It might still be convenient to consider, eg. whether a future
implementor that *does* use CoMI can just treat your /j resource as
an
alias to /c/abc0815 and be done with it, because a POST there with
the
binary data has the desired effect.
I am not very familiar with CoMI so it could be a dummy question, but
would we need to specify in the draft to allow this?
In principle the the length of the path will be shorter to start off
with. eg /c/FG0
Interesting thought to reduce the already reduced path.
If there are only so many comi data nodes in the server, say 4-5, then
the server might make
referral to the comi data entries form the short path's like /j, or /j1
till jn if there are n data points.
the client should also do this translation in the other direction from
/c/FG0.
Just a note in the text that this a convoluted possibility would be
sufficient IMO.
Peter
_______________________________________________
6tisch mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6tisch
