16 bits is done as 5 red 6 green 5 blue. That means you can't really have true greys because the color values are spread out differently among the color channels.
With a 5-6-5 scheme you can always turn off the least significant green bit. That gives you perfect emulation of a 5-5-5 scheme. Hence, 5-6-5 can give you perfect grays, just like 5-5-5.
Both halves have antialiasing, but the left half (which the author of the article dislikes) is the one I prefer, by far. Anyway, I got motivated to get antialiasing working in Gnome. I compiled freetype with the patented hinting turned on (legal for a little while longer since I'm in Denmark), and I'm very impressed.
Indeed.I was agreeing with David Chester that the improvement was immense, just wondering why he had put the "after" image on the left. Then I figured out he didn't, he's just one of those guys who likes his text blurry and pretty instead of sharp and ugly.
Antialiasing is a neat tool, used in moderation. I just dream of 200dpi screens where it's pointless.
Right now the major impediment to sharing movies is the storage (and bandwidth, sometimes) requirements. A 150GB hard drive can only hold about 15 DVD's, so sharing of full-quality DVD's is rather expensive. As soon as 9GB+ rewritable disks become cheap or hard drives double in size a few times, DVD sharing will become popular.
Going with a high-bandwidth encoding of HDTV would ensure that only the people buying the HDTV-DVD's would get the best quality. Choosing to go with a low-bandwidth encoding ensures that sharing full-quality HDTV-DVD's will become widespread quite soon.
I expect that Warner Bros will regret this decision in a few years.
Many of the posts, including the parent, talk about ineffective use of bandwidth because of collisions. Gigabit ethernet does not support CSMA/CD at all, it's simply gone, done away with, finished. All the problems with congestion are gone.
Gigabit ethernet is certainly a viable possibility for hard drives and such (see iSCSI), but it is expensive compared to i1394, and it does not carry power. There are Power over Ethernet solutions, but they cannot work with Gigabit Ethernet, since Gigabit Ethernet needs all 8 wires.
Does anyone see anything wrong in my assumptions or calculations? It seems like their figures are very hyped-up and don't have basis in reality. Even assuming they are talking about dry sodium borohydrate (not including the water), you would need 37.8 g of it to be the equivalent of 24 g of gasoline. This is still about 1.6 times as much mass and 1.1 times as much volume.
They may be making the assumption that you burn the hydrogen in the fuel cell and the gasoline in an internal combustion engine. The difference in efficiency should make up for the dfference in stored energy.
If they can then figure out a way to extract the water from the exhaust, they should be all set.
67M AUD / 876 GWh = 0.00007 AUD/kWh over a 20 year period
You're off by three orders of magnitude, sorry. 67*10^6AUD / 876*10^9GWh = 0.000076AUD/Wh = 0.076AUD/kWh = 0.039USD/kWh. Which is what I got originally, with different assumptions.
Ok, $1.72USD/W peak. I make a rash assumption and say that it does 50% of peak on average (since it supposedly works day and night, and unlike windmills the air flow would not change that much, so it does not make sense to put in a peak capacity much larger than average -- then the wind would never get strong enough to make the peak capacity useful.) Furthermore I assume that investors want a 9% return on their initial investment each year. Lastly I assume that the yearly maintenance is 1% of initial investment.
Therefore $3440USD/(kW average capacity) invested, $344USD/(kW average capacity) expected gross, divided by 24*365 hours a year, equals 0,039USD/kWh. This compares favourably to wind energy in Denmark at least, and it would be much more reliable than wind energy.
Of course the assumptions may throw this off by an order of magnitude, but the calculation shows that this is much cheaper than solar cells (especially solar cells with battery backup).
Ok, you have figured that you only actually get 4% of the solar power as electricity. Perhaps you can tell me why I should care -- the solar power is free, coming up with a way to turn even 1% of it cheaply into electricity would be a great boon.
Efficiency would only matter if the land area was expensive. That's why this is proposed in Australia, not Hong Kong.
The temperature differential depends on whether the air is saturated with water or not. Saturated air is much harder to cool. I imagine that the air there would be far from saturation, so a differential of 10C pr. 1km sounds about right.
