Substitute "roguelike" for "text adventure" and I agree with you. Diablo 2 "puzzles" (more accurately described as quests or problems) are all "kill this monster", and I'll give you a skill, or add sockets to your item. Quite similar to nethack, where the quest is to go kill the minion of the enemy god, and get your artifact, and basically all you do is run around and kill monsters and get treasure. Text adventures have puzzles. I haven't really played any, but if they're anything like Sierra games without graphics, there is little resemblance to Diablo 2. AFAIK, they are not stat-building games like nethack and Diablo.
Dude, this is the wrong kind of thread for advocacy. Nobody's even talking about getting a new system or doing anything major that Linux would rock at. I'm a huge fan of Debian, Free software, and GNU, and I have made advocacy posts on/. and on mailing lists, but I've learned not to overdo it. Just coming out of the blue with how great GNU/Linux is when people are just talking about some details of Windoze just pisses people off. If this thread was about some huge annoyance in windoze that Debian doesn't have, or something windoze can't do at all, then maybe it might be appropriate to do some advocacy, but think before you post. Most people here are aware of GNU/Linux, so pissing people off is worse than saying nothing.
It's all about muscle memory. When I'm typing a word,like "line", that's similar to "linux", I invariably type "linux" and have to fix it. (My muscle memory doesn't type GNU/Linux by itself, though...) Same thing with words close to my first or last name. I touch type pretty much the Official Correct Way, but for keys between my hands I don't always do it the way they say. I learned by basically pressing whatever key with whatever finger, so that certainly develops muscle memory for some strings (e.g. "grep", and "/usr/share/doc" (actually it was/usr/doc when I started, but it's one of the strings I type often enough these days to be muscle-accelerated)).
To keep the climate from changing faster than the ecosystem can adapt, we'd better get serious about greenhouse gases. I can just see oil tycoons saying "see, it's not our fault, so keep burnin' oil", but that's the exact opposite of the sensible course of action. To the extent that humanity can control the environment to keep it good, we should do what we can.
Think about it. If advertizers thought it was a good idea to broadcast adds from pop machines, they could do it with normal loudspeakers. They're smart enough to know how much that would piss people off, and lead to the machine getting unplugged, so they don't do it. There are laws against noise pollution, and so on. It might catch on in stores like the article suggested, though, and I don't like that prospect. I already detest shopping, and HSS would certainly make me dislike it even more. It would be a lot harder to smash the emitter in a store than in a soda machine.
You're right (in your reply to your own message) about the scary prospects of this in the hands of unscrupulous individuals in positions of power, such as the CIA.
> Would public protests against the government eliminated by these devices?
If it's a lot worse than tear gas, using it will create sympathy for protesters who are attacked with it. If it isn't, then people will add earplugs to their G8 welcoming committee survival kits, along with wet cloths to protect against gas, and thick clothing to pretect against police brutality. (if earplugs or headphone-style ear protectors work against this.)
I just hope this doesn't cause long-term hearing damage, but if it does, maybe people could sue the police for brutality or something.
> What if they modulated it to go through space as a laser beam or something?
You know in Star Trek, when they remodulate their stuff whenever they want to do something? They're just making that up. Unless you can think of a way to convert a laser beam into sound, other than setting up a receiver, amp, and speaker (HSS or otherwise), you're not going to get anywhere with this. If you need to put something somewhere to receive the satellite signal, it might as well be a normal satellite phone.
Feel free to correct my ignorance, but as I understand it X was never more than a network-transparent way to stack bitmaps on top of each other.
X has draw-line, draw-rectangle, draw-text, etc. calls. Run x11perf and check out all the different things you can ask the X server to draw with a single request. Running xterm or emacs over a network works pretty well, because they send the text as ASCII, and the server puts the bitmaps into the framebuffer. Qt and GTK AFAIK don't take advantage of the draw-whatever ops, and just pass bitmaps, which is probably the source of your confusion. Try running some Xaw apps, like xedit:).
Re:Perhaps X/XFree time is past?
on
XFree86 Politics
·
· Score: 1
I think the situation you describe (with slowness over a network) is due to Qt and GTK not really being designed with remote X in mind. AFAIK, X used to work well without a huge amount of network traffic, back when programs didn't just draw everything to a bitmap. X has more complicated operations that programs can use, but they're not flexible enough for drawing all the eye candy that GTK and Qt want. Programs like emacs and xterm work fine remotely, since they draw text in a window, and X has calls to draw text with a given font. The font bitmaps don't go back and forth between client and server, only the ASCII/Unicode/whatever. (emacs and xterm probably also draw their menus with line-drawing ops, rather than making a bitmap and sending it to the server.)
