On Mon Aug 18 2014 at 10:27:12 AM Ilya Grigorik <igrigo...@google.com> wrote:
> On Mon, Aug 18, 2014 at 12:22 AM, Sigbjorn Finne <s...@opera.com> wrote: > >> Thanks much for clarifying. For implementations that don't delay >> initiation (I know of none that do, but perhaps there are), this would >> amount to "Beacon-Age: 0" for all practical purposes. Which brings up the >> second question raised initially - what benefit does this bring? >> > > Initial implementations don't delay, but I think we should (*strongly*) > encourage this UA behavior - energy savings on mobile, etc. To make things > easier in the long run, we should also make it clear to analytics vendors, > etc, that this is the intended behavior... Hence the header, even if its > "0" today. > > That said, I'm torn on the "Beacon-Age: 0". On the one hand, saved bytes > are good, but I do think it could make things a bit more difficult.. If > some UA decides to skip the header (bug or otherwise), then as a developer > I have to resort to UA detection (ugh) to winnow out valid vs. invalid > cases. A guaranteed header makes things much easier. > You don't need UA detection. Just interpret the lack of a Beacon-Age header as zero. Adam
