The absolute power limit is 1 watt for unlicensed 2.4 ghz equipment. The EIRP limit is 4 watts for point-to-multipoint, but can be higher for point-to-point. (Disclaimer: I'm not an RF engineer, so read the actual fcc regulations before you try it.)
You can find transmit power / recieve sensitivity numbers for a lot of prism2 cards here.
I suspect most of these numbers are gleaned from product brochures, so you mileage may vary.
I'm running Mandrake 9.2 at work and home. In both cases, the default Apache config runs an open http proxy on port 80. This is a pretty bad security model. It also prevents you from posting to Slashdot. Is this a problem in Mandrake 10?
Well, I'm running Mandrake 10.0, httpd is running, and I can still post to slashdot, so apparently it must be fixed.
IIRC, the maximum tx power in the 2.4ghz band for unlicensed users is 1 watt, or 4 watts EIRP. For point-to-point links, though, you can trade 1db of power reduction for 3 db of antenna gain, allowing much higher EIRP.
It couldn't possibly be omnidirectional. That would break the laws of physics. To get a boost in signal strength you must either make it more directional or increase the power itself with an amplifier.
"Omnidirectional" is somewhat of a misnomer. Omnis transmit a pancake-shaped signal -- good signal in all directions in the same plane, but very little signal up and down. What you're refering to is called an isotropic radiator.
The super cantenna is only 12 db. 17 is more impressive, and should result in greater range.
Range itself is hard to compare, as it depends on environment, the radios used (cheap 35 mW? 200 mW with good receive sensitivity?), whether the same antenna is used on both ends, and the subjective evaluation of what exactly constitutes a "useable signal".
Modern spicers make all fusion splicing child's play.
If you can afford 5 grand or so for a fusion splicer... (Not a problem if installing fiber is your job, but it's a large barrier to entry for a hobbyist.)
single mode end point equipment is more expensive
Need not be. Using WDM single mode may actually end up saving you money.
The price difference isn't that big, but SM is usually a bit more expensive (and, for the extremely budget conscious (me), much harder to find on ebay, though I did pick up a nice rack of 6 100mbps SM media converters for about ten bucks, so there are exceptions. On the other hand, people can't hardly give MM 10-base-FL gear away.)
The fiber optic cable will outlast you and me.
I assume this is true for good, high quality direct-bury cable. What if I just burried zipcord or even *the horror* bare, unbuffered fiber in the ground? Would it decompose in a year or two, or keep working for thousands of years? What if it was left in direct sunlight? These are the questions your average network enthusiast might want to know before running fiber to the neighbor's house to, say, share an internet connection, play multiplayer games, or route to a dead spot in a community wireless network.
Ethernet over fiber may be good for small networks
Ethernet is good for any size of network with proper engineering and planning.
Pure ethernet uses the spanning tree protocol for routing, which often chooses suboptimal routes. If you stuck an IP router at the major junctions, then yes, ethernet would scale as well as anything else you might think to use.
I'd be interested in panel displays with no trim on the side so they could be placed adjacent to each other for a larger screen. Does anyone know if that's possible with current technology, or if anyone makes that now? (Okay, okay, what I really want is something I can roll up like a poster, but I don't expect that to happen any time soon.)
Someone has to mention it sooner or later, and it might as well be me: LaTeX!
I've been using latex for a few months, including writing a research paper. Placement of figures can be a pain, but for text it is reasonably intuitive. All the formatting trivia is abstracted away, and you can concentrate on just writing. It's math features are nice, too, if you need them.
How about "Church of the Heroic Cranial Ellipsoid" ?.
Maybe this will be a less controversial subject once the new Zelda comes out. On the other hand, I still occasionally talk to people who don't consider the Adventures of Link a real Zelda game because it was a side scroller.
Even with warping, it still took too long to get anywhere, in my opinion.
Was the remake of Majora's mask etc good? I was a little sad that they didn't increase the framerate a little for Ocarina of Time, but it was still fun to play the Master Quest version even though it made my eyes hurt.
I've been wondering why there isn't a good "do it yourself community fiber network" howto, at least as far as I know. I've been experimenting with fiber as a hobby for the last few months just for fun. Its amazing what you can find on ebay.
A few observations:
Single mode is vastly superior to multimode in terms of both range (20-100 km vs 2km) and theoretical maximum throughput (terabits per second versus gigabits per second).
On the other hand, single mode is harder to work with, and the end point equipment is more expensive. A new 100mbps fiber-to-copper ethernet converter is about $150 for MM and about $250 for SM. Gigabit gear is more expensive, but not terribly prohibitive.