I'm not saying that GTK and Qt are bad, I'm saying that X's design doesn't provide the capabilities for good network transparency AND eye candy. X is showing its age, and to do all the eye candy people want these days, we have to brute-force it.
That's how I understand the situation. I probably got some of it a bit wrong, so feel free to correct me instead of flaming if I said something nasty by mistake.
Those dict-protocol URIs didn't come out right, but they don't work with mozilla anyway. They're supposed to be like dict://dict.org/d:metaphor, as in rfc2229.
Chevy commercial: Like a rock... Ohhhh, like a rock... Like a rock...
People who speak in metaphors should shampoo my crotch.
Metaphor: You are a slow 286. Simile: You are like a slow 286.
Before you complain, check up on what you're talking about. Try reading some Ray Bradbury, so you can tell whether it's just bad metaphors and similes you hate, or whether it's the figure of speech itself. (Bradbury's metaphors are for the most part good.:)
You can actually boot up a kernel, _then_ compile new modules for it, and load them, so yes, you can install completely new device drivers after the fact. The only exception I've seen is that if you don't have IPv6 as a module when you build the kernel, you can't load the module. You have to build a new vmlinuz as well as ipv6.o. (This is with 2.4.18 or so, and it's probably a bug, but I haven't reported it.)
> Or if it trashes a hardware patent they rely on (let's not red-herring on software patents).
Patents can't stop something from being open source, because patents are all about "this is how to do such-and-such, but you can't do it unless I say you can". Patents are totally open. There is debate over whether open source code that implements patented algorithms is compatible with the GPL, but RT-Linux (IIRC) calls itself GPL and uses a patented method of running an OS on top of a real-time interrupt-handling layer. (I haven't payed attention to this for over a year, so further developments have probably occured.)
If you have Free code that uses patents, you can fix bugs in it or modify it to use different APIs, or even just recompile it against a new ABI, but you can't use it in other projects.
There are other kinds of "intellectual property" that prevent open-sourcing, though.
> So I can't hack the source code for fun, well boo hoo, but at least I have a kickass driver.
Besides that, if you have a kernel crash, you can't tell whether it was a result of some code in NVidia module or some experimental code in some other part of the kernel, or even some interaction between them. Some kernel developers are reportedly not at all eager to even look at crashed people are having if they are using the NVidia kernel module. (Yes, there is a kernel module. It's big (over a hundred kB, IIRC), too, and that's a lot of code of unknown function to have doing stuff in kernel-space.)
The NVidia stuff comes with source for wrapper functions that interface between the kernel and the already-compiled stuff, but nobody knows exactly what other parts aspects of the kernel they depend on. (i.e. what you can change without breaking them.) One reason Linus doesn't officially provide a binary module interface is that he doesn't want to be pinned down by having to maintain compatibility with it. My point is that you can't easily do kernel development while using the NVidia modules. Fortunately, there is accelerated 2D support in XFree86 using all Free code, so even those stuck with NV cards can try experimental kernels, or even hack around with them, as long as they're willing to give up hardware 3D. If not for this, the situation would be quite bad, esp. for laptop users who can't change their video hardware. It's not good to have things adding obstacles to people trying new code, because stuff has to get tested before it can be called stable. Even so, lots of people don't want to give up hardware 3D. (BTW, I may have exagerated how bad things are. I haven't even tried to use NVidia's closed source stuff with a 2.5 kernel, so if it actually works, people could at least be testing stuff. Still, if it crashes, it's harder to figure out why.)
> GPL is a fine thing, but it can't solve all the world's problems, so give nVIDIA a break, at least they provide Linux drivers.
Well, it's better than nothing, but it would be a bad thing if this started to get widespread. Not everybody needs to go nuts over Freeness, but somebody has to, or everyone's freedom will diminish over time. Besides, the problem with NV is not just Freeness, it's the consequences for kernel development, because the closed-source module is so big and complicated.
Interesting set of messages. One thing not mentioned on that page is that between then and now, DRM (the Direct Rendering Manager stuff) is in the official kernel. I don't follow the kernel list, so I don't know if it had to get stuffed down Linus' throat or if he loves it, but it's in there now. I haven't hacked around with any of this stuff, so unfortunately I can't add anything useful beyond that...
