Single weights of the Helvetica Now typeface are available for $/EUR 35 or GBP 30 each. The complete typeface family is available for $/EUR 299 or GBP 249.
(I had to change the currency symbols because Slashdot still doesn't handle non-ASCII characters correctly.)
Yes and there's no such thing as x86-64 either, Intel marketing decided it was to be called "IA32-T" or some other nonsense. It's best to ignore the vendor's PR names in such things, particularly when they're as inane as "AArch64".
I fry celery in fat all the time when making soup. Chop some onions, grate some celery into a mush, and cook both in olive oil. Then you go on to add the other ingredients. You don't burn or blacken it though, so perhaps nitrosamines cannot form.
You're missing the point. If you can vote at home with your PC or phone, anyone can ask to look over your shoulder while you do it. Or they can bribe or threaten you to let them look.
A secret ballot is just that -- entirely secret. Nobody is allowed in the voting booth with you. You cannot prove how you voted, even if you want to.
Perhaps by "blockchain voting" they mean you still go to a polling station to cast your vote in secret, but it's all fully verifiable because blockchain. Actually, cryptographers have already come up with systems where you can check your vote was counted without being able to prove what it was (taking home some kind of paper receipt). But I think of all the problems faced by democracy this isn't the most pressing. Those who want to stuff the ballots would just ignore the result of the cryptographic verification, after all, it's too complex for anyone without a PhD to understand (or they could just introduce their own fake 'experts').
(This is also why postal voting is flawed, in my view -- for people who can't attend on election day, they should be able to visit a special polling station a week or two in advance, but not to just mail in a vote which anybody could have "supervised". There have certainly been cases of fraud with postal votes, too.)
Those are run time checks. They are a great way of smoke-testing your code for memory access errors that can't be detected at compile time (which in standard C is most of them). Of course, while this testing can prove the existence of bugs, it cannot prove their nonexistence.
The paper is about a different and complementary approach, compile-time checking. Here you prove that certain types of unsafe access cannot occur based on the source code. The prior art cited in the paper is other approaches of this type.
The pro cards (Quadro, and also some NVS cards) nowadays come with four DisplayPort outputs. The consumer cards tend to have one DisplayPort and a mishmash of HDMI and DVI. That is annoying, since there are converters from DisplayPort to both of these but not the other way round. The GeForce cards also tend to be dual-slot, power-hungry monsters while the lower-end Quadro cards fit in a single slot and don't need auxiliary power.
So I think there are reasons to prefer the Quadro line, apart from the question of "certification" (which does look like like a bit of a racket).
In principle you can license and build a GeForce card with four DisplayPort connectors, but they are very hard to find. The best I was able to buy was a GeForce 1060 with dual DisplayPorts.
Ah, of course you're right. Microsoft ported lots of the classic DOS utilities to native Windows versions, including the command prompt itself (COMMAND.COM became CMD.EXE) but porting EDLIN would have been a step too far.
You can blame AMD for not including a way to go back from V86 mode to x86_64 mode, I think.
Recent WIndows versions no longer include EDLIN. It was there in NT 4.0 but is missing in Windows 10, so must have been dropped somewhere along the way.
If yours have the detachable cable (SDL connector) you can buy a replacement cable that has USB on the other end. Inside it contains a "Soarer's converter" which is the best PS/2 to USB converter.
I'm curious - I've used Model M and Model F keyboards for, oh, nearly 25 years now and I've never noticed springs starting to "wear". Maybe if you popped the keys off all the time to mess around, you might risk damaging a spring. Otherwise they seem well protected and certainly don't appear to suffer any ill effects from being pinged back and forth thousands of times a day. What's the typical failure?
Well strictly speaking the string handling is more to do with the C standard library than the language. You could use a library providing a sensible string type, and the main restriction imposed by the language would be that literal strings cannot contain the 0 byte. It's dispiriting that after so many decades the C standard library remains stuck without a decent set of string functions. (There are plenty of libraries to choose from but the real value comes in picking one and using it everywhere; it's a social rather than a technical issue.) I read that C++'s STL also lacked a string type and some lobbying was needed to get the standards committee to define one.
