My CSci teachers would get around this by making their own homework instead of just using what the book gave them
How is this 'getting around' anything? That's how university courses are supposed to work! They're not like school where everyone is teaching the same material, they're supposed to be tailored to the lecturer. That's why you go to a university - i.e. a place where people at the top of their field are conducting original research - so that you can be taught by experts in the material that you're learning. If they're just going to tell you to read a book and do the problems in it then you may as well just sit the exams as an external candidate, get the degree and save yourself 90% of the cost.
The only practical solution to this is to legislate that all elections are State-funded, and any outside money found in any campaign is the end of that campaign
That's also open to abuse, because how do you determine who qualifies for state funding? If the barrier to entry is too high, then you risk keeping the same oligarchy you have (or swapping it for a new one). If it's too low, then people can easily use it to funnel money to their own pockets (e.g. set up a leaflet printing company, nominate 10 candidates that all use you to print their own content-free leaflets, profit). Or they can simply flood the system with generic like-the-opposition candidates. For example, in the UK a few elections ago we had a Literal Democrat candidate, who got enough votes to get their deposit back because people confused him with the Liberal Democrats and split the Liberal Democrat vote just enough to cause them to lose the election to the Conservatives (who weren't running against a Conservatory Party candidate in that election). If you have a constituency that is 60% liberal and 40% conservative and the barrier to entry is sufficiently low, the best strategy for the conservatives is to back half a dozen liberal candidates...
Yep, the libraries have 1 or 2 copies. If everyone didn't buy the books, good luck getting hold of one of those copies
We had classes for about 100. Each had about 5 course texts, but they had overlapping content so you rarely needed to read more than one. Actually, between the lecture notes and the material presented in the lectures, you only needed to read a book if you were aiming to get a first or a decent 2.1 - you could get a 2.2 or just scrape a 2.1 without reading any of the books.
Another scam is that a new edition will be released every 2 years or so, with the difference being a few different problems, maybe some case studies, that kind of stuff
Does this actually matter? The only assessed problems for us were ones set by the lecturer, not ones in a specific book. If any specific cases was going to be required for the exam or coursework then they were covered in lectures and handouts. Textbooks were for more general background.
Except for the profs that wrote the book, then you HAD TO HAVE THE LATEST BOOK!
One of my books is used as the course text for a module in a canadian university. I think I make under £50/year from sales to their students, if that. Even if you get all of your students to buy your book, it's not worth it.
Books? which at four-year public colleges in 2011-12 is $1,168 per year
Seriously? Do US universities get away with this? I didn't have to buy any books for my degree. Anything that was recommended reading was available in the university library, so the only reason to actually buy a copy was if you wanted to be able to refer to it in the future. I did buy a few books, but most of them were from the second-hand shop on campus so cost £5 or under.
There's also the difference in size. The pictures really don't do the Kindle justice - I played with one in an airport shop and they're astonishingly thin and light. I have an iRex iLiad, which is one of the first generation of eInk devices. You could easily fit two or three 6" Kindles inside it. You could probably slip the kindle into a coat pocket without noticing. Doing that with the iLiad or a modern tablet would be difficult. The smallest tablet I own is a Nokia 770, and it fits, but it's much more bulky than a kindle. Most of the difference is the battery - a TFT screen draws a lot more power than eInk so an eInk device can get away with a tiny battery in comparison with a tablet (and a less power-hungry CPU, since you're not going to be playing back videos on the eBook reader).
You're missing the point: if you can get 15% on a credit card, you can almost certainly get under 10% on an unsecured personal loan, and less on one where you put up some collateral. Credit cards are very rarely the cheapest way of borrowing money, so using them to borrow money is generally a bad idea.
I think the grandparent's point was that you should pay off your credit card every month and just use it as you would use a debit card (but with better buyer protection), rather than borrow money on it. I have no idea what the interest rate on my credit card is - about 15% I think - because it is paid automatically from my bank account by direct debit (unless I dispute some charges) every month so I never pay any interest. If I need to borrow money, I will do so via some other mechanism. Actually, I bought my house with an offset mortgage, so I have a fairly low interest line of credit up to the amount I have already 'repaid' (i.e. put in a savings account linked to the mortgage).
