[Ssc-dev] incorrect location on infographic + old location cache data in j3m

Nathan of Guardian nathan at guardianproject.info
Mon May 18 22:49:19 EDT 2015


I think we need to clear the existing cache right when we start the
camera, to make sure there is no lingering data. In addition, we can
possibly filter all the events in the log to remove items outside of a
specified time window. I will look into this before Wednesday.

On Mon, May 18, 2015, at 10:23 PM, Martin Tzanev wrote:
> Just a bit more on this: the pic was taken on Sunday in Inwood, but the
> infographic shows location Brooklyn. I hadn't been in Brooklyn since
> Friday
> and I took a bunch of pics on Sat in Staten Island (with regular camera,
> not cameraV) which all show correct location. So, I'm wondering where
> camerav is pulling that old location from, since it seems that the phone
> had updated it's location multiple times over the weekend.
> M.
> On May 18, 2015 6:32 PM, "Martin Tzanev" <martin at witness.org> wrote:
> 
> > Hi Nathan,
> >
> > Attached is an example of a pic & j3m that contains 2-day old location
> > data.
> > The exif locatioin is correct and at some point the location data in
> > j3m updates to correct too, but I see three issues here:
> > 1. how can we avoid collecting old location cache data
> > 2. the infographic should display the locations closer to the time the
> > pic/video was taken (currently it displays the location from two days
> > earlier)
> > 3. this infographic shows just a dot even though there are multiple
> > location points in j3m (for other media, the infographic shows a line
> > but locations shown are from a video that I shot like 10-15 mins
> > earlier)
> >
> > I got a lot of this over the weekend, most were not as dramatic as two
> > day old data more like 15-20 mins earlier, so the videos that I shot
> > this weekend, all got inaccurate locations in the infographic.
> >
> > I have a lot more feedback that I will right up tomorrow.
> > m.
> >


-- 
  Nathan of Guardian
  nathan at guardianproject.info


More information about the Ssc-dev mailing list