You're not going to do a 1000 pages technical document in 3 different versions, released in 6 languages, and authored by 4 tech writers with InDesign.
Now, it would brain-dead to use FrameMaker to create your 16 pages flashy marketing magazine & 1 page foldable product flyers, in PDF and interactive web version...
I was about to tell you that afaik, circa 2006, framemaker was a dead product, but I checked and, hooray, Adobe changed its mind and released new versions. Great news.
Framemaker is the way to go for tech doc, IMO. I was so sad that Adobe was letting the product dying...
They'll do their capital gain via tax-less offshore companies, and own shares in those companies.
Also, I guess it would be possible to build a derivative product that will have a long term gain correlated to a specific short-term gain of some stock.
The document is fascinating, and quite easy to read. I had to go to page 18 to find the first mention of an exposure that wasn't looking like an insider trade (ie: lending money to people that owned the bank).
Indeed, the easiest way to rob a bank it to own one...
I doubt that high prices in France come from EU regulation. They come from high taxes, and by the fact that a lot of big retail stores are colluding to force high prices without troubles (due to the fact that they are very close to the govt. For instance, the hadopi law was designed by the head of the biggest music retailer).
Almost every part of the article is filled with inaccuracies.
Quake BSP pre-processing "would identify empty space inside and outside of the border. This highly effective technique reduced the amount of polygons usually by half and sometimes by much more." That is not what binary space partitioning is about.
It is extremely obvious that the guy writing the article have zero clues about how a 3d engine work. For instance, he states that, with Renderware "a developer could, [..] change the color of a character's clothing without altering the underlying code and rendering the scene all over again". As if the scene isn't rendered each frame anyway...
There is some seriously severe misunderstanding in there...
Thx for the link. Too bad the video is powered by "The video you are trying to watch is currently unavailable" brightcove.com flash player of doom... Never had that piece of shit display a single video on firefox, nor give any adequate error message...
> I looked at both of these, and none of the (bad) cases they list apply to documenting hardware.
So maybe you should read them again. From the very page linked by the grand-parent:
Wikipedia is not an indiscriminate collection of information => See Also Notability
General notability guideline => Topic have "Significant Coverage"
"Significant coverage" means that sources address the subject directly in detail, and no original research is needed to extract the content. Significant coverage is more than trivial but may be less than exclusive.[1]"
How can you argue that original research aiming at listing the connections of every obscure hardware connector under the sun should be in wikipedia ?
Btw, you say "I often use Wikipedia's pages on Unicode encoding, do you also mean that they should be removed?". Well, original poster never said or implied that Unicode had to be removed, and by doing so, you are using Schopenhauer's stratagem 24, the false syllogism (look it up, or read it at http://coolhaus.de/art-of-controversy/erist24.htm ).
Ok, I have moderated this thread, so my mod will be lost, but I just can't avoid replying. If that's trolling, well done:-)
> You are given the power to abuse something and you think nobody will notice....so you do
No, you don't.
I've been in the power of abusing for years, and I have never done it. It never even crossed my mind. When I had my own company I had countless opportunity to slightly (or grossly) abuse, and never did. I never ever had the company reimburse stuff like my subway tickets, or meals that weren't clearly business meals (ie: with prospects or customers).
I've seen a lot of people abusing the system for years and years. We had expenses in euros, but reimbursements in pounds, and I could see people, working for me, that were carefully choosing the exchange rate they could get away with, and being proud of that (because it was tolerated that you could use a slightly higher exchange rate to compensate the fact that your bank add fees when you pay in foreign country).
You may argue that that was because it was my own company, but I also worked for other people and never ever did any sort of reimbursement fee abuse.
And, while I've seen people abusing, I've also seen a lot of people deeply honest.
So, no, it is not because you can take company furniture to your home that you should. It is not because you can come late and leave early that you should. It is not because you can do personal web surfing at work that you should. It is not because you can telecommute and slack off that you should.