There is no advantage to running 64bit LINUX....FTFY
Not true at all. There are several advantages to running code compiled for x86-64 instead of for x86. If you're on *NIX, then one of the big ones is that x86-64 has rip-relative addressing, while x86 lacks eip-relative addressing. In position-independent code, this means that any call to a static function or access to a static variable is cheaper. Windows doesn't use position-independent code for DLLs on 32-bit (it tries to load them at a fixed address and performs load-time relocations if it can't), so that's not a problem, but on *NIX.
There are other advantages. One of the big ones is that anything targeting x86-64 will default to using SSE for floating point instead of x87. This is generally at least slightly faster in the silicon, but it's also much easier to generate good SSE code for the register-register SSE coprocessor (with lots of registers!) than for the stack-based x87 (with only a handful).
Then, of course, there's the fact that you have a lot more general-purpose registers with x86-64, so you need far fewer stack spills to do the same thing.
Not that they shouldn't care, mind you, as that is some seriously monolithic code.
They're talking about using link-time optimisation (LTO). That means that you compile each compilation unit to an intermediate representation and then run optimisations on the whole program. This takes a staggering amount of memory (which is why no one bothered with it off high-end workstations with very expensive compilers until very recently), but can sometimes be worth it. It actually helps modularity, because you can keep your source code nicely separated into independent components without worrying about efficiency, and you can do better data hiding.
As a trivial example, consider an accessor method. Something like getFoo() { return foo; }. Without LTO, you'd want to put the declaration of the class and this method in the header so that every compilation unit could separately inline it. This reduces modularity, but it saves you the cost of a function call just to access a field in a structure. With LTO, you can make the class opaque (if you're a C++ junkie, using the Pimpl pattern, if you're a C programmer by just not declaring the struct in the header). You'll get a single copy of the function emitted, and then the link-time optimiser will inline it everywhere.
Yes it does, but it's horrible. You lose the flat address space and you have to use API calls to manually switch banks of memory in. Windows has been able to do this since the late '90s, but I don't know of any software that has actually made use of it.
And then you get fun things like the Thor DVD that contained a filename with some invalid characters that broke libdvdnav, so nothing open source can play it back (I think there's a work-around in svn). I sent it back as defective when I rented it, with a note saying that their attempt at copy protection made me - a paying customer - unable to watch it (not entirely true - I watched it on my Mac, but that was extra effort).
a 1 hour TV show is NOT WORTH more than 0.99 to most people
I'd say it's not even worth that. The TV I watch is typically on DVDs. I don't remember the last time I watched a 1-hour show - perhaps you mean a 1 US-TV-Hour show, which is 40 minutes. I get four of these on a DVD. For the price of my rental plan, I pay about 10p (about 15) per episode. Online distribution means that they're not paying the cost of shipping me a shiny disk. If I can make a copy and watch it again[1] then that's worth slightly more, but I rarely want to watch a TV show more than once. 20 per episode, or maybe $5 for a season for a DRM-free download would put things in the impulse buy category - and I'd even be prepared to pay it before they even started filming the season for a lot of shows.
GOG.com has learned that this is an incredibly valuable price point. I have about ten games that I've bought from there and not yet played. Most of them cost about $3 on one of their weekend promotions. I bought them because they looked like they might be fun. Some that I've bought like this before have turned out to be quite boring and I've deleted them after 15 minutes. Some have been so much fun I've played them solidly for days. At this price, I'm willing to take a risk - if it's rubbish then I've only lost the price of a pint of beer.
[1] Legally, I mean - doing this with DVDs is trivial, but not legal if they are rented.
Or Smalltalk. And if you want Smalltalk to scale, VMWare bought Gemstone fairly recently and will sell you a deployment where you can have as many Smalltalk images as you want all using transactional memory to share state and happily scaling to a few million concurrent instances without any significant effort.
Is it? PHP has two problems: it makes it easy to write huge spaghetti projects that are expensive to maintain, and it's slow. If you make it fast, all that you do is encourage people to write even more spaghetti projects that are expensive to maintain. I suppose it's a good thing if you're a consultant...
No, you must use your product to advertise theirs, you can't use theirs to advertise yours. It's the same as the old BSD license. You must include 'built using Zend's stuff' somewhere in your documentation, but you can't say 'really amazing because it uses Zend's stuff'.
