Re: [gpsd-dev] PPS over USB

[tz]

I think it is something like your attempt to stastically clean up USB polling jitter - in the case of the real world, we can get statistics and derive information.

But even more, this is like having weather stations.  Weather cannot be studied in a lab.  You need thermometers, anemometers, hygrometers, and barometers scattered over a wide area.  You can calibrate instruments in a lab, but they aren't useful for observing as there is no weather in a lab.

The delay might vary with time or have something unexpected.  The internet design has a lot of compensation and error correction which masks problems.

On May 8, 2012 6:07 AM, "Ed W" <address@hidden> wrote:

On 08/05/2012 08:58, Dave Taht wrote:

The internet escaped the lab a long time ago. Yes, lab work is needed.
But for a change, I'd like the reality to be tested, and the lab work
to line up with reality. And the tools used, to be trustable. And the
results, repeatable.

I think we agree except on the point that the real world is repeatable.  That's the very thing which it isn't in my opinion and where the models come in.  Lets not let this deteriorate into postulating the grand theory of everything - what I mean by models is usually extracting some facet of reality into a repeatable situation and then analysing just that piece

Ed W