Lately, I've seen this "you are stupid if you don't abuse the system" mentally creeping in higher and higher levels of the society due to the fact that we blindly reward success without regard to any other metric. I believe that it is the root of most of the world problems.
It is not because you are given the power to wreck the world economy for you profit and get away with it that you should.
> I believe that is exactly the wrong strategy, if 15/16 4KB 'sectors' in a 256KB block are TRIMMED you would NOT want to use that block as the one to overwrite because it will soon be freed and can be background erased.
(Note: there is not 16 4KiB pages in 256K, but 64)
Depends on what the metric for success is, in particular how do you weight in the cost of wear versus the write performance.
The downsides of your strategy are as follow:
* You will wear the disk a bit faster (you are going to ERASE blocks a bit more often, as the block you are erasing will contain less free pages) * You will get a worse throughput if you write a big file, because you are going to ERASE more blocks, with no TRIM intervening.
On the upside, you maybe get a better throughput if there are a lot of simultaneous write/delete on a disk with mostly free space.
As I said in my original post, you can have a lot of different algorithms. For instance, you can also relocate a page that is the only one in its block into a free page, to be able to asynchronously ERASE the block. You'll add a little wear, but will always have blocks available... You could use a dynamic watermark under which you relocate content and ERASE the block. Everything is possible, with very different performance metrics on different benchmarks.
Once upon a time, a technical subject on/. gave insightful and informative responses that were modded up. Time changes, I guess.
The "fragmentation" that SSD drive have don't really come from wear leveling, or from having to find some place to write things, but from the following properties:
* Filesystems read and write 4KiB pages. * SSD can read many time 4KiB pages FAST, can write ONCE 4KiB pages FAST, but can only erase a whole 512KiB blocks SLOWLY.
When the drive is mostly empty, the SSD have no trouble finding blanks area to store the 4KiB write from the OS (he can even cheat with wear leveling to re-locate 4K pages to blank spaces when the OS re-write the same block). After some usage, ALL THE DRIVE HAVE BEEN WRITTEN TO ONCE. From the point of view of the SSD all the disk is full. From the point of view of the filesystem, there is unallocated space (for instance, space occupied for files that have been deleted).
At this point, when the OS send a write command to a specific page, the SSD is forced to to the following:
* read the 512KiB block that contain the page * erase the block (SLOW) * modify the page * write back the 512KiB block
Of course, various kludges/caches are used to limit the issue, but the end result is here: writes are getting slow, and small writes are getting very slow.
The TRIM command is a command that tell the SSD drive that some 4KiB page can be safely erased (because it contains data from a delete file, for instance), and the SSD stores a map of the TRIM status of each page.
Then the SSD can do one of the following two things:
* If all the pages of a block are TRIMed, it can asynchronously erase the block. So, the next 4KiB write can be relocated to that block with free space, and also the 127 next 4KiB writes. * If a write request come and there is no space to write data to, the drive can READ/ERASE/MODIFY/WRITE the block with most TRIMed space, which will speed up the next few writes. (of course, you can have more complex algorithms to pre-erase at the cost of additional wear)
Security patches are good. For instance, java has a remote execution issue that is 5 month old. See this blog
In that page, you have a link [BEWARE, DON'T CLICK], which will execute arbitrary code (the guy says it is harmless, I believe him, but you don't have to), on your fully-patched, up-to-date OSX. I checked it, it works.
So, well, I, for one, guess that a high bug fix list is a good thing. I wish that Apple fix list count was one higher.
> So if you want greatsite.com but thats taken then register blah.com
You obviously haven't even tried to register a domain name. All 4 letters domain were already gone maybe 10 years ago (there was a/. article on that), and getting a barely pronounceable 5 letters domain is really really really difficult.
The issue with the cram game, is that Ia m pretty sure I can code it in C and have it faster than the Mathematica version, so that would make it a bad example for the point at hand.