The fiber itself is cheap (one article I read indicated that the wholesale cost of SM fiber is about $15 a km). All the cladding and armor they put on it makes it expensive (A dollar or two per foot for direct-bury cable with a dozen or so fiber strands).
It takes 2 fibers to make a connection, usually. This is called duplex.
It is possible to run multiple connections on different wavelengths. This is called wavelength division multiplexing. DWDM systems sometimes have over a hundred separate channels.
I don't know much about durability.
Ethernet over fiber may be good for small networks, but it requires active electronics (and reliable power) at each junction. Depending on application, this may not be a problem.
Take a look at fiberdyne's webpage if you're curious about approximate equipment costs. They seem to sell almost everything related to fiber. Here's another page with a decent fiber tutorial.
I don't understand why a "realistic" Zelda game is a good idea.
Because its hard to take a hero seriously who's head is a perfect sphere. (I have nothing against non-photorealistic rendering, it just needs to be done right. Windwaker was an experiment. They got most things right, but they also got a few things wrong, including Link's head and the whole sailing-forever-to-get-anywhere thing. And no, I'm not sorry I bought it, but it from my perspective it could have been better than it was. I expect this to be the perpetual vi versus emacs debate of the Nintendo world for quite some time to come.)
I just watched the trailer. Very nice. I'm looking forward to it coming out.
As other replies said, this is an easy read. Just to sum up for the lazy, though:
Most of the text relates specifically and exclusively to audio CDs. Specifically, it will require CDs with copy protection to be labeled as such.
The last bit is more general though:
SEC. 5. FAIR USE AMENDMENTS.
(a) SCIENTIFIC RESEARCH- Subsections (a)(2)(A) and (b)(1)(A) of section 1201 of title 17, United States Code, are each amended by inserting after `title' in subsection (a)(2)(A) and after `thereof' in subsection (b)(1)(A) the following: `unless the person is acting solely in furtherance of scientific research into technological protection measures'.
(b) FAIR USE RESTORATION- Section 1201(c) of title 17, United States Code, is amended--
(1) in paragraph (1), by inserting before the period at the end the following:
and it is not a violation of this section to circumvent a technological measure in connection with access to, or the use of, a work if such circumvention does not result in an infringement of the copyright in the work'; and
(2) by adding at the end the following new paragraph:
(5) It shall not be a violation of this title to manufacture, distribute, or make noninfringing use of a hardware or software product capable of enabling significant noninfringing use of a copyrighted work.
It seems this allows certain kinds of research, in addition to making it legal to use and make products that circumvent copy protection if the purpose is not copyright infringement. It seems that breaking a copy protection technology for the purposes of copyright violation is still illegal under both the DMCA and plain old copyright law (but someone correct me if I'm wrong).
I'd also like to point out (because I like to pick nits) the shortage of details in the 321 studios press release:
The DMCRA... would re-affirm consumer fair use rights and balance the otherwise one-sided protection afforded copyright owners under current interpretations of the Digital Millennium Copyright Act (DMCA).
It would have been a simple matter to explain how the DMCRA would re-affirm fair use and balance. You'd think people these days were afraid of using concrete facts in the things they write for fear of educating their readers.
OpenGL, a Primer, Edward Angel. It's short, easy to understand, and tells you what you need to know to get reasonably complicated OpenGL programs working, without requiring prior knowledge of GL.
-jim
Re:Old news... (but still very cool)
on
Perfect Digital Skin
·
· Score: 4, Informative
This is also mentioned in his
book, (2001), which I highly recommend to anyone interested in raytracing. It's short and about as easy to understand as photon mapping could possibly be.
He has a lot of stuff on his webpage, too, including videos of computer-generated smoke, light through translucent materials, and a good global illumination demo.
For a simpler explaination of what this is all about, there's a photon mapping entry at wikipedia.
Beats me, but there's a paper if you're interested.
Note: raytracers can be slow, but doubling the complexity of the scene does not double the rendering time, as with scanline renderers, assuming the raytracer uses a reasonable bounding volume system.
Raytracing is slower than scanline rendering. Period. No exceptions. If you can come up to an exception to this, I would love to see the renderer that can raytrace a scene faster than Blender could render it.
How about this?
~1 billion triangles, interactive framerates.
I'm not really a graphics expert, but I'll go out on a limb here and say that the trouble with using hybrid scanline/raytracing rendering is that scanline rendering has worse computational complexity.
Rendering a grove of oak trees (down to the very last leaf and twig) from the light source may take a lot longer with scanline rendering, and more memory compared to a raytracer, particularly one that allows the oak trees to be defined recursively.
