1) Thanks for this test. So far I see no results that would appear to be a hard stop. But in the other thread, you mentioned that your Epic GPS accuracy was not as good as your G1. So I am still concerned. I think performance on this new, expensive, high-end Android should be at least as good as the performance of the oldest Android model.
Also, driving is the least demanding test of the low-level GPS, and there are many applications for it besides vehicular navigation. On other Galaxy S platforms, the GPS tracks seem to get smoothed or interpolated, which in driving tests seems
okay most of the time because the motion of the vehicle ovewhelms the underlying errrors in the smoothing algorithm. The result is that the GPS seems great while moving pretty straight, but can veer off-course on turns, corners and stops. You did mention similar symptoms. (And then navigation programs mask errors further by snap-to behavior layered onto the smoothed GPS output.)
I know it must seem that I am never satisfied, but I am always interested in more rigorous testing. Often a better My Tracks test is over a walking course, preferably along with a benchmark device such as the G1. And the whole track -- not just isolated snapshots -- then can be shared with us via upload to Google.
For example, look at my own tests of the Vibrant vs G1 here
(If you are worried about your own privacy, just test in some location other than your home. I went to an apartment complex in my neighborhood.)
2) Another set of GPS-bug symptoms reported on other Galaxy S platforms is the failure to use and lock on to enough satellites. That is best seen not through tracking or navigation apps, but by utilities that read and report the detailed performance of the GPS from the Android system. The two most popular utilities for this testing are GPS Test and GPS Status. Among other things, they will tell you how many satellites the GPS thinks it sees, their signal strength as SNR, and how many are actually being used for a fix. Could you try those utilites and report the results?
3) You haven't told us exactly how your test unit came into your hands. But we do know from other intelligence that at least some pre-sales demo units were selectively shipped before the latest Samsung software was installed. So it is useful to know the build date of your unit, as shown by internal file date stamps. (Perhaps this can clear up some confusion about other reported tests.)
4) From the latest build dumps we have seen, it appears that the Epic -- uniquely among Samsung S variants -- has a system utility called GpsSetup. Can you see any evidence that this utility is installed on your unit? If so, what does it do? Are there configurable settings? If so, what are the default settings, and what settings are you using?
Thanks again for your work.