The way I understood the position of the original poster is that C can be faster than every other (higher) language, and that the only reason to choose some language over C is because the source code will be smaller / more maintainable (ie: the impedance mismatch with the problem domain will be lower).
The point I am trying to make is that, if the chosen language matches a complex problem domain, then the only solution would be to re-implement everything in C and tune it to hell to match the "higher language" implementation.
Of course, my example were extreme, as they are basically library calls, but we can do some "more realistic" examples (but at the end, everything will boil down to explicit or implicit library calls):
1) A unification function. Here is some full LISP code for unification (http://web.cecs.pdx.edu/~mperkows/CLASS_ROBOTICS/LISP/tanimoto/UNIFY.CL)
Here is someone posting some "similar" C code. (http://coding.derkeiler.com/Archive/Prolog/comp.lang.prolog/2006-04/msg00084.html)
Unfortunately, as said in the post, the C code lacks variable handling which is "for free" in the lisp version. Oops. To get that working, you basically have to implement a chunk of lisp. And if you want your unification routine to outperform the lisp one on every input, you may have to get smart in memory copy (no duplication of nodes during recursion) and smart in memory usage (deleting unneeded node when necessary). This means a garbage collector. Oops again.
At the end, you'll end up with a full lisp interpreter in your C code.
2) Software using self-modifying code.
LISP, SmallTalk, Javascript or even Java, have the possibility of adding code to a running program. If we take JIT versions of LISP or Java, we can build code that evaluate function with a much better performance than any C parser / interpreter, because it will be JITed. The only solution in C would be to use something like LLVM, but I think that using such huge library would be cheating.
3) A scheme interpreter. This isn't very difficult to write in a LISP dialect, and extremely hard to do in C, for obvious reasons.
Of course, all those examples will seem weird and contrived to a C coder, precisely because it doesn't match C problem domain. Which was my original point.
The challenge is, verbatim: "And that is my challenge: one, relatively small app written in any higher language will be faster when written in C".
So, you think to think that, because FactorInteger is coded in C, it should not be used. Then what about that simple program ?
G[f_]:= Integrate[f, x]
You pass it a function f, and it integrates it. And no, Mathematica's Integrate is NOT written in C.
Or maybe I can't do that, because I relie on "good math libraries" ? Rules out an option pricer, then. And I guess I cannot post any Prolog program that uses unification too ? Because that is hard to do in C.
If one morph the challenge into "write a simple app that don't take advantage of anything that would be long to reproduce from scratch with a C compiler", OF COURSE, it can be rewritten in C, even if it is a bit longer and unmaintainable.
F gives the number of distinct prime factors in x. For instance F[3] is 1, F[4] is 1 too (because 4 only have 2 as a factor] and F[12322] is 3 (12322=2*61*101)
This is a small app. Write it in C faster than Mathematica.
To give you a benchmark, an old Core2Duo take a little more than 2 seconds to answer to F[33622711352843762413542039873684935021017]
> Firefox should be a little better at sandboxing plugins, but they can't be blamed for Adobe's crap.
Yes they can. They should do whatever is necessary to avoid leaving a ghost process that have to be manually killed. I had to explain to my 9 year old why he couldn't launch a browser on his ubuntu eeepc. It seems clear to him that the windows machine he uses for games is less buggy than his linux machine.
Your argument is bullshit, as it can be used to prtoect almost any release of information. For instance, I could argue that images of nazi camps should have never been released, as it dishearten the family of the guards.
Obama said he would release the images. He didn't. He either lied or flipped. No excuses.
Stalker is quite a good adaptation of the already good novel "The Roadside Picnic".... ("good adaptation" being defined as "absolutely mind blowingly extraordinary superior")
> we would be so saturated in $1 DVDs in discount stores
I hope not. I hope we would be saturated with 20c legal downloads of good quality DiVX. Or 3$/month all-five-year-old-movies-you-can see subscription plans.