Incidentally, there's a clever trick with photon mapping that allows you to avoid tracing shadow rays at all: whenever a photon hits something, you project a photon with negative intensity onto the object behind it. When you add the negative value from the photon map to the plain diffuse lighting of the object, you get a proper shadow. I don't know if it's any faster than plain shadow ray testing, but its a cool idea nevertheless. (This is described in Henrik Jensen's photon mapping book - an interesting read that I would recommend to anyone who wants to know about global illumination.)
Scanline rendering might be good for texture intepolation between sample points or other tedious tasks, though.
On to Final Fantasy quality graphics (or better)!*
I'm increasingly convinced that the next big revolution in graphics is going to be real time raytracing. You can only do so much by increasing polygon counts, but a good ray tracer can do so much more: accurate reflections (environment mapping doesn't count) off of multiple objects, soft shadows, motion blur, focal blur, global illumination with photon mapping, subsurface light scattering, volume rendering, simple CSG, smaller memory footprint for a similar model, and better computational complexity: polygonal renderers are O(N) with the size of the scene, whereas raytracers are O(log N), assuming a good bounding volume heirarchy. Eventually, they will catch up and become faster than polygonal renderers.
-jim
* - FF was a pretty good movie, but alas, good technology can't make up for the limitations of the plot.
Slashdot Mirror
The absolute power limit is 1 watt for unlicensed 2.4 ghz equipment. The EIRP limit is 4 watts for point-to-multipoint, but can be higher for point-to-point. (Disclaimer: I'm not an RF engineer, so read the actual fcc regulations before you try it.)
more information
-jim
You can find transmit power / recieve sensitivity numbers for a lot of prism2 cards here. I suspect most of these numbers are gleaned from product brochures, so you mileage may vary.
-jim
Well, I'm running Mandrake 10.0, httpd is running, and I can still post to slashdot, so apparently it must be fixed.
-jim
IIRC, the maximum tx power in the 2.4ghz band for unlicensed users is 1 watt, or 4 watts EIRP. For point-to-point links, though, you can trade 1db of power reduction for 3 db of antenna gain, allowing much higher EIRP.
More info is here
-jim
Try looking at hyperlink, superpass, and netnimble for comparison.
-jim
"Omnidirectional" is somewhat of a misnomer. Omnis transmit a pancake-shaped signal -- good signal in all directions in the same plane, but very little signal up and down. What you're refering to is called an isotropic radiator.
-jim
The super cantenna is only 12 db. 17 is more impressive, and should result in greater range.
Range itself is hard to compare, as it depends on environment, the radios used (cheap 35 mW? 200 mW with good receive sensitivity?), whether the same antenna is used on both ends, and the subjective evaluation of what exactly constitutes a "useable signal".
-jim
Thanks for the input, it's quite informative.
-jim
I'd be interested in panel displays with no trim on the side so they could be placed adjacent to each other for a larger screen. Does anyone know if that's possible with current technology, or if anyone makes that now? (Okay, okay, what I really want is something I can roll up like a poster, but I don't expect that to happen any time soon.)
-jim
Wow, ~20mbps download. Thanks.
-jim
Someone has to mention it sooner or later, and it might as well be me: LaTeX!
I've been using latex for a few months, including writing a research paper. Placement of figures can be a pain, but for text it is reasonably intuitive. All the formatting trivia is abstracted away, and you can concentrate on just writing. It's math features are nice, too, if you need them.
-jim
How about "Church of the Heroic Cranial Ellipsoid" ?.
Maybe this will be a less controversial subject once the new Zelda comes out. On the other hand, I still occasionally talk to people who don't consider the Adventures of Link a real Zelda game because it was a side scroller.
-jim
Even with warping, it still took too long to get anywhere, in my opinion.
Was the remake of Majora's mask etc good? I was a little sad that they didn't increase the framerate a little for Ocarina of Time, but it was still fun to play the Master Quest version even though it made my eyes hurt.
-jim
I've been wondering why there isn't a good "do it yourself community fiber network" howto, at least as far as I know. I've been experimenting with fiber as a hobby for the last few months just for fun. Its amazing what you can find on ebay.
A few observations:
Single mode is vastly superior to multimode in terms of both range (20-100 km vs 2km) and theoretical maximum throughput (terabits per second versus gigabits per second).
On the other hand, single mode is harder to work with, and the end point equipment is more expensive. A new 100mbps fiber-to-copper ethernet converter is about $150 for MM and about $250 for SM. Gigabit gear is more expensive, but not terribly prohibitive.
The fiber itself is cheap (one article I read indicated that the wholesale cost of SM fiber is about $15 a km). All the cladding and armor they put on it makes it expensive (A dollar or two per foot for direct-bury cable with a dozen or so fiber strands).
