RDF Messages spec draft for tomorrow's meeting

Dear RSP CG members,

I would like to submit something to be discussed at tomorrow's meeting
(that I won't be able to join, sadly). It's a spec draft for "RDF
Messages", a simple set of abstractions for handling sequences (streams) of
RDF datasets: https://www.pieter.pm/rdf-messages/

This spec is intentionally lightweight and does not concern itself with
time, as time is not relevant to the entire streaming tech stack (e.g.,
message brokers, serializers, some event processors). We imagine that specs
that do introduce the notion of time would be built on top of this.

This is intended to cover a very wide variety of use cases, essentially
everything that imec/Uni Gent has encountered with LDES, and everything
that NeverBlink has encountered with Jelly. This includes non-real-time
streams, or "logs", such as lists of Nanopublications. Of course, this also
does include real-time streams, such as those coming from IoT devices.

This draft was started and led by Pieter Colpaert, I helped here and there.
Pieter Bonte provided very useful feedback for us. Our reasoning is
documented in closed GitHub issues:
https://github.com/pietercolpaert/rdf-messages/issues?q=is%3Aissue%20state%3Aclosed

Like I wrote, I won't be able to attend tomorrow's meeting, as I'm busy
with ISWC. In my place Anastasiya from NeverBlink will attend. In short, my
stance is that we should form a TF around this spec draft. I would consider
it also reasonable to form other TFs that would develop specs based on this
one. I can offer to lead the TF for RDF Messages, but I would also be happy
to give up the mantle to anyone else that is willing to do so.

I think the important work to be done aside from what we propose here would
be (1) to define serialization formats for RDF Message Logs, and (2) define
profiles for higher-level interpretations of RDF streams, that do include
time.

Warm greetings from Nara!

-- 
Piotr Sowiński
Co-founder @ NeverBlink
https://neverblink.eu