You'd need a more complicated power supply to efficiently use power from an ultracap, since they really are capacitors. The voltage across the terminals is linearly proportional to the remaining charge (and thus the square root of the remaining energy). To power a fixed-voltage circuit, you'd need a flexible DC-DC converter. OTOH, I don't think it would be much of a problem to build such a supply.
Other than that, I think you're misinterpreting the message: power is energy/time. 10 times the power means you can charge or discharge fast. The page you linked doesn't say anything about energy storage capacity, but I'd guess it's less than chemical batteries. If you get an electron of charge per molecule of whatever's in your battery, you'd be hard-pressed to match that with a capacitor, no matter how much effective surface area your carbon has. The page doesn't say anything about being good for powering electric vehicles on its own, just for handling the high-current initial acceleration and regenerative braking.
> al-Muhajir travelled to Pakistan for some time...
I didn't know that, sorry. I'd heard some about the case before, and I thought I knew some things, but I guess I was wrong. Sorry about that, but I've read the article now.
> These criteria clearly match Mr. al-Muhajir.
Only if you count Al-Qaeda as an enemy army the US is at war with, rather than a criminal/terrorist organization. Al-Qaeda is not part of any country's military force, and AFAIK, there is no group of people anywhere that has a government (official country or not) that has Al-Qaeda as it's military.
Do you really think that the law of war makes it illegal to fight without a uniform if you own a uniform but left it home?
No, I'm not trying to pull some stupid technical argument like that.
Another point I might add: Just because you have laws that allow you to be really harsh, you don't have to apply them when they're not necessary. Not everyone thinks it's necessary to be so nasty about locking people up without giving them their day in court for a very long time.
Non sequitur. If you don't like what some group is up to, it makes sense to boycott their stuff whether or not that country is boycotting anybody else.
You can disaprove of Israel's actions and policies without wanting Arabs to destroy it! Israel is a modern democracy, and Israelis by and large are not too brainwashed or dogmatic, so it makes sense to try to influence them. Changing Arab opinion is not easy, because parents keep bringing up their kids to hate the "great satan" and all that stupid conspiracy crap.
802.11g is _not_ just dual band a/b. It is 802.11a's OFDM (Orthogonal Frequency Division Multiplexing, which is a really good way to send bits without multipath interference problems) at 2.4GHz, the same band as 802.11b. It is backwards compatible with 11b. The draft standard includes some optional modulations, like PBCC at 22Mb/s, which is already used by Texas Instruments chipsets, notably in D-Link's "802.11b+" gear. Thus, TI's 802.11g gear will be able to talk to their b+ gear at 22. (Very little hope of chipsets from other manufacturers doing PBCC, though, so they would only talk to "b+" gear at 11Mb/s.)
> You're mighty quick to try to twist someone's words. Of course the presumption of innocence applies...
I was trying to gain some insight into your thoughts from the words that you chose. If I was mistaken, then I guess it would seem to you that your words had been twisted. Sorry about that, but you've been sounding so hawkish that I wanted to remind you (and everyone reading this) that I thought it necessary to remind everyone that we don't even have any idea what this guy actually did. Most people don't like the idea of people being hauled off like that based on unproven suspicion, especially during peace time. Even if you can twist an interpretation of law to apply to this situation, most people still don't like it. (I read your journal entry that you linked to, and your posts, and I don't think the Ex Parte Quirin decision is relevant, because Padilla/al-Muhajir has been a US resident for the whole time, unlike the guys that went to Germany and later returned on a submarine as saboteurs. The decision seems to hinge on the fact that they left, trained with the German army, then came back as soldiers, but didn't wear uniforms, thus making them illegal combatants. Since that important detail is not the same for the Padilla case, the decision is not directly relevant.)
One problem with making (written) hard SF into a movie is that movies have trouble explaining anything without having the characters standing there talking about it. The descriptive language used in stories conveys the physical principles at work, etc. (e.g. the story might talk about blueshifted light, while in a movie you'd just see blue stars, without realizing they were blue because of gravity or doppler shift, or whatever was going on.) It's dumb when characters stand there discussing something that should be common knowledge for all the characters, so you know they're saying it for the audience's benefit. They might as well have a very short educational piece with diagrams and stuff instead of having the characters explain it. Either that, or everything gets Trekkified, so there is no apparent science in action, and the audience's knowledge of physical principles is no longer useful in seeing where the story is going, or anticipating what the characters should do.