It would be so easy to find interesting stuff that nobody would even bother to "pirate" anymore. Fan of 50's SciFi movies ? Here are a few thousands, for your viewing pleasure. Fan of wildlife documentaries ? Here a few other thousands.
You would not pay for the privilege of downloading, but for the privilege of not having to mess around with files, and stuff. Just point, click and watch. Right now, I know many people that download because they have no other way to get their hands on new material that they may or may not like.
In such an ecosystem, new work could be available on some sort of premium plans, but, as it will have to compete with older good stuff, it would have to be reasonably priced. And yes, it means that it'll start to be thought to make 300M$ movies. But it'll be easier to make 1M$ ones.
Of course, there is zero chance to ever see such thing implemented: first the copyright of the stuff since 1923 will not retroactively expire. Second, the morons hijacking the culture in the name of "entertainment industry" are currently selling mp3 songs at 1.29$, which should be the digital price of the whole album.
Oh, I get it, you didn't touched the original install. Sorry, misread.
Well, I almost wanted to do that (but by booting a rescuecd and dd'ing the partitions over to some server), but, at the end, I figured out that, if I have a problem, I would just blank the disks before sending the machine back (if I can).
I don't think they can legally void the hardware warranty just because of some software installation. At least, not where I live.
Just glanced through your linked article. I installed ubuntu on a eeepc 900, and it was painless (I used the eeepc distrib, and then upgraded to more mainstream ubuntu)
On the netbook-remix interface, I removed the "maximizer" program from the session, and now windows aren't covering all the desktop, and the box is roughly usable.
Slashdot Mirror
You're not going to do a 1000 pages technical document in 3 different versions, released in 6 languages, and authored by 4 tech writers with InDesign.
Now, it would brain-dead to use FrameMaker to create your 16 pages flashy marketing magazine & 1 page foldable product flyers, in PDF and interactive web version...
> By no means is it perfect. (It's from Microsoft, after all.)
Of course, "Word" & "Perfect" can't be from Microsoft.
It was from SSI, then WordPerfect Corporation, and then Corel. /pun
I was about to tell you that afaik, circa 2006, framemaker was a dead product, but I checked and, hooray, Adobe changed its mind and released new versions. Great news.
Framemaker is the way to go for tech doc, IMO. I was so sad that Adobe was letting the product dying...
They'll do their capital gain via tax-less offshore companies, and own shares in those companies.
Also, I guess it would be possible to build a derivative product that will have a long term gain correlated to a specific short-term gain of some stock.
The document is fascinating, and quite easy to read. I had to go to page 18 to find the first mention of an exposure that wasn't looking like an insider trade (ie: lending money to people that owned the bank).
Indeed, the easiest way to rob a bank it to own one...
I doubt that high prices in France come from EU regulation. They come from high taxes, and by the fact that a lot of big retail stores are colluding to force high prices without troubles (due to the fact that they are very close to the govt. For instance, the hadopi law was designed by the head of the biggest music retailer).
Almost every part of the article is filled with inaccuracies.
Quake BSP pre-processing "would identify empty space inside and outside of the border. This highly effective technique reduced the amount of polygons usually by half and sometimes by much more." That is not what binary space partitioning is about.
It is extremely obvious that the guy writing the article have zero clues about how a 3d engine work. For instance, he states that, with Renderware "a developer could, [..] change the color of a character's clothing without altering the underlying code and rendering the scene all over again". As if the scene isn't rendered each frame anyway...
There is some seriously severe misunderstanding in there...
Thx for the link. Too bad the video is powered by "The video you are trying to watch is currently unavailable" brightcove.com flash player of doom... Never had that piece of shit display a single video on firefox, nor give any adequate error message...
> I looked at both of these, and none of the (bad) cases they list apply to documenting hardware.