I find the ODbL's contract terms problematic. It purports to be a contract, not a licence, and to impose additional restrictions beyond what copyright law allows it to enforce. In my view that makes it non-free. I greatly prefer the style used by the GPL and other free licences: "You are not required to accept this License, since you have not signed it." Then the licence is a grant of permission to do something which would not otherwise be allowed -- so it's legally much more watertight. (If you go to court arguing that the licence is invalid, you are admitting that you don't have any right to copy the work it covers.)
The rationale for this is that in some countries map data may not be covered by copyright. This is dubious -- legal research I commissioned from US and English lawyers concluded that while it's not entirely clear-cut, there are certainly strong precedents for map data falling within copyright. In my personal view, I think that even one country decided, as a matter of policy, to exclude map data from the scope of copyright and have it be in the public domain, that is a decision for the legislature of that country, and not something the licence on an open data project should try to block.
There are plenty of other licences which can be used for data, require attribution, and don't have the problematic contract terms and overreach of the ODbL. The OSM foundation should dual-license under one of these.
Sometimes I wish that if the optimizer finds something really juicy (like eliminating a dozen lines of code because it can prove they will never be called by assuming that undefined behaviour won't be triggered) it would just refrain from optimizing the code out and instead emit a diagnostic telling the programmer to apply that optimization in the source code. You could then review the code and either say 'gee, the compiler is right, I will delete that whole branch', or 'ouch, that should not be happening... let me find where I am relying on undefined behaviour', or in some cases add an explicit compiler hint to allow the optimizer to take an axe to the code without bothering you further.
It would have to be a heuristic, and would produce some false positives (where the optimizer is fine, and should just do its job without chatter), but in the cases it did work it would avoid a lot of acrimony and language-lawyering.
The most valuable company is surely Aramco, or more correctly Saudi Aramco, valued at between 2 and 10 trillion dollars. It still begins with A, more or less.
The headline should have talked about the second most valuable *listed* company. Although a small part of Saudi Aramco shares is expected to be listed soon.
I can think of one example, block-sorting file compression as used by bzip2. If you are making a tarball of several files, it will presumably help to have files with similar content next to each other in the archive. (The block-sorting process would rearrange them later to some extent, but I believe that compression will be a bit better if they are already close, particularly if the input data is much bigger than the sorting window.)
So you first sort by some measure that only approximates sorting by content. That is, you sort by filename first. It's a guess that similar filenames will have similar content. That could be completely wrong, of course, but the heuristic tends to work in practice.
(This applies to most forms of compressed archive, even if the compression algorithm doesn't involve sorting, but here you have an example of where it's worth sorting by some approximate key before sorting again.)
Another example may be bulk insertion into a relational database table, where inserting rows in sorted order is fastest, but you may not have the memory to sort them on the client. If you could sort them by something that tends to correlate to the final ordering, you will typically get some speedup in practice.
I have been in a London taxi where the driver had Google Maps open -- not, he explained, to navigate the route, but to show where traffic congestion was.
I don't think it is correct to say that the GPU needs to be four times as powerful for a display with four times as many pixels. For 3d gaming, yes. But for text rendering, I don't think that a higher resolution web page takes four times the processor power to lay out. The same number of character glyphs need to be painted on the screen; they are bigger, but still nothing that challenges even the weakest GPU. The difficult part is rendering the glyphs at the higher resolution, and that is CPU-bound. Sure, a more powerful GPU would help, but it doesn't need to be four times as powerful.
As anecdotal evidence I note that for a couple of years I ran two 5k monitors from a single NVS 510 desktop video card. This is a pretty weak GPU, much less powerful than a Geforce or Quadro. Performance was fine, since I didn't do gaming but only text. (I've since upgraded to Quadro K1200 cards, faster but still hardly monsters.)
Slashdot Mirror
(I had to change the currency symbols because Slashdot still doesn't handle non-ASCII characters correctly.)
Yes and there's no such thing as x86-64 either, Intel marketing decided it was to be called "IA32-T" or some other nonsense. It's best to ignore the vendor's PR names in such things, particularly when they're as inane as "AArch64".
