... not a search engine. In fact it submits the search to MSN, Netscape, Lycos and Looksmart and clusters the result into categories.
My question is: if it really is interested in providing the future of web searches -- why not cluster on top of google results which are accepted to be *the* best? Them not using google underneath and instead using Lycos etc shakes my confidence in them.
Compressed XML is comparable in size to binary formats. I remember there being some comparision of XML representations of objects in compressed format and the same objects in binary representations.
XML is highly compressible. Not only are <=/> characters and tagnames are repeated all over the place, being mainly in printable ASCII means most characters use 7bits.
Throwing a compress/decompress around the RPC process often adds very little extra processing and gives you the advantage of a very extensible, self describing, structured data.
The gamma reflecting surface is the inner surface of the bomb container -- which is about a feet or so clear of the imploding ball. So the implosion as such would not harm it. I assume you mean the explosion would destrow the reflector -- which is true; however particulate matter from the explosion (gasses, debris etc) travel slower than the EM radiation and the billionth of a second difference between the radiation being reflected of the inner surface and the expanding gasses blowing out the shell is enough for the radiation to turn the foam coating around the fussion candle into plasma yielding in a compression of the fusion chamber starting the chain reaction. I don't see anything incorrect here.
We always say security through obscurity is bogus. Case in point -- closed source software, squlching of bug/expolit reports, use of the DMCA to silence hackers instead of fixing the exploits et al.
When it comes to national security, what makes people think secrecy makes the nation any more secure?
Planet-Lab is not an emulator. It is a group of computers distributed across the globe that use the real internet. People write distributed apps and run them on different vantage points spread out on the net to measure real internet performance, test their app etc.
Unless MS releases an official MSN client for Linux you should read the ban as -- "Linux users: pay $199 to buy Windows and MSN messenger or else pay DMCA fines."
Do you suppose people will bitch if hotmail became a paid service?
Whops. Mod me down on that last one. The image I described was the opaque image that is being used as a background, and the bufferring threw off the printf vs. x sync I am guessing. On closer examination... imlib does seem to work... but doesn't display anything while its doing stuff... only the final image.
I noticed that as well. I commented out test 1 since with renderaccel it was at par with imlib, but test2 was slow. As for test 2, I moved the imlib test to the top, followed by the xrender and then xrender offscreen.
Surprise. I think I see something I shouldn't be seeing. After imlib claims that it is done I suspend the app and in my window I see 4 unequal sized rectangles slightly covering each other. on the top is a black-green diagonal gradient (NorthEast). under it is a light blue-dark blue NW gradient. Under the blue is a black-red southward gradient and covering that is a longish red-yellow Eastward gradient.
Me thnks imlib had a brainfart... or graphics didn't load or something. Can someone confirm this?
The NVidia drivers say something about Render Accleration as someone already pointed out. However, there is definitely some glitch somewhere. I tried the benckmark with the RenderAccel both turned off and on on my GeForce 3 with the 4496 drivers and perceived no significant difference in the tests except for test 1. (11s for no accel, 2.5s for accel, 0.62 for imlib2). The rest of the tests sucked for the driver (11s, 215s, 183s, 356s for tests 2 to 5 -- both with and without render accel as opposed to 0.21s, 4.5s, 2.7s, 5.8s for imlib2).
I use Xinerama with the secondary display on an ATI 98 Pro (Yay for college tuitions). One thing I did notice was that even in render-acclerated mode, if I drag the window to the middle straddling the screen split, the images display on both sides -- though ATI's side is scaled down even at the same resolution for some reason). However, if I use a gl application (glxgears, mplayers -vo gl2 etc) then straddling the screen only gives half a display on the GeForce board. So in this case, X is either not using XRender either because of NVidia drivers or is picking the lower of the capabilities of the video card. Or is doing something in the middle causing the GeForce and ATI displays to be different.
I wonder if there is any way to explicitly force X to use the hardware for XRender as you can do with GL.
