Ark. Sounds great on paper, dooms you to extinction because you have nobody left to do the jobs that actually were essential after all.
No. Government's job is to represent the people and serve the people. It is not to serve businesses, who are NOT people.
Trump, however, is in an interesting position. Back Sanders and risk alienating voters, or back both one of his enemies and lower wages for his core supporters, risking alienating voters.
We live in a society. The college education of one pacts directly the well-being of thousands of others and the profitability not just of where they work but dozens of companies enmeshed in a complex web.
We are all just nodes in the web, ultimately it is the web that exists. There is no individual responsibility because individuals affect too many others and you can't calculate the effect.
Free university is affordable, Britain could afford it without trouble when the Thatcherite aim was 100% in college. It's affordable because the total benefit vibrates through the web and raises the profitability of many and not just the one.
This hasn't been the experience in America, to the same degree, because humanities were at the expense of STEM (neoclassical education requires university involve both sides, so there are no pure liberal arts graduates - that's why it worked in Britain for so long), because education at school has too high a religious content (so the first year is catch-up) and because there's a fundamental mistrust of higher learning. Egg-heads are looked down upon. Tales of success are dominated by drop-outs. Inventors are sneered upon, whereas patent trolls are held in high regard.
If you want a functional, profitable web, that's how not to go about it.
The Internet is being run on wired land lines. That's how it works.
There's a difference between setting up dedicated bandwidth (say, via a protocol like RSVP) for an emergency phone call or telerobotic surgery (and, yes, space has been cleared for the latter on the public Internet) and confiscating bandwidth you've bought because you're successful.
Not confiscating because someone else needs it (I've pointed out elsewhere how to do weighted round robin and other fair service management) but because you're seen as a problem.
The correct approach is to divide bandwidth rationally. If you've bought N% of the total downstream pipe, you are guaranteed to be able to use up to N% of the upstream pipe. What you don't use should be made available to those with extra demand. Apply at each router/switch. It's not an expensive algorithm.
No throttling, just a fair division of resources.
Throttling means providing a site with less than that N%. Throttling when popular means seeking to make a site unpopular. That's why you would do this. It does not mean sharing, it means confining. What I described would be sharing, but it isn't throttling. Even if you added RED.
In the case of video sharing sites, I have no sympathy at all with ISPs or with MPAA. They created this mess by blocking multicast and web caching to the home because they couldn't bill it. If multicast had been widely available then multiple people streaming the same thing at more or less the same time would not occupy any more of the net than one. If caching had remained in place, the bulk of the Internet would have remained clear.
This is a self-inflicted problem and the ISPs should sit down with the MPAA to figure out how to undo their mistakes.
Unusually for them, the vendors are almost innocent.
On the day our beloved folk hero, Servalan (aka Jaqueline Pearce) died. The fact that she was notoriously sexy in her youth in no way relates to the version number.
Seriously, a revamped design? You're infinitely better off with a modular design and a scripted UI. That way, you can upgrade modules independently and the design is whatever the user wants.
1. Ted Neilson's Xanadu. Never got off the ground.
2. IPv6. The original spec required all Internet traffic to be over IPSec, with server networks using digital certificates to prove their identity.
3. Class 3 SSL certificates. These were certificates released if the person could prove their identity and could prove they had the right to the certificate. Nothing more for user certificates, proof of ownership of domain name and business for server certificates.
4. Smart web pages. If you're using AJAX and servlets, everything can be done in data, you don't need to mess with the URL.
The result? A few of these are utilized, but most webmasters either don't understand the technology, won't use it or have been ordered not to by their boss.
Hacking the URI bar won't change that.
If Google wanted a better system, they'd start by looking at TUBA, one of the IPng/IP6 candidates. If you can uniquely express a resource with an address, you can give it a name. TUBA has infinitely variable length addresses, so it's easy to code a directory path into the physical address. It's an address so it can be given a unique name.
Your webserver is now a virtual network rather than a filesystem.
Nope. Because, for example in Britain, the police are largely disarmed too. And we like it that way. Total balance. No unfair advantage, no might makes right mentality by either side, civility rules UK.
But I don't expect that to cut any ice with other societies. Different cultures have different views. That's ok, as long as they keep their noses out of ours.
And of course that plays both ways. Brits should be wary of criticizing American culture, it has diverged in the last 300 years and is NOT just Britain with a funny accent. It's as foreign as Africa.