Depends on how it's distributed. If it's a perfectly even distribution, or skewed slightly south, then some parts of Canada may become suitable for human habitation. If it's focussed more in the north, then expect large amounts of flooding as the snow and ice melt. Unfortunately, an average increase is likely to mean some places getting significantly warmer and others getting significantly colder - if you pump a huge amount of energy into a chaotic system, the results are (by definition) difficult to predict...
A better approach would be to move something like Cairo to the server, so you could upload bezier paths and reference them by name (actually, just arbitrary snippets of PDF data would be ideal, but then you've practically reinvented Quartz). The problem with uploading 'fonts' is that you need to specify a font format. Most fonts these days will be in TrueType format, or maybe OpenType format, but PDFs embed PostScript fonts (type 1, 2, or 3...), DVIs use MetaFont, and so on. You're requiring the display server to have a complex font parser (see the recent Windows security hole and various past *NIX ones for why this is a bad idea) and you're limiting the kind of fonts that can supported.
If you just provide the glyphs to the server and then composite them there then you only use a small amount of bandwidth to render each one. The current approach means that you need to send them again if you render at a different size, but that's not a huge use of bandwidth - you're still typically only sending a couple of dozen monochrome glyph masks per font / size. Sending them as a simple vector representation (e.g. PostScript bezier drawing commands) would be more efficient, but not by much.
That said, if I were redesigning X11, I'd replace the entire drawing system with something that let you send PDF objects that would be referenced by a unique ID (as they are in a PDF file, in fact) and composited wherever they were required.
The best fix would be to adopt an X12 protocol which puts proper font rendering back in the server
X11 already has proper font rendering on the server via the XRender extension. The server handles the rendering, the client provides the glyphs. This is superior to the old approach of making the server provide the glyphs because it means that you can install an application + fonts on a machine and use it anywhere without having to install the fonts on every X server (which is hard if they're dumb terminals) and means that you can do things like embed fonts in documents.
The second is flexibility in window managers
And compositing managers. Wayland uses a one-size-fits all approach to both, while X11 is modular.
Agreed, giving away everything is probably not the wisest move
It can be. Software itself is not valuable, being able to use software to solve business problems is valuable. Most off-the-shelf software does not actually do what the business wants. There is, for example, a huge secondary market on top of Microsoft Office building things like complex templates and wizards for Word, business-specific database applications, and so on. This is part of why OpenOffice has such a hard time competing: in a big company the cost of the MS Office licenses can be quite small compared to the cost of the customisation, and even if they were paid to use OpenOffice the cost of rewriting all of their custom stuff would make MS Office cheaper.
This is even more true for open source software. It changes the economics back to a more rational basis. Rather than doing the difficult bit (writing the software) for free and charging for the easy bit (copying the software), you charge for the difficult bit and let people do the easy bit for free. Companies like Asterisk make a lot of money by giving everything away for free and getting people to pay for the bits that don't exist yet. The fact that their product is Free Software and they have competitors able to sell exactly the same thing even works in their favour, because they can offer a second source and say to their customers 'if you're not happy with the job we do, there's no lock in and there are a load of other companies who would be happy to have your business.'
Red Hat gives away their core product, but people pay so that the bugs or missing features that affect them get higher priority.
If an upgrade destroys user's data, then it's a bug. Removing the ability to set the confirmation status of a new appointment is a design decision. Stripping this data from existing appointments is a data-loss causing bug. Raskin's First Law: a program shall not harm a user's data, nor through inaction allow a user's data to come to harm. Any violation of this is a bug.
If he's talking about ordinals then he's referring to the largest value that the octet can store, so if it starts at 0 then this would be increased from 255 to 256. If he's talking about cardinals then he's referring to the range of values that it can store, so this increases from 256 to 257. The phrase 'goes to 257' could be interpreted in either way - it either stores values up to 257 or it stores up to 257 unique values.
I agree. 10.5 removed the ability to mark appointments as tentative / confirmed and, worse, removed this data from existing appointments. I filed a bug, marked it as 'data loss' and had it closed with 'works as expected'. I gave up filing Apple bugs around then.
