Thank you all for the responses.
I am convinced that this type of connection should be deemed a PR connection
unless it can be shown otherwise without tests. The fact that a channel is
used, and the fact that there is no stiffener plate at the column, both mean
that the load path is not at all straightforward, and it is not readily
apparent that the connection can develop the strength of the beam.
This means that testing of this connection is required to determine its
strength and stiffness. We all know this can be quite expensive.
The main dilemma seems to be as Charles Greenlaw described. In this case, a
loosely-defined differentiation between two categories, which was made by
AISC, was presumed in the UBC to be a well-defined differentiation. I
suppose it was a compromise to prevent the requirement that ALL connection
configurations be tested.
Mark Swingle, SE
These are my own opinions....