Given we don't see politicians publishing their bank account details and credit card numbers, we can assume politicians really don't understand the consequences of their proposals.
Ignorance is a really bad place to be making decisions from.
Yep, nothing obvious to show it's anything other than a toy os or that it can do anything I asked. If you're going to pull the sarcasm, best be sure there's some meat on the bone.
I take it, then, that you cannot show me an OS that can perform the operations listed, or that has the level of provable reliability, or that has similar performance characteristics.
Software capable of running theorems on software haven't been around that long, to the point of being usable I'd say they're newer than Rust.
There is no "standard C", there are only C standards. And almost nobody complies with them. So your argument is a little dubious. Everything out there is a subset.
And, yes, subsets are a valid dialect as long as there's enough of the core there to be that language and not another. Verified C and SystemC are dialects. You understand, I hope, that programmers have recognized language dialects as being versions of the language since the 1960s. You don't get to come along and just blow away the concept because you don't like it.
Especially when it means there are no C compilers, if you do that.
My point is that the languages are not equivalent if you cannot provide the same proofs of correctness, the same hard real-time guarantees, the same compactness, the same level of performance and the same underlying capabilities.
This isn't about operating systems, this is about language equivalence.
In mathematics, equivalent isn't the same as equal. If you can't match SEL4's proofs, you can't claim the same or better defect density and you can't claim the software satisfies the design constraints. All you can do is test and claim that it's not failed so far. Those aren't the same thing.
If you can't write a hard real time OS that can guarantee that the start and end deadlines are satisfied at least as well as the best of those I've listed, your underlying language is not equivalent. It cannot do the same things done in C/C++. There's latency and latency is an often ignored factor in programming.
If you can't put your OS onto an FPGA, it lacks the compactness that is possible in C. Space is another ignored factor, owing to most environments being huge. In fact, as SystemC is a C dialect, I can compile an OS onto an ASIC. Can you do likewise?
Performance isn't just about latency. It's about efficiency. Occam is far more efficient than C, at a lot of things, and should be the ideal language for an OS. But because it relies on certain properties being static, it's very inefficient at highly dynamic loads. I love Occam but you'd need to modify it to run Linux as well as C despite it being a bettera language. I have seen claims Rust would be better, but no evidence. I'm offering the chance for you to show me some.
Underlying capabilities. C is happy with me using RDMA to transfer data directly into my program without the program knowing about it. The whole point of Java is the sandbox, RDMA would negate it. In C, I can switch processor rings and put data directly onto the PCIe or PCI-X bus, edit the contents of flash memory chips and upload new microcode to fix yet another defect in silicon. I can write code for managing the hot-swapping of CPUs or memory. Can I do all this in Rust?
If there's anything I can't do in this list in Rust, it can be an excellent language (it is) but it's clearly not able to reimplement feature-for-feature the operating systems I've listed.
Slashdot Mirror
Thargons.
This is clearly an attempt to create a Thargoid base on Saturn, or maybe it's one of their ships.
On being asked for a statement, Commander Jameson repeatedly stated "It isn't my fault!".
That's easy. A.
Now, what is the airspeed of an unladen swallow?
So far, no evidence the alleged ISP net tech is either working for an ISP or a network engineer. Still, plenty of time. The questions are carefully selected to examine the range of understanding in the specific area under dispute as well as the history of network engineering.
I'll add a few more. If there's no response by the claimant, then I might turn these into a drinking game.
7. Name the congestion control schemes relating to UDP that are also names of colours? (This is fair. If you're into drinking games, you'll know why.)
8. Name the author of the Multicast section in LARTC.
9. Name the predecessor to VIC.
10. What is the command to enable ECN on Cisco routers?
11. What is PIM-BiDi?
12. What are the current RFCs for IGMPv3 and MLD?
These aren't intended to be derogatory or insulting, I genuinely want a clear picture of who is debating this. An unquestionable demonstration of the level of understanding they have. How they interpret this is their problem, I'm interested and I follow the SOP of any geek who is interested, I ask questions.
There are no trick questions, there are no traps, these are simple, honest questions.
Trump despises Bezos and cares nothing for the unemployed.
Cutting from the low-earners won't change anything, since that increases the tax they pay.
Cutting a percentage point or two from all those earning $10 million a year or more to and diverting that to paying the low earners will eliminate the tax bill and will cost less than the tax would.
That would raise the profits of the company and share value. It's the bonuses that are the big source of income for big earners, so bigger bonuses will cut into, or eliminate, the loss from regular wages.
