[trustable-software] t.opinion isn't t.feedback

Paul Albertella paul.albertella at codethink.co.uk
Thu Nov 28 10:03:33 GMT 2019


I wrote:
 > Do we need to have two terms, to allow us to distinguish between these
 > two types of input, or can we come up with a term that encompasses
 > both?

After discussing this with a colleague (Daniel Silverstone, take a bow), 
we have another suggestion:

1) We retain t.opinion, but explicitly define it as a recorded statement 
from a human relating to a t.proposal, which is necessarily subjective.

2) We add a new element called t.score, which represents recorded input 
to an evaluation process that can be mechanically processed. This may be 
human in origin (e.g. a record of a show of hands in a meeting, or an 
approval registered through a code review system), or it may come from a 
computer system (e.g. pass/fail logged by an automated process, based on 
the results of a test execution).

* A t.score MAY be supported or accompanied by a t.opinion.
* A t.score MAY be supported by other recorded evidence (e.g. full test
   results or a log file).

I hope that this makes a clearer distinction between the different 
classes of evidence that we want to consider, while being sufficiently 
generic to enable us to model any type of process.



More information about the trustable-software mailing list