Closed toddbaert closed 11 months ago
I think it would be helpful to include glossary entries for client-side
and server-side
that reference static/dynamic context, respectively. That would allow us to refer to SDKs as being either client-side
or server-side
while still keeping the more technically correct terminology in the spec.
This PR adds new points to support client use-cases. It does this primarily by defining and specifying a new static-context paradigm.
There's a few requirements that now "fork" on the condition that the implementation targets the dynamic context (server) or static context (client) modes. There's no functional changes specified for dynamic context implementations.
For more justification on the above, see the issue here, and the OFEPs linked therein.
Please note that there is some "number thrashing" with these changes. I tried to minimize it, but I think it's more important to have a coherent and sensible spec than to make sure to maintain the mapping between section numbers and content. I found that RFC2616 changed the content of their numbered sections revision to revision as well, so this isn't unprecedented.