So maybe you should read them again. From the very page linked by the grand-parent:
Wikipedia is not an indiscriminate collection of information => See Also Notability
General notability guideline => Topic have "Significant Coverage"
"Significant coverage" means that sources address the subject directly in detail, and no original research is needed to extract the content. Significant coverage is more than trivial but may be less than exclusive.[1]"
How can you argue that original research aiming at listing the connections of every obscure hardware connector under the sun should be in wikipedia ?
Btw, you say "I often use Wikipedia's pages on Unicode encoding, do you also mean that they should be removed?". Well, original poster never said or implied that Unicode had to be removed, and by doing so, you are using Schopenhauer's stratagem 24, the false syllogism (look it up, or read it at http://coolhaus.de/art-of-controversy/erist24.htm ).
Ok, I have moderated this thread, so my mod will be lost, but I just can't avoid replying. If that's trolling, well done :-)
> You are given the power to abuse something and you think nobody will notice....so you do
No, you don't.
I've been in the power of abusing for years, and I have never done it. It never even crossed my mind. When I had my own company I had countless opportunity to slightly (or grossly) abuse, and never did. I never ever had the company reimburse stuff like my subway tickets, or meals that weren't clearly business meals (ie: with prospects or customers).
I've seen a lot of people abusing the system for years and years. We had expenses in euros, but reimbursements in pounds, and I could see people, working for me, that were carefully choosing the exchange rate they could get away with, and being proud of that (because it was tolerated that you could use a slightly higher exchange rate to compensate the fact that your bank add fees when you pay in foreign country).
You may argue that that was because it was my own company, but I also worked for other people and never ever did any sort of reimbursement fee abuse.
And, while I've seen people abusing, I've also seen a lot of people deeply honest.
So, no, it is not because you can take company furniture to your home that you should. It is not because you can come late and leave early that you should. It is not because you can do personal web surfing at work that you should. It is not because you can telecommute and slack off that you should.
Lately, I've seen this "you are stupid if you don't abuse the system" mentally creeping in higher and higher levels of the society due to the fact that we blindly reward success without regard to any other metric. I believe that it is the root of most of the world problems.
It is not because you are given the power to wreck the world economy for you profit and get away with it that you should.
> I believe that is exactly the wrong strategy, if 15/16 4KB 'sectors' in a 256KB block are TRIMMED you would NOT want to use that block as the one to overwrite because it will soon be freed and can be background erased.
(Note: there is not 16 4KiB pages in 256K, but 64)
Depends on what the metric for success is, in particular how do you weight in the cost of wear versus the write performance.
The downsides of your strategy are as follow:
* You will wear the disk a bit faster (you are going to ERASE blocks a bit more often, as the block you are erasing will contain less free pages)
* You will get a worse throughput if you write a big file, because you are going to ERASE more blocks, with no TRIM intervening.
On the upside, you maybe get a better throughput if there are a lot of simultaneous write/delete on a disk with mostly free space.
As I said in my original post, you can have a lot of different algorithms. For instance, you can also relocate a page that is the only one in its block into a free page, to be able to asynchronously ERASE the block. You'll add a little wear, but will always have blocks available... You could use a dynamic watermark under which you relocate content and ERASE the block. Everything is possible, with very different performance metrics on different benchmarks.
Once upon a time, a technical subject on /. gave insightful and informative responses that were modded up. Time changes, I guess.
The "fragmentation" that SSD drive have don't really come from wear leveling, or from having to find some place to write things, but from the following properties:
* Filesystems read and write 4KiB pages.
* SSD can read many time 4KiB pages FAST, can write ONCE 4KiB pages FAST, but can only erase a whole 512KiB blocks SLOWLY.
When the drive is mostly empty, the SSD have no trouble finding blanks area to store the 4KiB write from the OS (he can even cheat with wear leveling to re-locate 4K pages to blank spaces when the OS re-write the same block). After some usage, ALL THE DRIVE HAVE BEEN WRITTEN TO ONCE. From the point of view of the SSD all the disk is full. From the point of view of the filesystem, there is unallocated space (for instance, space occupied for files that have been deleted).