If society is so decrepit that OAPs have to work in hostile, abusive environments to buy food, then society doesn't get to complain when the environments are made less hostile and abusive.
The problem needs to be fixed. If you don't fix it, the worst possible people to get involved will. So next time fix things before they think it's a good idea. You get to choose your reaction times here.
And when all employers you can work for set unlivable wages, who do you work for?
Your proposal is like sending telephone sanitisers into space on a B Ark. Sounds great on paper, dooms you to extinction because you have nobody left to do the jobs that actually were essential after all.
No. Government's job is to represent the people and serve the people. It is not to serve businesses, who are NOT people.
It would break the PHP on some news sites.
He's bdoing this for social justice, not Trump.
Trump, however, is in an interesting position. Back Sanders and risk alienating voters, or back both one of his enemies and lower wages for his core supporters, risking alienating voters.
It's down to who he hates more right now.
We live in a society. The college education of one pacts directly the well-being of thousands of others and the profitability not just of where they work but dozens of companies enmeshed in a complex web.
We are all just nodes in the web, ultimately it is the web that exists. There is no individual responsibility because individuals affect too many others and you can't calculate the effect.
Free university is affordable, Britain could afford it without trouble when the Thatcherite aim was 100% in college. It's affordable because the total benefit vibrates through the web and raises the profitability of many and not just the one.
This hasn't been the experience in America, to the same degree, because humanities were at the expense of STEM (neoclassical education requires university involve both sides, so there are no pure liberal arts graduates - that's why it worked in Britain for so long), because education at school has too high a religious content (so the first year is catch-up) and because there's a fundamental mistrust of higher learning. Egg-heads are looked down upon. Tales of success are dominated by drop-outs. Inventors are sneered upon, whereas patent trolls are held in high regard.
If you want a functional, profitable web, that's how not to go about it.
employee = malloc (sizeof(job));
free(employee);
Is being placed under the rocket vents the same thing as fired?
Ok, a pop quiz
1. What's the multicast address range for IPv6?
2. What were the years the mbone and 6bone went native on the backbone?
3. What is the relationship between the TTL and hop count on a multicast group?
4. What happens when you switch off multicast routing on an interface?
5. Who had the first IPv6 hub in the United Kingdom?
6. Who are the various authors of the major multicast how-tos?
The Internet is being run on wired land lines. That's how it works.
There's a difference between setting up dedicated bandwidth (say, via a protocol like RSVP) for an emergency phone call or telerobotic surgery (and, yes, space has been cleared for the latter on the public Internet) and confiscating bandwidth you've bought because you're successful.
Not confiscating because someone else needs it (I've pointed out elsewhere how to do weighted round robin and other fair service management) but because you're seen as a problem.
According to Ars Technica, ISPs like Comcast have no problem destroying the equipment of rival ISPs. And I don't mean figuratively.
In that case, they already see themselves as above the law.
No.
The correct approach is to divide bandwidth rationally. If you've bought N% of the total downstream pipe, you are guaranteed to be able to use up to N% of the upstream pipe. What you don't use should be made available to those with extra demand. Apply at each router/switch. It's not an expensive algorithm.
No throttling, just a fair division of resources.
Throttling means providing a site with less than that N%. Throttling when popular means seeking to make a site unpopular. That's why you would do this. It does not mean sharing, it means confining. What I described would be sharing, but it isn't throttling. Even if you added RED.
In the case of video sharing sites, I have no sympathy at all with ISPs or with MPAA. They created this mess by blocking multicast and web caching to the home because they couldn't bill it. If multicast had been widely available then multiple people streaming the same thing at more or less the same time would not occupy any more of the net than one. If caching had remained in place, the bulk of the Internet would have remained clear.
This is a self-inflicted problem and the ISPs should sit down with the MPAA to figure out how to undo their mistakes.
Unusually for them, the vendors are almost innocent.
On the day our beloved folk hero, Servalan (aka Jaqueline Pearce) died. The fact that she was notoriously sexy in her youth in no way relates to the version number.
Seriously, a revamped design? You're infinitely better off with a modular design and a scripted UI. That way, you can upgrade modules independently and the design is whatever the user wants.
Servalan, I think, would have approved.
1. Ted Neilson's Xanadu. Never got off the ground.
2. IPv6. The original spec required all Internet traffic to be over IPSec, with server networks using digital certificates to prove their identity.