ARM classic or ARM64? (Of course anything is better than x86 family, but that's a low bar to clear.)
Thanks -- so frying implies temperatures higher than water's boiling point. I'd never worked out the difference between "fry" and "saute" until now.
I fry celery in fat all the time when making soup. Chop some onions, grate some celery into a mush, and cook both in olive oil. Then you go on to add the other ingredients. You don't burn or blacken it though, so perhaps nitrosamines cannot form.
You're missing the point. If you can vote at home with your PC or phone, anyone can ask to look over your shoulder while you do it. Or they can bribe or threaten you to let them look. A secret ballot is just that -- entirely secret. Nobody is allowed in the voting booth with you. You cannot prove how you voted, even if you want to. Perhaps by "blockchain voting" they mean you still go to a polling station to cast your vote in secret, but it's all fully verifiable because blockchain. Actually, cryptographers have already come up with systems where you can check your vote was counted without being able to prove what it was (taking home some kind of paper receipt). But I think of all the problems faced by democracy this isn't the most pressing. Those who want to stuff the ballots would just ignore the result of the cryptographic verification, after all, it's too complex for anyone without a PhD to understand (or they could just introduce their own fake 'experts'). (This is also why postal voting is flawed, in my view -- for people who can't attend on election day, they should be able to visit a special polling station a week or two in advance, but not to just mail in a vote which anybody could have "supervised". There have certainly been cases of fraud with postal votes, too.)
https://science.slashdot.org/s...
Sounds to me like "the thin, monotonous whine of blasphemous flutes from inconceivable, unlighted chambers beyond Time".
Sorry, you do have a point, since their project includes run time checking with some performance penalty.
Those are run time checks. They are a great way of smoke-testing your code for memory access errors that can't be detected at compile time (which in standard C is most of them). Of course, while this testing can prove the existence of bugs, it cannot prove their nonexistence. The paper is about a different and complementary approach, compile-time checking. Here you prove that certain types of unsafe access cannot occur based on the source code. The prior art cited in the paper is other approaches of this type.
The pro cards (Quadro, and also some NVS cards) nowadays come with four DisplayPort outputs. The consumer cards tend to have one DisplayPort and a mishmash of HDMI and DVI. That is annoying, since there are converters from DisplayPort to both of these but not the other way round. The GeForce cards also tend to be dual-slot, power-hungry monsters while the lower-end Quadro cards fit in a single slot and don't need auxiliary power.
So I think there are reasons to prefer the Quadro line, apart from the question of "certification" (which does look like like a bit of a racket).
In principle you can license and build a GeForce card with four DisplayPort connectors, but they are very hard to find. The best I was able to buy was a GeForce 1060 with dual DisplayPorts.
Ah, of course you're right. Microsoft ported lots of the classic DOS utilities to native Windows versions, including the command prompt itself (COMMAND.COM became CMD.EXE) but porting EDLIN would have been a step too far. You can blame AMD for not including a way to go back from V86 mode to x86_64 mode, I think.
They're both the same socket, BGA 1440. (It would be crazy to have two different motherboard designs in the same laptop model.)
Recent WIndows versions no longer include EDLIN. It was there in NT 4.0 but is missing in Windows 10, so must have been dropped somewhere along the way.
Ctrl-Alt-Del followed by Enter also locks the screen quickly.
If yours have the detachable cable (SDL connector) you can buy a replacement cable that has USB on the other end. Inside it contains a "Soarer's converter" which is the best PS/2 to USB converter.
I'm curious - I've used Model M and Model F keyboards for, oh, nearly 25 years now and I've never noticed springs starting to "wear". Maybe if you popped the keys off all the time to mess around, you might risk damaging a spring. Otherwise they seem well protected and certainly don't appear to suffer any ill effects from being pinged back and forth thousands of times a day. What's the typical failure?
Well strictly speaking the string handling is more to do with the C standard library than the language. You could use a library providing a sensible string type, and the main restriction imposed by the language would be that literal strings cannot contain the 0 byte. It's dispiriting that after so many decades the C standard library remains stuck without a decent set of string functions. (There are plenty of libraries to choose from but the real value comes in picking one and using it everywhere; it's a social rather than a technical issue.) I read that C++'s STL also lacked a string type and some lobbying was needed to get the standards committee to define one.