It takes 2 fibers to make a connection, usually. This is called duplex.
It is possible to run multiple connections on different wavelengths. This is called wavelength division multiplexing. DWDM systems sometimes have over a hundred separate channels.
I don't know much about durability.
Ethernet over fiber may be good for small networks, but it requires active electronics (and reliable power) at each junction. Depending on application, this may not be a problem.
Take a look at fiberdyne's webpage if you're curious about approximate equipment costs. They seem to sell almost everything related to fiber. Here's another page with a decent fiber tutorial.
-jim
Because its hard to take a hero seriously who's head is a perfect sphere. (I have nothing against non-photorealistic rendering, it just needs to be done right. Windwaker was an experiment. They got most things right, but they also got a few things wrong, including Link's head and the whole sailing-forever-to-get-anywhere thing. And no, I'm not sorry I bought it, but it from my perspective it could have been better than it was. I expect this to be the perpetual vi versus emacs debate of the Nintendo world for quite some time to come.)
I just watched the trailer. Very nice. I'm looking forward to it coming out.
-jim
Good, I'm looking forward to reading it.
-jim
This is one of the best collections of graphics algorithms on the net I'm aware of:
comp.graphics.algorithms FAQ
Another favorite of mine is Ray Tracing News, but there haven't been any new issues in a few years.
What other people's favorite collections of algorithms?
-jim
As other replies said, this is an easy read. Just to sum up for the lazy, though:
Most of the text relates specifically and exclusively to audio CDs. Specifically, it will require CDs with copy protection to be labeled as such.
The last bit is more general though:
It seems this allows certain kinds of research, in addition to making it legal to use and make products that circumvent copy protection if the purpose is not copyright infringement. It seems that breaking a copy protection technology for the purposes of copyright violation is still illegal under both the DMCA and plain old copyright law (but someone correct me if I'm wrong).
I'd also like to point out (because I like to pick nits) the shortage of details in the 321 studios press release:
It would have been a simple matter to explain how the DMCRA would re-affirm fair use and balance. You'd think people these days were afraid of using concrete facts in the things they write for fear of educating their readers.
-jim
No, but there is one that models the effects of relativity.
-jim
OpenGL, a Primer, Edward Angel. It's short, easy to understand, and tells you what you need to know to get reasonably complicated OpenGL programs working, without requiring prior knowledge of GL.
-jim
This is also mentioned in his book, (2001), which I highly recommend to anyone interested in raytracing. It's short and about as easy to understand as photon mapping could possibly be.
He has a lot of stuff on his webpage, too, including videos of computer-generated smoke, light through translucent materials, and a good global illumination demo.
For a simpler explaination of what this is all about, there's a photon mapping entry at wikipedia.
-jim
Beats me, but there's a paper if you're interested.
Note: raytracers can be slow, but doubling the complexity of the scene does not double the rendering time, as with scanline renderers, assuming the raytracer uses a reasonable bounding volume system.
-jim
~1 billion triangles, interactive framerates.
-jim
I'm not really a graphics expert, but I'll go out on a limb here and say that the trouble with using hybrid scanline/raytracing rendering is that scanline rendering has worse computational complexity.
Rendering a grove of oak trees (down to the very last leaf and twig) from the light source may take a lot longer with scanline rendering, and more memory compared to a raytracer, particularly one that allows the oak trees to be defined recursively.
Incidentally, there's a clever trick with photon mapping that allows you to avoid tracing shadow rays at all: whenever a photon hits something, you project a photon with negative intensity onto the object behind it. When you add the negative value from the photon map to the plain diffuse lighting of the object, you get a proper shadow. I don't know if it's any faster than plain shadow ray testing, but its a cool idea nevertheless. (This is described in Henrik Jensen's photon mapping book - an interesting read that I would recommend to anyone who wants to know about global illumination.)
Scanline rendering might be good for texture intepolation between sample points or other tedious tasks, though.
-jim
On to Final Fantasy quality graphics (or better)!*
I'm increasingly convinced that the next big revolution in graphics is going to be real time raytracing. You can only do so much by increasing polygon counts, but a good ray tracer can do so much more: accurate reflections (environment mapping doesn't count) off of multiple objects, soft shadows, motion blur, focal blur, global illumination with photon mapping, subsurface light scattering, volume rendering, simple CSG, smaller memory footprint for a similar model, and better computational complexity: polygonal renderers are O(N) with the size of the scene, whereas raytracers are O(log N), assuming a good bounding volume heirarchy. Eventually, they will catch up and become faster than polygonal renderers.
-jim
* - FF was a pretty good movie, but alas, good technology can't make up for the limitations of the plot.