GTFS-rt - are stop_time_update.SKIPPED values propagated downstream?

[barbeau]

See https://groups.google.com/forum/#!topic/gtfs-realtime/teaI9X1FoGE.

[barbeau]

An analysis of how many agencies use the SKIPPED value for any StopTimeUpdate would be useful.

[minhhpham]

We performed an analysis using transit-feed-quality-calculator on 9/29/2018 and found 12 agencies that use SKIPPED value for any StopTimeUpdate:

Feed	Location
Metro Transit Trip Updates	Madison, WI, USA
Thunder Bay Transit Trip Updates	Thunder Bay, ON, Canada
Nashville MTA Trip Updates	Nashville, TN, USA
OVapi Train Trip Updates	The Netherlands
OVapi Trip Updates	The Netherlands
CT Transit Hartford Trip Updates	Connecticut, USA
VIA Trip Updates	San Antonio, TX, USA
ETS Trip Updates	Edmonton, AB, Canada
YRTViva Trip Updates	York, Toronto, ON, Canada
MetroTransit Trip Updates	Halifax, NS, Canada
Kingston Transit Trip Updates	Kingston, ON, Canada
MBTA Trip Updates	Boston, MA, USA

The branch of this analysis is located at https://github.com/CUTR-at-USF/gtfs-realtime-validator/tree/v26-schedule_relationship-SKIPPED

[minhhpham]

Snapshots of these feeds are available at https://gist.github.com/minhhpham/65dafc594ff8a1fafca30fe0319eed88

[minhhpham]

An update when contacting what agencies mean when using skipped:

• Kingston, Metro Transit, ETS, ThunderBay, OVapi, VIA, and MBTA provide entire stop_sequence in their trip_update *i.e., StopTimeUpdates for all stops in the trip), so their intention is clear (i.e., nothing is propagated).
• CT Transit: confirmed SKIPPED not propagated
• MTA: confirmed SKIPPED not propagated

[barbeau]

I've opened a proposal PR at google/transit#139

[barbeau]

google/transit#139 was approved - new defined behavior is that SKIPPED values are not propagated, but delays (including those calculated from time values) are propagated over SKIPPED stops.