At this point, when the OS send a write command to a specific page, the SSD is forced to to the following:
* read the 512KiB block that contain the page
* erase the block (SLOW)
* modify the page
* write back the 512KiB block
Of course, various kludges/caches are used to limit the issue, but the end result is here: writes are getting slow, and small writes are getting very slow.
The TRIM command is a command that tell the SSD drive that some 4KiB page can be safely erased (because it contains data from a delete file, for instance), and the SSD stores a map of the TRIM status of each page.
Then the SSD can do one of the following two things:
* If all the pages of a block are TRIMed, it can asynchronously erase the block. So, the next 4KiB write can be relocated to that block with free space, and also the 127 next 4KiB writes.
* If a write request come and there is no space to write data to, the drive can READ/ERASE/MODIFY/WRITE the block with most TRIMed space, which will speed up the next few writes.
(of course, you can have more complex algorithms to pre-erase at the cost of additional wear)
Security patches are good. For instance, java has a remote execution issue that is 5 month old. See this blog
In that page, you have a link [BEWARE, DON'T CLICK], which will execute arbitrary code (the guy says it is harmless, I believe him, but you don't have to), on your fully-patched, up-to-date OSX. I checked it, it works.
So, well, I, for one, guess that a high bug fix list is a good thing. I wish that Apple fix list count was one higher.
My bad, it was 3 letters domain names that were gone 10 years ago, alongside all dictionary words.
But the last 4 letter domain was registered around the end of 2007
> So if you want greatsite.com but thats taken then register blah.com
You obviously haven't even tried to register a domain name. All 4 letters domain were already gone maybe 10 years ago (there was a /. article on that), and getting a barely pronounceable 5 letters domain is really really really difficult.
The issue with the cram game, is that Ia m pretty sure I can code it in C and have it faster than the Mathematica version, so that would make it a bad example for the point at hand.
The way I understood the position of the original poster is that C can be faster than every other (higher) language, and that the only reason to choose some language over C is because the source code will be smaller / more maintainable (ie: the impedance mismatch with the problem domain will be lower).
The point I am trying to make is that, if the chosen language matches a complex problem domain, then the only solution would be to re-implement everything in C and tune it to hell to match the "higher language" implementation.
Of course, my example were extreme, as they are basically library calls, but we can do some "more realistic" examples (but at the end, everything will boil down to explicit or implicit library calls):
1) A unification function. Here is some full LISP code for unification (http://web.cecs.pdx.edu/~mperkows/CLASS_ROBOTICS/LISP/tanimoto/UNIFY.CL)
Here is someone posting some "similar" C code. (http://coding.derkeiler.com/Archive/Prolog/comp.lang.prolog/2006-04/msg00084.html)
Unfortunately, as said in the post, the C code lacks variable handling which is "for free" in the lisp version. Oops. To get that working, you basically have to implement a chunk of lisp. And if you want your unification routine to outperform the lisp one on every input, you may have to get smart in memory copy (no duplication of nodes during recursion) and smart in memory usage (deleting unneeded node when necessary). This means a garbage collector. Oops again.
At the end, you'll end up with a full lisp interpreter in your C code.
2) Software using self-modifying code.
LISP, SmallTalk, Javascript or even Java, have the possibility of adding code to a running program. If we take JIT versions of LISP or Java, we can build code that evaluate function with a much better performance than any C parser / interpreter, because it will be JITed. The only solution in C would be to use something like LLVM, but I think that using such huge library would be cheating.
3) A scheme interpreter. This isn't very difficult to write in a LISP dialect, and extremely hard to do in C, for obvious reasons.
Of course, all those examples will seem weird and contrived to a C coder, precisely because it doesn't match C problem domain. Which was my original point.
And your point is ?
The challenge is, verbatim: "And that is my challenge: one, relatively small app written in any higher language will be faster when written in C".