Slashdot Mirror
My CSci teachers would get around this by making their own homework instead of just using what the book gave them
How is this 'getting around' anything? That's how university courses are supposed to work! They're not like school where everyone is teaching the same material, they're supposed to be tailored to the lecturer. That's why you go to a university - i.e. a place where people at the top of their field are conducting original research - so that you can be taught by experts in the material that you're learning. If they're just going to tell you to read a book and do the problems in it then you may as well just sit the exams as an external candidate, get the degree and save yourself 90% of the cost.
Law is a bit of a special case. The books that you buy during a law degree are usually the books that every lawyer needs to own for regular reference.
The only practical solution to this is to legislate that all elections are State-funded, and any outside money found in any campaign is the end of that campaign
That's also open to abuse, because how do you determine who qualifies for state funding? If the barrier to entry is too high, then you risk keeping the same oligarchy you have (or swapping it for a new one). If it's too low, then people can easily use it to funnel money to their own pockets (e.g. set up a leaflet printing company, nominate 10 candidates that all use you to print their own content-free leaflets, profit). Or they can simply flood the system with generic like-the-opposition candidates. For example, in the UK a few elections ago we had a Literal Democrat candidate, who got enough votes to get their deposit back because people confused him with the Liberal Democrats and split the Liberal Democrat vote just enough to cause them to lose the election to the Conservatives (who weren't running against a Conservatory Party candidate in that election). If you have a constituency that is 60% liberal and 40% conservative and the barrier to entry is sufficiently low, the best strategy for the conservatives is to back half a dozen liberal candidates...
Yep, the libraries have 1 or 2 copies. If everyone didn't buy the books, good luck getting hold of one of those copies
We had classes for about 100. Each had about 5 course texts, but they had overlapping content so you rarely needed to read more than one. Actually, between the lecture notes and the material presented in the lectures, you only needed to read a book if you were aiming to get a first or a decent 2.1 - you could get a 2.2 or just scrape a 2.1 without reading any of the books.
Another scam is that a new edition will be released every 2 years or so, with the difference being a few different problems, maybe some case studies, that kind of stuff
Does this actually matter? The only assessed problems for us were ones set by the lecturer, not ones in a specific book. If any specific cases was going to be required for the exam or coursework then they were covered in lectures and handouts. Textbooks were for more general background.
Except for the profs that wrote the book, then you HAD TO HAVE THE LATEST BOOK!
One of my books is used as the course text for a module in a canadian university. I think I make under £50/year from sales to their students, if that. Even if you get all of your students to buy your book, it's not worth it.
Books? which at four-year public colleges in 2011-12 is $1,168 per year
Seriously? Do US universities get away with this? I didn't have to buy any books for my degree. Anything that was recommended reading was available in the university library, so the only reason to actually buy a copy was if you wanted to be able to refer to it in the future. I did buy a few books, but most of them were from the second-hand shop on campus so cost £5 or under.
There's also the difference in size. The pictures really don't do the Kindle justice - I played with one in an airport shop and they're astonishingly thin and light. I have an iRex iLiad, which is one of the first generation of eInk devices. You could easily fit two or three 6" Kindles inside it. You could probably slip the kindle into a coat pocket without noticing. Doing that with the iLiad or a modern tablet would be difficult. The smallest tablet I own is a Nokia 770, and it fits, but it's much more bulky than a kindle. Most of the difference is the battery - a TFT screen draws a lot more power than eInk so an eInk device can get away with a tiny battery in comparison with a tablet (and a less power-hungry CPU, since you're not going to be playing back videos on the eBook reader).
You're missing the point: if you can get 15% on a credit card, you can almost certainly get under 10% on an unsecured personal loan, and less on one where you put up some collateral. Credit cards are very rarely the cheapest way of borrowing money, so using them to borrow money is generally a bad idea.
I think the grandparent's point was that you should pay off your credit card every month and just use it as you would use a debit card (but with better buyer protection), rather than borrow money on it. I have no idea what the interest rate on my credit card is - about 15% I think - because it is paid automatically from my bank account by direct debit (unless I dispute some charges) every month so I never pay any interest. If I need to borrow money, I will do so via some other mechanism. Actually, I bought my house with an offset mortgage, so I have a fairly low interest line of credit up to the amount I have already 'repaid' (i.e. put in a savings account linked to the mortgage).
