Starting to get my head around this again (I think :).
Correct me if I get any of this wrong +Gary Hodgson
The heart of TTN is essentially the JSON metadata description of a "thing", which can be made available in any way that the document can be served up. Everything else we've been discussing (the DHT index, client software, BitTorrent delivery of files, etc.) is sort of conveniences and accessories to a description standard that could span any method of hosting and sharing models, would you agree?
If so I think a focus on nailing down the spec should be first order of business, since there's existing tech that can handle the rest (even if it's not ideal), and once that's established tools for generating the JSON (or facilitating the generation of it) should probably come next.
I think that the distributed index stuff and all that is super-important long-term, but now that I see how close we are to a usable foundation I'd love to start seeing that data in the wild, and learning what tools are needed by makers to adopt sharing this way into their workflow.
I also wonder if theres more existing tech we could co-opt to save development time. One thing that occurred to me tonight was that the TTN JSON is reminiscent of RSS with attachments, and it made me wonder if we could apply podcasting tech to the problem, either directly or conceptually? It seems like podcasts follow a similar publishing model (at least originally) and share similar discovery and syndication issues. I also thought it would be so cool to be able to subscribe a printer to a makers feed and have it print new things automatically as soon as they are published :)
I thought a lot about leveraging BitTorrent protocol for delivery of model files, and I do think this is the way to go, even if it means having somewhat specialized nodes out there that participate in the distributed back-end network but present a more traditional web interface for users who do not have the ability to download and seed a BitTorrent network.
This got me thinking about BitTorrent trackers and how to avoid them, Id love to be able to have something like a Piratebox that could synchronize with the network, host gobs of thing files and then be pulled from the Internet and used independently to continue serving.
Let me know if this is in line with your thinking, or maybe Im way off course :)