So, you think to think that, because FactorInteger is coded in C, it should not be used. Then what about that simple program ?
G[f_] := Integrate[f, x]
You pass it a function f, and it integrates it. And no, Mathematica's Integrate is NOT written in C.
Or maybe I can't do that, because I relie on "good math libraries" ? Rules out an option pricer, then. And I guess I cannot post any Prolog program that uses unification too ? Because that is hard to do in C.
If one morph the challenge into "write a simple app that don't take advantage of anything that would be long to reproduce from scratch with a C compiler", OF COURSE, it can be rewritten in C, even if it is a bit longer and unmaintainable.
This is because notepad.exe is badly programmed.
Normally, you get the mouse down event, and set up a timer to do the scrolling. You cancel the timer when you get the mouse up event.
Some badly coded apps don't do the timer stuff and use the mouse moved event. They get the behavior you describe if the user stop moving the mouse.
Ok. Here is a simple Mathematica program:
F[x_] := Length[FactorInteger[x]]
F gives the number of distinct prime factors in x. For instance F[3] is 1, F[4] is 1 too (because 4 only have 2 as a factor] and F[12322] is 3 (12322=2*61*101)
This is a small app. Write it in C faster than Mathematica.
To give you a benchmark, an old Core2Duo take a little more than 2 seconds to answer to F[33622711352843762413542039873684935021017]
(which is 2)
> Firefox should be a little better at sandboxing plugins, but they can't be blamed for Adobe's crap.
Yes they can. They should do whatever is necessary to avoid leaving a ghost process that have to be manually killed. I had to explain to my 9 year old why he couldn't launch a browser on his ubuntu eeepc. It seems clear to him that the windows machine he uses for games is less buggy than his linux machine.
And I can't blame him.
Your argument is bullshit, as it can be used to prtoect almost any release of information. For instance, I could argue that images of nazi camps should have never been released, as it dishearten the family of the guards.
Obama said he would release the images. He didn't. He either lied or flipped. No excuses.
Stalker is quite a good adaptation of the already good novel "The Roadside Picnic".... ("good adaptation" being defined as "absolutely mind blowingly extraordinary superior")
> we would be so saturated in $1 DVDs in discount stores
I hope not. I hope we would be saturated with 20c legal downloads of good quality DiVX. Or 3$/month all-five-year-old-movies-you-can see subscription plans.
It would be so easy to find interesting stuff that nobody would even bother to "pirate" anymore. Fan of 50's SciFi movies ? Here are a few thousands, for your viewing pleasure. Fan of wildlife documentaries ? Here a few other thousands.
You would not pay for the privilege of downloading, but for the privilege of not having to mess around with files, and stuff. Just point, click and watch. Right now, I know many people that download because they have no other way to get their hands on new material that they may or may not like.
In such an ecosystem, new work could be available on some sort of premium plans, but, as it will have to compete with older good stuff, it would have to be reasonably priced. And yes, it means that it'll start to be thought to make 300M$ movies. But it'll be easier to make 1M$ ones.
Of course, there is zero chance to ever see such thing implemented: first the copyright of the stuff since 1923 will not retroactively expire. Second, the morons hijacking the culture in the name of "entertainment industry" are currently selling mp3 songs at 1.29$, which should be the digital price of the whole album.
Oh, I get it, you didn't touched the original install. Sorry, misread.
Well, I almost wanted to do that (but by booting a rescuecd and dd'ing the partitions over to some server), but, at the end, I figured out that, if I have a problem, I would just blank the disks before sending the machine back (if I can).
I don't think they can legally void the hardware warranty just because of some software installation. At least, not where I live.
Just glanced through your linked article. I installed ubuntu on a eeepc 900, and it was painless (I used the eeepc distrib, and then upgraded to more mainstream ubuntu)
On the netbook-remix interface, I removed the "maximizer" program from the session, and now windows aren't covering all the desktop, and the box is roughly usable.