If you put in fiber, it will be multimode fiber, because it is cheaper and you can get fiber ports for multimode relatively cheaply. Right now you can do 1 Gbps over multimode. However, it isn't clear that future high speed nets will work over multimode. Singlemode fiber is too expensive to really consider, except possibly for putting in dark fiber and hoping that the singlemode fiber ports will magically become cheap some day
Cat5e gets you to 1 Gbps, and it's cheap. It's highly likely that it will be possible to go faster than 1 Gbps on it, even if you might never reach more than 5 Gbps. Just put in plenty of Cat5e and conduits so you can put in whichever fiber standard turns out to win in the end, if you should need to.
108 MPH sounds very low. Most cars with large engines here are limited to 250km/h though. The simple reason is that cars here have to have tires and brakes that are rated for their top speed. Tires and brakes rated for more than 250km/h are awfully expensive.
This is the first reasonable digital audio/tv device I have seen. The ones before have been tied to proprietary formats, ridiculously expensive, or limited in capacity or quality. Of course the hard drive is too small at 20GB, but that can be remedied by putting a file server in another room (so you don't have to listen to the noise). It would have been nice to have a DVD drive, but again, it's just a matter of copying the DVD to the file server.
If they added a VideoLan Client, it would be practically perfect. VLC by itself probably doesn't support the hardware MPEG decoder, and the CPU isn't fast enough to do DVD decoding in software.
Should the GNOME board consist of community members or GNOME hackers? If the latter, it's going to be impossible not to have strong representation by Ximian and Red Hat. If the latter, merely difficult.
It would be unfair if there was not strong representation by RedHat and Ximian, since those two companies contribute so much. 2/3 of the seats just seems closer to "overwhelming" than to "strong".
Since I posted I was made aware of the GNOME Advisory Board. It seems that the Advisory Board ensures that other companies and organizations are heard, so I do not seem to have much to worry about.
In retrospect the protection against company dominations could have been stronger. Of those elected:
4 are employed by Ximian
3 are employed by RedHat
3 are without affiliation
1 is employed by Compaq
As it is, Ximian and RedHat together have almost 2/3 of the seats. Both are respectable and honourable companies and I am sure that they will try to keep the viewpoints of the whole GNOME community in mind when they decide issues, but it is hard for them to represent important GNOME backers such as Sun Microsystems.
One can hope that the GNOME Board will consult with the greater community when they are facing important issues. They could invite people with different affiliations as non-voting guests for select meetings, perhaps.
Actually my argument does not hinge on whether proprietary software is similar to slavery. I have never claimed the similarity, and you clearly demonstrate that they are not the same. Try this:
"Either X is a net win for both Y and Z on pure self-interest grounds or it is not. If it is, you cannot lose; if it is not you cannot (and should not) win."
This assumes that X cannot be so great a benefit for Z that the loss to Y is insignificant in comparison. Clearly when you fill in X="abolishment of slavery", Y="slave owners", Z="slaves", you see that there is a case in which X can (and should) win. This is a counterexample, so therefore the original argument made by ESR is invalid.
I am sure that it is possible to come up with counterexamples that have different values of X, Y, and Z. However, I could not quickly think of one, and one counterexample is enough for the invalidation.
Saying that thesis work even comes close to slavery seems a bit extreme to me.
"Either open source is a net win for both producers and consumers on pure self-interest grounds or it is not. If it is, you cannot lose; if it is not you cannot (and should not) win."
Either abolishment of slavery is a net win for both slave owners and slaves on pure self-interest grounds or it is not. If it is, you cannot lose; if it is not you cannot (and should not) win.
Not all text in all message boxes can be copied. An example is the error message "You are synchronizing already. Please try again later." from Outlook.
Finally someone mentions Kay. I was getting worried. Ursula K. LeGuin is probably my favourite author, but Guy Gavriel Kay is a close competitor. I've read a lot by Marion Zimmer Bradley and Anne McCaffrey, but I've tired of Marion and Anne just writes so much that quality suffers. I've never read anything by Harlan Ellison, I'll have to check him out now.
Anyway, I don't think C.S. Lewis and Ray Bradbury have been mentioned yet, and they do deserve mention.