3. Class 3 SSL certificates. These were certificates released if the person could prove their identity and could prove they had the right to the certificate. Nothing more for user certificates, proof of ownership of domain name and business for server certificates.
4. Smart web pages. If you're using AJAX and servlets, everything can be done in data, you don't need to mess with the URL.
The result? A few of these are utilized, but most webmasters either don't understand the technology, won't use it or have been ordered not to by their boss.
Hacking the URI bar won't change that.
If Google wanted a better system, they'd start by looking at TUBA, one of the IPng/IP6 candidates. If you can uniquely express a resource with an address, you can give it a name. TUBA has infinitely variable length addresses, so it's easy to code a directory path into the physical address. It's an address so it can be given a unique name.
Your webserver is now a virtual network rather than a filesystem.
That doesn't sound like what they're doing.
Yes, but only Class 3 matters and everyone bought the cheapo ones instead.
Maybe so, but Voltaire never said it. It's a fictional quote by a biographer.
Nope. Because, for example in Britain, the police are largely disarmed too. And we like it that way. Total balance. No unfair advantage, no might makes right mentality by either side, civility rules UK.
But I don't expect that to cut any ice with other societies. Different cultures have different views. That's ok, as long as they keep their noses out of ours.
And of course that plays both ways. Brits should be wary of criticizing American culture, it has diverged in the last 300 years and is NOT just Britain with a funny accent. It's as foreign as Africa.
If there's a backdoor, there's no encryption.
Given we don't see politicians publishing their bank account details and credit card numbers, we can assume politicians really don't understand the consequences of their proposals.
Ignorance is a really bad place to be making decisions from.
Ah. So you're saying I can drop a GCC program in Green Hills, PGC or Intel?
Optimist.
Your sarcasm has a defect.
Yep, nothing obvious to show it's anything other than a toy os or that it can do anything I asked. If you're going to pull the sarcasm, best be sure there's some meat on the bone.
I take it, then, that you cannot show me an OS that can perform the operations listed, or that has the level of provable reliability, or that has similar performance characteristics.
Much better just to say so.
Software capable of running theorems on software haven't been around that long, to the point of being usable I'd say they're newer than Rust.
There is no "standard C", there are only C standards. And almost nobody complies with them. So your argument is a little dubious. Everything out there is a subset.
And, yes, subsets are a valid dialect as long as there's enough of the core there to be that language and not another. Verified C and SystemC are dialects. You understand, I hope, that programmers have recognized language dialects as being versions of the language since the 1960s. You don't get to come along and just blow away the concept because you don't like it.
Especially when it means there are no C compilers, if you do that.
My point is that the languages are not equivalent if you cannot provide the same proofs of correctness, the same hard real-time guarantees, the same compactness, the same level of performance and the same underlying capabilities.
This isn't about operating systems, this is about language equivalence.
In mathematics, equivalent isn't the same as equal. If you can't match SEL4's proofs, you can't claim the same or better defect density and you can't claim the software satisfies the design constraints. All you can do is test and claim that it's not failed so far. Those aren't the same thing.
If you can't write a hard real time OS that can guarantee that the start and end deadlines are satisfied at least as well as the best of those I've listed, your underlying language is not equivalent. It cannot do the same things done in C/C++. There's latency and latency is an often ignored factor in programming.
If you can't put your OS onto an FPGA, it lacks the compactness that is possible in C. Space is another ignored factor, owing to most environments being huge. In fact, as SystemC is a C dialect, I can compile an OS onto an ASIC. Can you do likewise?
Performance isn't just about latency. It's about efficiency. Occam is far more efficient than C, at a lot of things, and should be the ideal language for an OS. But because it relies on certain properties being static, it's very inefficient at highly dynamic loads. I love Occam but you'd need to modify it to run Linux as well as C despite it being a bettera language. I have seen claims Rust would be better, but no evidence. I'm offering the chance for you to show me some.
Underlying capabilities. C is happy with me using RDMA to transfer data directly into my program without the program knowing about it. The whole point of Java is the sandbox, RDMA would negate it. In C, I can switch processor rings and put data directly onto the PCIe or PCI-X bus, edit the contents of flash memory chips and upload new microcode to fix yet another defect in silicon. I can write code for managing the hot-swapping of CPUs or memory. Can I do all this in Rust?
If there's anything I can't do in this list in Rust, it can be an excellent language (it is) but it's clearly not able to reimplement feature-for-feature the operating systems I've listed.