Slashdot Mirror
Substitute "roguelike" for "text adventure" and I agree with you. Diablo 2 "puzzles" (more accurately described as quests or problems) are all "kill this monster", and I'll give you a skill, or add sockets to your item. Quite similar to nethack, where the quest is to go kill the minion of the enemy god, and get your artifact, and basically all you do is run around and kill monsters and get treasure. Text adventures have puzzles. I haven't really played any, but if they're anything like Sierra games without graphics, there is little resemblance to Diablo 2. AFAIK, they are not stat-building games like nethack and Diablo.
rock over london, rock over chicago.
Dude, this is the wrong kind of thread for advocacy. Nobody's even talking about getting a new system or doing anything major that Linux would rock at. I'm a huge fan of Debian, Free software, and GNU, and I have made advocacy posts on /. and on mailing lists, but I've learned not to overdo it. Just coming out of the blue with how great GNU/Linux is when people are just talking about some details of Windoze just pisses people off. If this thread was about some huge annoyance in windoze that Debian doesn't have, or something windoze can't do at all, then maybe it might be appropriate to do some advocacy, but think before you post. Most people here are aware of GNU/Linux, so pissing people off is worse than saying nothing.
It's all about muscle memory. When I'm typing a word ,like "line", that's similar to "linux", I invariably type "linux" and have to fix it. (My muscle memory doesn't type GNU/Linux by itself, though...) Same thing with words close to my first or last name. I touch type pretty much the Official Correct Way, but for keys between my hands I don't always do it the way they say. I learned by basically pressing whatever key with whatever finger, so that certainly develops muscle memory for some strings (e.g. "grep", and "/usr/share/doc" (actually it was /usr/doc when I started, but it's one of the strings I type often enough these days to be muscle-accelerated)).
To keep the climate from changing faster than the ecosystem can adapt, we'd better get serious about greenhouse gases. I can just see oil tycoons saying "see, it's not our fault, so keep burnin' oil", but that's the exact opposite of the sensible course of action. To the extent that humanity can control the environment to keep it good, we should do what we can.
Think about it. If advertizers thought it was a good idea to broadcast adds from pop machines, they could do it with normal loudspeakers. They're smart enough to know how much that would piss people off, and lead to the machine getting unplugged, so they don't do it. There are laws against noise pollution, and so on. It might catch on in stores like the article suggested, though, and I don't like that prospect. I already detest shopping, and HSS would certainly make me dislike it even more. It would be a lot harder to smash the emitter in a store than in a soda machine.
You're right (in your reply to your own message) about the scary prospects of this in the hands of unscrupulous individuals in positions of power, such as the CIA.
> Would public protests against the government eliminated by these devices?
If it's a lot worse than tear gas, using it will create sympathy for protesters who are attacked with it. If it isn't, then people will add earplugs to their G8 welcoming committee survival kits, along with wet cloths to protect against gas, and thick clothing to pretect against police brutality. (if earplugs or headphone-style ear protectors work against this.)
I just hope this doesn't cause long-term hearing damage, but if it does, maybe people could sue the police for brutality or something.
> What if they modulated it to go through space as a laser beam or something?
You know in Star Trek, when they remodulate their stuff whenever they want to do something? They're just making that up. Unless you can think of a way to convert a laser beam into sound, other than setting up a receiver, amp, and speaker (HSS or otherwise), you're not going to get anywhere with this. If you need to put something somewhere to receive the satellite signal, it might as well be a normal satellite phone.
I _knew_ it was a Freudian typo influenced by years of playing AD&D.
X has draw-line, draw-rectangle, draw-text, etc. calls. Run x11perf and check out all the different things you can ask the X server to draw with a single request. Running xterm or emacs over a network works pretty well, because they send the text as ASCII, and the server puts the bitmaps into the framebuffer. Qt and GTK AFAIK don't take advantage of the draw-whatever ops, and just pass bitmaps, which is probably the source of your confusion. Try running some Xaw apps, like xedit
I think the situation you describe (with slowness over a network) is due to Qt and GTK not really being designed with remote X in mind. AFAIK, X used to work well without a huge amount of network traffic, back when programs didn't just draw everything to a bitmap. X has more complicated operations that programs can use, but they're not flexible enough for drawing all the eye candy that GTK and Qt want. Programs like emacs and xterm work fine remotely, since they draw text in a window, and X has calls to draw text with a given font. The font bitmaps don't go back and forth between client and server, only the ASCII/Unicode/whatever. (emacs and xterm probably also draw their menus with line-drawing ops, rather than making a bitmap and sending it to the server.)