It is almost impossible to use overlapping windows in Windows, although it would be a very convenient thing to do. Often you can't copy text in Windows (As an example, most error messages can't be selected as text). Therefore you want to keep the error message in front of the text editor, so you can type what it says in the bug report.
Unfortunately, when you select the text editor and try to type, the text editor jumps to front. Absolutely annoying. Then you try to move the error message, only to find out that the position on the screen is fixed. Argh!
Modal dialogs are about as annoying as popup ads. Click-to-front is evil. The only way to work in Windows and not go insane is to pretend that each (maximized) application has its own workspace and that you switch workspaces with alt-tab.
You probably missed the fact that the dual bands are 900 snd 1900MHz. Which makes the device useful in the US and passable for travellers to Europe, but the lack of 1800MHz support will kill sales in Europe.
I searched Norwegian news on this. The last I could find was this: Dagbladet 28th of June 2001. A rough translation of title and subtitle goes:
"New [police] interviews of DVD-Jon. One and a half year after the search of DVD-breaker Jon Johansen's home, the police still work on the charges against the 16 year old."
I would certainly enjoy hearing about new developments.
Slashdot Mirror
Both halves have antialiasing, but the left half (which the author of the article dislikes) is the one I prefer, by far. Anyway, I got motivated to get antialiasing working in Gnome. I compiled freetype with the patented hinting turned on (legal for a little while longer since I'm in Denmark), and I'm very impressed.
Indeed.I was agreeing with David Chester that the improvement was immense, just wondering why he had put the "after" image on the left. Then I figured out he didn't, he's just one of those guys who likes his text blurry and pretty instead of sharp and ugly.
Antialiasing is a neat tool, used in moderation. I just dream of 200dpi screens where it's pointless.
Right now the major impediment to sharing movies is the storage (and bandwidth, sometimes) requirements. A 150GB hard drive can only hold about 15 DVD's, so sharing of full-quality DVD's is rather expensive. As soon as 9GB+ rewritable disks become cheap or hard drives double in size a few times, DVD sharing will become popular.
Going with a high-bandwidth encoding of HDTV would ensure that only the people buying the HDTV-DVD's would get the best quality. Choosing to go with a low-bandwidth encoding ensures that sharing full-quality HDTV-DVD's will become widespread quite soon.
I expect that Warner Bros will regret this decision in a few years.
Many of the posts, including the parent, talk about ineffective use of bandwidth because of collisions. Gigabit ethernet does not support CSMA/CD at all, it's simply gone, done away with, finished. All the problems with congestion are gone.
Gigabit ethernet is certainly a viable possibility for hard drives and such (see iSCSI), but it is expensive compared to i1394, and it does not carry power. There are Power over Ethernet solutions, but they cannot work with Gigabit Ethernet, since Gigabit Ethernet needs all 8 wires.
If they can then figure out a way to extract the water from the exhaust, they should be all set.
Dang, here we go again:
67*10^6AUD / 876*10^9Wh = 0.000076AUD/Wh = 0.076AUD/kWh = 0.039USD/kWh
At least the result was correct in the end.
Ok, $1.72USD/W peak. I make a rash assumption and say that it does 50% of peak on average (since it supposedly works day and night, and unlike windmills the air flow would not change that much, so it does not make sense to put in a peak capacity much larger than average -- then the wind would never get strong enough to make the peak capacity useful.) Furthermore I assume that investors want a 9% return on their initial investment each year. Lastly I assume that the yearly maintenance is 1% of initial investment.
Therefore $3440USD/(kW average capacity) invested, $344USD/(kW average capacity) expected gross, divided by 24*365 hours a year, equals 0,039USD/kWh. This compares favourably to wind energy in Denmark at least, and it would be much more reliable than wind energy.
Of course the assumptions may throw this off by an order of magnitude, but the calculation shows that this is much cheaper than solar cells (especially solar cells with battery backup).
Ok, you have figured that you only actually get 4% of the solar power as electricity. Perhaps you can tell me why I should care -- the solar power is free, coming up with a way to turn even 1% of it cheaply into electricity would be a great boon.