Slashdot Mirror
... not a search engine. In fact it submits the search to MSN, Netscape, Lycos and Looksmart and clusters the result into categories.
My question is: if it really is interested in providing the future of web searches -- why not cluster on top of google results which are accepted to be *the* best? Them not using google underneath and instead using Lycos etc shakes my confidence in them.
Compressed XML is comparable in size to binary formats. I remember there being some comparision of XML representations of objects in compressed format and the same objects in binary representations.
XML is highly compressible. Not only are <=/> characters and tagnames are repeated all over the place, being mainly in printable ASCII means most characters use 7bits.
Throwing a compress/decompress around the RPC process often adds very little extra processing and gives you the advantage of a very extensible, self describing, structured data.
The gamma reflecting surface is the inner surface of the bomb container -- which is about a feet or so clear of the imploding ball. So the implosion as such would not harm it. I assume you mean the explosion would destrow the reflector -- which is true; however particulate matter from the explosion (gasses, debris etc) travel slower than the EM radiation and the billionth of a second difference between the radiation being reflected of the inner surface and the expanding gasses blowing out the shell is enough for the radiation to turn the foam coating around the fussion candle into plasma yielding in a compression of the fusion chamber starting the chain reaction. I don't see anything incorrect here.
We always say security through obscurity is bogus. Case in point -- closed source software, squlching of bug/expolit reports, use of the DMCA to silence hackers instead of fixing the exploits et al.
When it comes to national security, what makes people think secrecy makes the nation any more secure?
the Planetlab Internet emulator
Planet-Lab is not an emulator. It is a group of computers distributed across the globe that use the real internet. People write distributed apps and run them on different vantage points spread out on the net to measure real internet performance, test their app etc.
Unless MS releases an official MSN client for Linux you should read the ban as -- "Linux users: pay $199 to buy Windows and MSN messenger or else pay DMCA fines."
Do you suppose people will bitch if hotmail became a paid service?
Check the SCO OpenServer machine. You'll find it.
Whops. Mod me down on that last one. The image I described was the opaque image that is being used as a background, and the bufferring threw off the printf vs. x sync I am guessing. On closer examination ... imlib does seem to work ... but doesn't display anything while its doing stuff ... only the final image.
I noticed that as well. I commented out test 1 since with renderaccel it was at par with imlib, but test2 was slow. As for test 2, I moved the imlib test to the top, followed by the xrender and then xrender offscreen.
Surprise. I think I see something I shouldn't be seeing. After imlib claims that it is done I suspend the app and in my window I see 4 unequal sized rectangles slightly covering each other. on the top is a black-green diagonal gradient (NorthEast). under it is a light blue-dark blue NW gradient. Under the blue is a black-red southward gradient and covering that is a longish red-yellow Eastward gradient.
Me thnks imlib had a brainfart ... or graphics didn't load or something. Can someone confirm this?
The NVidia drivers say something about Render Accleration as someone already pointed out. However, there is definitely some glitch somewhere. I tried the benckmark with the RenderAccel both turned off and on on my GeForce 3 with the 4496 drivers and perceived no significant difference in the tests except for test 1. (11s for no accel, 2.5s for accel, 0.62 for imlib2). The rest of the tests sucked for the driver (11s, 215s, 183s, 356s for tests 2 to 5 -- both with and without render accel as opposed to 0.21s, 4.5s, 2.7s, 5.8s for imlib2).
I use Xinerama with the secondary display on an ATI 98 Pro (Yay for college tuitions). One thing I did notice was that even in render-acclerated mode, if I drag the window to the middle straddling the screen split, the images display on both sides -- though ATI's side is scaled down even at the same resolution for some reason). However, if I use a gl application (glxgears, mplayers -vo gl2 etc) then straddling the screen only gives half a display on the GeForce board. So in this case, X is either not using XRender either because of NVidia drivers or is picking the lower of the capabilities of the video card. Or is doing something in the middle causing the GeForce and ATI displays to be different.
I wonder if there is any way to explicitly force X to use the hardware for XRender as you can do with GL.