There is no advantage to running 64bit LINUX....FTFY
Not true at all. There are several advantages to running code compiled for x86-64 instead of for x86. If you're on *NIX, then one of the big ones is that x86-64 has rip-relative addressing, while x86 lacks eip-relative addressing. In position-independent code, this means that any call to a static function or access to a static variable is cheaper. Windows doesn't use position-independent code for DLLs on 32-bit (it tries to load them at a fixed address and performs load-time relocations if it can't), so that's not a problem, but on *NIX.
There are other advantages. One of the big ones is that anything targeting x86-64 will default to using SSE for floating point instead of x87. This is generally at least slightly faster in the silicon, but it's also much easier to generate good SSE code for the register-register SSE coprocessor (with lots of registers!) than for the stack-based x87 (with only a handful).
Then, of course, there's the fact that you have a lot more general-purpose registers with x86-64, so you need far fewer stack spills to do the same thing.
Not that they shouldn't care, mind you, as that is some seriously monolithic code.
They're talking about using link-time optimisation (LTO). That means that you compile each compilation unit to an intermediate representation and then run optimisations on the whole program. This takes a staggering amount of memory (which is why no one bothered with it off high-end workstations with very expensive compilers until very recently), but can sometimes be worth it. It actually helps modularity, because you can keep your source code nicely separated into independent components without worrying about efficiency, and you can do better data hiding.
As a trivial example, consider an accessor method. Something like getFoo() { return foo; }. Without LTO, you'd want to put the declaration of the class and this method in the header so that every compilation unit could separately inline it. This reduces modularity, but it saves you the cost of a function call just to access a field in a structure. With LTO, you can make the class opaque (if you're a C++ junkie, using the Pimpl pattern, if you're a C programmer by just not declaring the struct in the header). You'll get a single copy of the function emitted, and then the link-time optimiser will inline it everywhere.
Yes it does, but it's horrible. You lose the flat address space and you have to use API calls to manually switch banks of memory in. Windows has been able to do this since the late '90s, but I don't know of any software that has actually made use of it.
And then you get fun things like the Thor DVD that contained a filename with some invalid characters that broke libdvdnav, so nothing open source can play it back (I think there's a work-around in svn). I sent it back as defective when I rented it, with a note saying that their attempt at copy protection made me - a paying customer - unable to watch it (not entirely true - I watched it on my Mac, but that was extra effort).
a 1 hour TV show is NOT WORTH more than 0.99 to most people
I'd say it's not even worth that. The TV I watch is typically on DVDs. I don't remember the last time I watched a 1-hour show - perhaps you mean a 1 US-TV-Hour show, which is 40 minutes. I get four of these on a DVD. For the price of my rental plan, I pay about 10p (about 15) per episode. Online distribution means that they're not paying the cost of shipping me a shiny disk. If I can make a copy and watch it again[1] then that's worth slightly more, but I rarely want to watch a TV show more than once. 20 per episode, or maybe $5 for a season for a DRM-free download would put things in the impulse buy category - and I'd even be prepared to pay it before they even started filming the season for a lot of shows.
GOG.com has learned that this is an incredibly valuable price point. I have about ten games that I've bought from there and not yet played. Most of them cost about $3 on one of their weekend promotions. I bought them because they looked like they might be fun. Some that I've bought like this before have turned out to be quite boring and I've deleted them after 15 minutes. Some have been so much fun I've played them solidly for days. At this price, I'm willing to take a risk - if it's rubbish then I've only lost the price of a pint of beer.
[1] Legally, I mean - doing this with DVDs is trivial, but not legal if they are rented.
Or Smalltalk. And if you want Smalltalk to scale, VMWare bought Gemstone fairly recently and will sell you a deployment where you can have as many Smalltalk images as you want all using transactional memory to share state and happily scaling to a few million concurrent instances without any significant effort.
And a JIT for php is a good thing
Is it? PHP has two problems: it makes it easy to write huge spaghetti projects that are expensive to maintain, and it's slow. If you make it fast, all that you do is encourage people to write even more spaghetti projects that are expensive to maintain. I suppose it's a good thing if you're a consultant...
No, you must use your product to advertise theirs, you can't use theirs to advertise yours. It's the same as the old BSD license. You must include 'built using Zend's stuff' somewhere in your documentation, but you can't say 'really amazing because it uses Zend's stuff'.
