Opened 11 years ago
#34 new enhancement
Semantics of History-Info values need to be documented explicitly
Reported by: | worley@… | Owned by: | |
---|---|---|---|
Priority: | critical | Milestone: | milestone1 |
Component: | rfc4244bis | Version: | |
Severity: | In WG Last Call | Keywords: | |
Cc: |
Description
The draft doesn't give a clear specification of History-Info *as a data structure*, that is, if one were to look at a request containing an H-I header, (1) How would one verify that it had been constructed correctly? and (2) What are the meanings of the various fields (in particular, their relationships)? Many of the components of History-Info are described, but their descriptions are not as exact as they need to be.
Presumably the details of the "procedures" set how History-Info actually works -- but that is the same problem as having a data structure in a program that is only "documented" by the code that modifies it. In particular, if there was specification of H-I as a data structure, that specification could be compared against the procedures to verify that both the data structure specification and the procedures are correct. Or more importantly, the reader could compare their interpretations of the two to verify that their understanding of both is correct.
(I am willing to help write this.)