I'm not saying that GTK and Qt are bad, I'm saying that X's design doesn't provide the capabilities for good network transparency AND eye candy. X is showing its age, and to do all the eye candy people want these days, we have to brute-force it.
That's how I understand the situation. I probably got some of it a bit wrong, so feel free to correct me instead of flaming if I said something nasty by mistake.
Those dict-protocol URIs didn't come out right, but they don't work with mozilla anyway. They're supposed to be like dict://dict.org/d:metaphor, as in rfc2229.
Metaphor: You are a slow 286.
Simile: You are like a slow 286.
Before you complain, check up on what you're talking about. Try reading some Ray Bradbury, so you can tell whether it's just bad metaphors and similes you hate, or whether it's the figure of speech itself. (Bradbury's metaphors are for the most part good.
You can actually boot up a kernel, _then_ compile new modules for it, and load them, so yes, you can install completely new device drivers after the fact. The only exception I've seen is that if you don't have IPv6 as a module when you build the kernel, you can't load the module. You have to build a new vmlinuz as well as ipv6.o. (This is with 2.4.18 or so, and it's probably a bug, but I haven't reported it.)
> Or if it trashes a hardware patent they rely on (let's not red-herring on software patents).
Patents can't stop something from being open source, because patents are all about "this is how to do such-and-such, but you can't do it unless I say you can". Patents are totally open. There is debate over whether open source code that implements patented algorithms is compatible with the GPL, but RT-Linux (IIRC) calls itself GPL and uses a patented method of running an OS on top of a real-time interrupt-handling layer. (I haven't payed attention to this for over a year, so further developments have probably occured.)
If you have Free code that uses patents, you can fix bugs in it or modify it to use different APIs, or even just recompile it against a new ABI, but you can't use it in other projects.
There are other kinds of "intellectual property" that prevent open-sourcing, though.
> So I can't hack the source code for fun, well boo hoo, but at least I have a kickass driver.
Besides that, if you have a kernel crash, you can't tell whether it was a result of some code in NVidia module or some experimental code in some other part of the kernel, or even some interaction between them. Some kernel developers are reportedly not at all eager to even look at crashed people are having if they are using the NVidia kernel module. (Yes, there is a kernel module. It's big (over a hundred kB, IIRC), too, and that's a lot of code of unknown function to have doing stuff in kernel-space.)
The NVidia stuff comes with source for wrapper functions that interface between the kernel and the already-compiled stuff, but nobody knows exactly what other parts aspects of the kernel they depend on. (i.e. what you can change without breaking them.) One reason Linus doesn't officially provide a binary module interface is that he doesn't want to be pinned down by having to maintain compatibility with it. My point is that you can't easily do kernel development while using the NVidia modules. Fortunately, there is accelerated 2D support in XFree86 using all Free code, so even those stuck with NV cards can try experimental kernels, or even hack around with them, as long as they're willing to give up hardware 3D. If not for this, the situation would be quite bad, esp. for laptop users who can't change their video hardware. It's not good to have things adding obstacles to people trying new code, because stuff has to get tested before it can be called stable. Even so, lots of people don't want to give up hardware 3D. (BTW, I may have exagerated how bad things are. I haven't even tried to use NVidia's closed source stuff with a 2.5 kernel, so if it actually works, people could at least be testing stuff. Still, if it crashes, it's harder to figure out why.)
> GPL is a fine thing, but it can't solve all the world's problems, so give nVIDIA a break, at least they provide Linux drivers.
Well, it's better than nothing, but it would be a bad thing if this started to get widespread. Not everybody needs to go nuts over Freeness, but somebody has to, or everyone's freedom will diminish over time. Besides, the problem with NV is not just Freeness, it's the consequences for kernel development, because the closed-source module is so big and complicated.
Interesting set of messages. One thing not mentioned on that page is that between then and now, DRM (the Direct Rendering Manager stuff) is in the official kernel. I don't follow the kernel list, so I don't know if it had to get stuffed down Linus' throat or if he loves it, but it's in there now. I haven't hacked around with any of this stuff, so unfortunately I can't add anything useful beyond that...
