BQ35100: FAQs, profile, EOS, calibration

Sorry for so many questions. (also so many other threads with somewhat related confused users)

1. Data sheet 8.2.2.6 Expected Device Usage Profiles + tech ref etc
   1. need for 5 hours rest - "Periods in between risk not allowing the battery to rest and any EOS-related data may be
      compromised." : How can one predict BQ35100 EOS performance with low but more frequent loads on the battery? e.g. is there a use case at 30min, 1 hr, 3hrs or is that just a deal breaker for all of them regardless?
   2. what does "for 10 s of ms" here mean?
2. tech ref 5.3 EOS - 
   1. ?When? does  the device have enough  information as per : "When the device has enough information to update Impedance, the present value of Impedance is
      copied to Previous Impedance in preparation for the new Impedance value to be updated."
   2. EOS trend detection - default is 20. Are customers (sometimes) to change this? where is this explained?
   3. note the typos as trent detection.
   4. note the min isn't really 1 if the default is 0, but are R short trend filter and R long trend filter explained somewhere else?
   5. I dont follow EOS smoothing. it doesnt seem to come up anywhere else. EOS mode does SOH first. we're not supposed to switch modes, but does Lifetime Data have to be on the whole life or only after EOS alert? 
3. Calibration
   1. the datasheet gives no indication that this is a step for each and every unit. We were quite (too) far into our evaluation before finding those details
   2. Is it correct that calibration done on the EVM via BQstudio with our batteries is irrelevant to what matters here? it needs to be done on our pcbs?
   3. I gather the +/-  125mV sense range is used to scale this calibration, so the sense resistor is selected via the largest measured currents, and a similarly large measure current is then applied at calibration? If this is correct - why does the techref 3.8 say typically use 1000mA?
   4. does calibration affect SOH more than it affects EOS?
4. Initial / Operating conditions
   
   1. 7.3.2.3.1 Initial EOS Learning - If EOS is triggered via the long and medium term impedance ratio - why is this so? just how compromised are you saying? how does this work with batteries that might be depassivated via loading them up or with batteries that aren't yet depassivated?
      "For optimal accuracy, the first event where the device updates its impedance value is required to be when the
      battery is full (a fresh battery). If the battery is partially discharged, then the accuracy of the EOS detection is
      compromised."
   2. If the orientation of the battery changes during operation (which affects some LiSoCl2 batteries more than others) - does EOS still work?
   3. for measurement with consistent pulses - just how consistent are we talking? assuming a radio based device awakes after rest as per 1a, some transmits / connects  will naturally vary. Should variable parts of load/cycle be avoided? Then should measurements (GE high) only be made when predictable (but now lower load for these kinds of devices)?
   4. Related to above : I haven't understood if only a subset of the load needs to be observed via GE high, or all/most pulses should be observed?

Thanks in advance