In #12, Vladimir implies that smooth software scaling is too slow. Is that claim still true in 2010? (And was it ever true?) Scaling can be done with a dozen instructions per pixel, at the very most. Considering the cost of decoding JPEG and PNG, I have a hard time believing this would be "slow". Or even noticeable on any desktop or laptop built in the past 5-8 years.
Chromium uses software scaling, and it is damn fast.
There is more and more discrepancy between webpage "sizes" these days. At one extreme, http://msdn.microsoft.com/en-us/library/default.aspx uses very small fonts (apparently using an absolute 'pt' size), and requires scaling on many setups, and good citizens, like google.com that simply use '100%' fonts.
Webpages that are scaled to make the text readable end up with horrid images with Firefox on Linux, this is a shame. Two years after the original report, the approach first considered is not getting anywhere.
In #12, Vladimir implies that smooth software scaling is too slow. Is that claim still true in 2010? (And was it ever true?) Scaling can be done with a dozen instructions per pixel, at the very most. Considering the cost of decoding JPEG and PNG, I have a hard time believing this would be "slow". Or even noticeable on any desktop or laptop built in the past 5-8 years.
Chromium uses software scaling, and it is damn fast.
There is more and more discrepancy between webpage "sizes" these days. At one extreme, http:// msdn.microsoft. com/en- us/library/ default. aspx uses very small fonts (apparently using an absolute 'pt' size), and requires scaling on many setups, and good citizens, like google.com that simply use '100%' fonts.
Webpages that are scaled to make the text readable end up with horrid images with Firefox on Linux, this is a shame. Two years after the original report, the approach first considered is not getting anywhere.
Please reconsider this question.