I'm afraid that there is no Chrome data set, only the frequency table result (which isn't useful here). The browser was instrumented to aggregate character counts without collecting raw data. On Sun, Aug 3, 2014 at 1:53 PM, Martin Thomson <martin.thomson@gmail.com> wrote: > > On Aug 3, 2014 10:24 AM, "Jeff Pinner" <jpinner@twitter.com> wrote: > > The huffman encoding was derived from Chrome. I'd be more interested > > if we could get numbers from the headers internal headers emitted from > > say GFE rather than from those emitted from Chrome. > > I'd have said that GFE data would exhibit biases peculiar to Google > properties that might be detrimental to other users. More so than the > Chrome data. > > The data set we got from the Chrome experiment might be less well suited > to private use - or probably more importantly, API use - than might be > ideal, but I don't see any easy way to avoid bias in collecting data from > other sources. >
Received on Monday, 4 August 2014 18:12:56 UTC