Efficiency would only matter if the land area was expensive. That's why this is proposed in Australia, not Hong Kong.
The temperature differential depends on whether the air is saturated with water or not. Saturated air is much harder to cool. I imagine that the air there would be far from saturation, so a differential of 10C pr. 1km sounds about right.
If you put in fiber, it will be multimode fiber, because it is cheaper and you can get fiber ports for multimode relatively cheaply. Right now you can do 1 Gbps over multimode. However, it isn't clear that future high speed nets will work over multimode. Singlemode fiber is too expensive to really consider, except possibly for putting in dark fiber and hoping that the singlemode fiber ports will magically become cheap some day
Cat5e gets you to 1 Gbps, and it's cheap. It's highly likely that it will be possible to go faster than 1 Gbps on it, even if you might never reach more than 5 Gbps. Just put in plenty of Cat5e and conduits so you can put in whichever fiber standard turns out to win in the end, if you should need to.
108 MPH sounds very low. Most cars with large engines here are limited to 250km/h though. The simple reason is that cars here have to have tires and brakes that are rated for their top speed. Tires and brakes rated for more than 250km/h are awfully expensive.
If they added a VideoLan Client, it would be practically perfect. VLC by itself probably doesn't support the hardware MPEG decoder, and the CPU isn't fast enough to do DVD decoding in software.
Since I posted I was made aware of the GNOME Advisory Board. It seems that the Advisory Board ensures that other companies and organizations are heard, so I do not seem to have much to worry about.
In retrospect the protection against company dominations could have been stronger. Of those elected:
4 are employed by Ximian
3 are employed by RedHat
3 are without affiliation
1 is employed by Compaq
As it is, Ximian and RedHat together have almost 2/3 of the seats. Both are respectable and honourable companies and I am sure that they will try to keep the viewpoints of the whole GNOME community in mind when they decide issues, but it is hard for them to represent important GNOME backers such as Sun Microsystems.
One can hope that the GNOME Board will consult with the greater community when they are facing important issues. They could invite people with different affiliations as non-voting guests for select meetings, perhaps.
"Either X is a net win for both Y and Z on pure self-interest grounds or it is not. If it is, you cannot lose; if it is not you cannot (and should not) win."
This assumes that X cannot be so great a benefit for Z that the loss to Y is insignificant in comparison. Clearly when you fill in X="abolishment of slavery", Y="slave owners", Z="slaves", you see that there is a case in which X can (and should) win. This is a counterexample, so therefore the original argument made by ESR is invalid.
I am sure that it is possible to come up with counterexamples that have different values of X, Y, and Z. However, I could not quickly think of one, and one counterexample is enough for the invalidation.
Saying that thesis work even comes close to slavery seems a bit extreme to me.
Not all text in all message boxes can be copied. An example is the error message "You are synchronizing already. Please try again later." from Outlook.
Finally someone mentions Kay. I was getting worried. Ursula K. LeGuin is probably my favourite author, but Guy Gavriel Kay is a close competitor. I've read a lot by Marion Zimmer Bradley and Anne McCaffrey, but I've tired of Marion and Anne just writes so much that quality suffers. I've never read anything by Harlan Ellison, I'll have to check him out now.
Anyway, I don't think C.S. Lewis and Ray Bradbury have been mentioned yet, and they do deserve mention.
Unfortunately, when you select the text editor and try to type, the text editor jumps to front. Absolutely annoying. Then you try to move the error message, only to find out that the position on the screen is fixed. Argh!
Modal dialogs are about as annoying as popup ads. Click-to-front is evil. The only way to work in Windows and not go insane is to pretend that each (maximized) application has its own workspace and that you switch workspaces with alt-tab.
You probably missed the fact that the dual bands are 900 snd 1900MHz. Which makes the device useful in the US and passable for travellers to Europe, but the lack of 1800MHz support will kill sales in Europe.
It's an old joke. I prefer this version from back in the Amiga days:
If Commodore had sold sushi, it would have marketed it as cold, dead fish.
"New [police] interviews of DVD-Jon. One and a half year after the search of DVD-breaker Jon Johansen's home, the police still work on the charges against the 16 year old."
I would certainly enjoy hearing about new developments.