Support for store-and-forward

[hzbarcea]

Is it in the scope of the spec to support a scenario where a notification receiver is temporarily offline, yet the notification sender stores it (with at ttl) and streams then once the notification receiver reestablishes connection?

[csarven]

I would say handling message failures is an acceptable case without implying that the Notification Protocol needs to define a retry policy.

IIRC, there was some discussion/notes along the lines of continuing messages from where they were left off - need to dig that up.

[TallTed]

I would certainly hope that the SNP (Solid Notification Protocol) would have robustness similar to SMTP or NNTP. I will also note that SMTP and NNTP routing interruptions of years have occurred, with messages continuing on their way upon restoration of communication (i.e., when the relevant machine(s) were rebooted and/or hubs/switches/routers reconnected), and that even this delayed delivery was appreciated by the recipients/readers.