SERVICE SILENT results

A report and discussion [1] about SERVICE SILENT centres around the fact 
that when the SERVICE calls fails, the effect of SILENT is to return a 
single row of no variables.  Some people are finding this counter 
intuitive, expecting no rows to be returned instead as if no matching 
occurred.

http://www.w3.org/2009/sparql/docs/fed/service.xml#serviceFailure
"""
The failed SERVICE clause is treated as if it had a result of a single 
solution with no bindings.
"""

Is there a reason for this particular design?  I can't recall one and I 
do find the expectation of no rows more logical.

I am guessing but may be the intention is that

SELECT *
{
    SERVICE <S> { ....} # SERVICE call fails in HTTP
    ?s ?p ?o
}

potentially returns something but now it will be an unconstrained 
something from "?s ?p ?o".

	Andy

[1] https://issues.apache.org/jira/browse/JENA-261