Depends on how it's distributed. If it's a perfectly even distribution, or skewed slightly south, then some parts of Canada may become suitable for human habitation. If it's focussed more in the north, then expect large amounts of flooding as the snow and ice melt. Unfortunately, an average increase is likely to mean some places getting significantly warmer and others getting significantly colder - if you pump a huge amount of energy into a chaotic system, the results are (by definition) difficult to predict...
A better approach would be to move something like Cairo to the server, so you could upload bezier paths and reference them by name (actually, just arbitrary snippets of PDF data would be ideal, but then you've practically reinvented Quartz). The problem with uploading 'fonts' is that you need to specify a font format. Most fonts these days will be in TrueType format, or maybe OpenType format, but PDFs embed PostScript fonts (type 1, 2, or 3...), DVIs use MetaFont, and so on. You're requiring the display server to have a complex font parser (see the recent Windows security hole and various past *NIX ones for why this is a bad idea) and you're limiting the kind of fonts that can supported.
If you just provide the glyphs to the server and then composite them there then you only use a small amount of bandwidth to render each one. The current approach means that you need to send them again if you render at a different size, but that's not a huge use of bandwidth - you're still typically only sending a couple of dozen monochrome glyph masks per font / size. Sending them as a simple vector representation (e.g. PostScript bezier drawing commands) would be more efficient, but not by much.
That said, if I were redesigning X11, I'd replace the entire drawing system with something that let you send PDF objects that would be referenced by a unique ID (as they are in a PDF file, in fact) and composited wherever they were required.
"Getting laid" simply doesn't take that much time.
In that case, you're probably doing it wrong...
The best fix would be to adopt an X12 protocol which puts proper font rendering back in the server
X11 already has proper font rendering on the server via the XRender extension. The server handles the rendering, the client provides the glyphs. This is superior to the old approach of making the server provide the glyphs because it means that you can install an application + fonts on a machine and use it anywhere without having to install the fonts on every X server (which is hard if they're dumb terminals) and means that you can do things like embed fonts in documents.
The second is flexibility in window managers
And compositing managers. Wayland uses a one-size-fits all approach to both, while X11 is modular.
Agreed, giving away everything is probably not the wisest move
It can be. Software itself is not valuable, being able to use software to solve business problems is valuable. Most off-the-shelf software does not actually do what the business wants. There is, for example, a huge secondary market on top of Microsoft Office building things like complex templates and wizards for Word, business-specific database applications, and so on. This is part of why OpenOffice has such a hard time competing: in a big company the cost of the MS Office licenses can be quite small compared to the cost of the customisation, and even if they were paid to use OpenOffice the cost of rewriting all of their custom stuff would make MS Office cheaper.
This is even more true for open source software. It changes the economics back to a more rational basis. Rather than doing the difficult bit (writing the software) for free and charging for the easy bit (copying the software), you charge for the difficult bit and let people do the easy bit for free. Companies like Asterisk make a lot of money by giving everything away for free and getting people to pay for the bits that don't exist yet. The fact that their product is Free Software and they have competitors able to sell exactly the same thing even works in their favour, because they can offer a second source and say to their customers 'if you're not happy with the job we do, there's no lock in and there are a load of other companies who would be happy to have your business.'
Red Hat gives away their core product, but people pay so that the bugs or missing features that affect them get higher priority.
You could, but a more logical thing to do would be enjoy some of the less risky things first...
If an upgrade destroys user's data, then it's a bug. Removing the ability to set the confirmation status of a new appointment is a design decision. Stripping this data from existing appointments is a data-loss causing bug. Raskin's First Law: a program shall not harm a user's data, nor through inaction allow a user's data to come to harm. Any violation of this is a bug.
If he's talking about ordinals then he's referring to the largest value that the octet can store, so if it starts at 0 then this would be increased from 255 to 256. If he's talking about cardinals then he's referring to the range of values that it can store, so this increases from 256 to 257. The phrase 'goes to 257' could be interpreted in either way - it either stores values up to 257 or it stores up to 257 unique values.
I agree. 10.5 removed the ability to mark appointments as tentative / confirmed and, worse, removed this data from existing appointments. I filed a bug, marked it as 'data loss' and had it closed with 'works as expected'. I gave up filing Apple bugs around then.