Apparently a good way to quieten a Model M (or F) is to insert a short strand of dental floss in the middle of each spring.
I find the ODbL's contract terms problematic. It purports to be a contract, not a licence, and to impose additional restrictions beyond what copyright law allows it to enforce. In my view that makes it non-free. I greatly prefer the style used by the GPL and other free licences: "You are not required to accept this License, since you have not signed it." Then the licence is a grant of permission to do something which would not otherwise be allowed -- so it's legally much more watertight. (If you go to court arguing that the licence is invalid, you are admitting that you don't have any right to copy the work it covers.) The rationale for this is that in some countries map data may not be covered by copyright. This is dubious -- legal research I commissioned from US and English lawyers concluded that while it's not entirely clear-cut, there are certainly strong precedents for map data falling within copyright. In my personal view, I think that even one country decided, as a matter of policy, to exclude map data from the scope of copyright and have it be in the public domain, that is a decision for the legislature of that country, and not something the licence on an open data project should try to block. There are plenty of other licences which can be used for data, require attribution, and don't have the problematic contract terms and overreach of the ODbL. The OSM foundation should dual-license under one of these.
Sometimes I wish that if the optimizer finds something really juicy (like eliminating a dozen lines of code because it can prove they will never be called by assuming that undefined behaviour won't be triggered) it would just refrain from optimizing the code out and instead emit a diagnostic telling the programmer to apply that optimization in the source code. You could then review the code and either say 'gee, the compiler is right, I will delete that whole branch', or 'ouch, that should not be happening... let me find where I am relying on undefined behaviour', or in some cases add an explicit compiler hint to allow the optimizer to take an axe to the code without bothering you further. It would have to be a heuristic, and would produce some false positives (where the optimizer is fine, and should just do its job without chatter), but in the cases it did work it would avoid a lot of acrimony and language-lawyering.
The most valuable company is surely Aramco, or more correctly Saudi Aramco, valued at between 2 and 10 trillion dollars. It still begins with A, more or less. The headline should have talked about the second most valuable *listed* company. Although a small part of Saudi Aramco shares is expected to be listed soon.
I can think of one example, block-sorting file compression as used by bzip2. If you are making a tarball of several files, it will presumably help to have files with similar content next to each other in the archive. (The block-sorting process would rearrange them later to some extent, but I believe that compression will be a bit better if they are already close, particularly if the input data is much bigger than the sorting window.)
So you first sort by some measure that only approximates sorting by content. That is, you sort by filename first. It's a guess that similar filenames will have similar content. That could be completely wrong, of course, but the heuristic tends to work in practice.
(This applies to most forms of compressed archive, even if the compression algorithm doesn't involve sorting, but here you have an example of where it's worth sorting by some approximate key before sorting again.)
Another example may be bulk insertion into a relational database table, where inserting rows in sorted order is fastest, but you may not have the memory to sort them on the client. If you could sort them by something that tends to correlate to the final ordering, you will typically get some speedup in practice.
I have been in a London taxi where the driver had Google Maps open -- not, he explained, to navigate the route, but to show where traffic congestion was.
I don't think it is correct to say that the GPU needs to be four times as powerful for a display with four times as many pixels. For 3d gaming, yes. But for text rendering, I don't think that a higher resolution web page takes four times the processor power to lay out. The same number of character glyphs need to be painted on the screen; they are bigger, but still nothing that challenges even the weakest GPU. The difficult part is rendering the glyphs at the higher resolution, and that is CPU-bound. Sure, a more powerful GPU would help, but it doesn't need to be four times as powerful. As anecdotal evidence I note that for a couple of years I ran two 5k monitors from a single NVS 510 desktop video card. This is a pretty weak GPU, much less powerful than a Geforce or Quadro. Performance was fine, since I didn't do gaming but only text. (I've since upgraded to Quadro K1200 cards, faster but still hardly monsters.)