You'd need a more complicated power supply to efficiently use power from an ultracap, since they really are capacitors. The voltage across the terminals is linearly proportional to the remaining charge (and thus the square root of the remaining energy). To power a fixed-voltage circuit, you'd need a flexible DC-DC converter. OTOH, I don't think it would be much of a problem to build such a supply.
Other than that, I think you're misinterpreting the message: power is energy/time. 10 times the power means you can charge or discharge fast. The page you linked doesn't say anything about energy storage capacity, but I'd guess it's less than chemical batteries. If you get an electron of charge per molecule of whatever's in your battery, you'd be hard-pressed to match that with a capacitor, no matter how much effective surface area your carbon has. The page doesn't say anything about being good for powering electric vehicles on its own, just for handling the high-current initial acceleration and regenerative braking.
I didn't know that, sorry. I'd heard some about the case before, and I thought I knew some things, but I guess I was wrong. Sorry about that, but I've read the article now.
> These criteria clearly match Mr. al-Muhajir.
Only if you count Al-Qaeda as an enemy army the US is at war with, rather than a criminal/terrorist organization. Al-Qaeda is not part of any country's military force, and AFAIK, there is no group of people anywhere that has a government (official country or not) that has Al-Qaeda as it's military.
No, I'm not trying to pull some stupid technical argument like that.
Another point I might add: Just because you have laws that allow you to be really harsh, you don't have to apply them when they're not necessary. Not everyone thinks it's necessary to be so nasty about locking people up without giving them their day in court for a very long time.
Non sequitur. If you don't like what some group is up to, it makes sense to boycott their stuff whether or not that country is boycotting anybody else.
You can disaprove of Israel's actions and policies without wanting Arabs to destroy it! Israel is a modern democracy, and Israelis by and large are not too brainwashed or dogmatic, so it makes sense to try to influence them. Changing Arab opinion is not easy, because parents keep bringing up their kids to hate the "great satan" and all that stupid conspiracy crap.
802.11g is _not_ just dual band a/b. It is 802.11a's OFDM (Orthogonal Frequency Division Multiplexing, which is a really good way to send bits without multipath interference problems) at 2.4GHz, the same band as 802.11b. It is backwards compatible with 11b. The draft standard includes some optional modulations, like PBCC at 22Mb/s, which is already used by Texas Instruments chipsets, notably in D-Link's "802.11b+" gear. Thus, TI's 802.11g gear will be able to talk to their b+ gear at 22. (Very little hope of chipsets from other manufacturers doing PBCC, though, so they would only talk to "b+" gear at 11Mb/s.)
> You're mighty quick to try to twist someone's words. Of course the presumption of innocence applies ...
I was trying to gain some insight into your thoughts from the words that you chose. If I was mistaken, then I guess it would seem to you that your words had been twisted. Sorry about that, but you've been sounding so hawkish that I wanted to remind you (and everyone reading this) that I thought it necessary to remind everyone that we don't even have any idea what this guy actually did. Most people don't like the idea of people being hauled off like that based on unproven suspicion, especially during peace time. Even if you can twist an interpretation of law to apply to this situation, most people still don't like it. (I read your journal entry that you linked to, and your posts, and I don't think the Ex Parte Quirin decision is relevant, because Padilla/al-Muhajir has been a US resident for the whole time, unlike the guys that went to Germany and later returned on a submarine as saboteurs. The decision seems to hinge on the fact that they left, trained with the German army, then came back as soldiers, but didn't wear uniforms, thus making them illegal combatants. Since that important detail is not the same for the Padilla case, the decision is not directly relevant.)
> Not a movie trailer, but you're on the right track.
/. doesn't let you play nethack while you wait 2 minutes so you can post again, or give you some quotes from zippy the pinhead to read....
Well, I think it's a pretty good analogy.
Too bad
How do the French^H^H^Hfreedom fries taste?
One problem with making (written) hard SF into a movie is that movies have trouble explaining anything without having the characters standing there talking about it. The descriptive language used in stories conveys the physical principles at work, etc. (e.g. the story might talk about blueshifted light, while in a movie you'd just see blue stars, without realizing they were blue because of gravity or doppler shift, or whatever was going on.) It's dumb when characters stand there discussing something that should be common knowledge for all the characters, so you know they're saying it for the audience's benefit. They might as well have a very short educational piece with diagrams and stuff instead of having the characters explain it. Either that, or everything gets Trekkified, so there is no apparent science in action, and the audience's knowledge of physical principles is no longer useful in seeing where the story is going, or anticipating what the characters should do.
Things can only happen if they're _possible_.