I continue to think more about how we make sense of data in various ways with this model. Especially in the face of very long threads of conversation. A combination of aggregates and search seems like a reasonable approach.
The right design for such a flexible system might be discoverable only through evolution!
My trouble seems to be that every case I come up with causes another piece of complexity to be added. At this point I think I need to draw a line. Search can wait. People can code their aggregates to put limits on how much data gets stored in a single entry (threads stop at 1000 for example).
They could build a server aggregate to “pin” messages that tell people to make new threads when they hit the limit — I don’t need to build a special feature for that!
But I’m certain search is coming. It’s only a matter of time. It can wait for a bit tho!
Anyway.. feels good to let go of complications. At least for now. Always possible I’m wrong. 😀