I think we're getting to a point where a distributed,

I think we’re getting to a point where a distributed, un-censorable object repository system will become a necessity Thingiverse is great, but we’ve already seen that they are willing to compromise content due to external pressures.

I believe any single site, system or company is vulnerable to this sort of pressure and furthermore I believe it is their right to choose what they host.

However, if history has taught us anything it is that allowing information to be at the mercy of any level of tyranny can lead to loss of knowledge; due to accidents, coincidence or in worse cases deliberate destruction.

What I propose as a remedy to this is a system of federating this information instead of relying on any single entity, vulnerable to the systemic weaknesses present in any type of organization known today. A federated system, distributed across physical and political regions in order to avoid constraints imposed by any type of external method of control or coercion.

At a less abstract level, this amounts to object repositories with standardized access methods and standardized catalog data that can be easily replicated to allow for indexing and searching across the entire federation of repository sites.

The technology for this certainly exists (either in parts or perhaps in whole), and I’m looking for feedback from this group to determine if perhaps this work is already underway. I am aware that there are other sites like Thingiverse out there, but at this point I’m unaware of any deliberate effort to publish their content in a way that allows for global search across sites (without relying on other centralized corporate entities like search engines) or replication designs intended to distribute models across sites in an automated fashion (necessary to prevent things from disappearing when a site or object is taken down voluntarily or by force).

I believe now is the time to design these standards, while the proliferation of “thing” repositories is small and the control mechanisms are still in their infancy. If we wait too long, we’ll all become too attached to our personal favorites and implementing a federated system will become more difficult. Doing something now also makes it harder for those who would desire to curtail such an effort if it has already demonstrated its value to a worldwide community.

I’m planning to elaborate on these ideas and investigate implementation options in the near future, so any information about simular or overlapping projects you can share is appreciated.

Consider looking into the formation of Archive of Our Own. Www.ao3.orgthrew were looking for a space for definitive works mostly fiction. But wanted many of the things you’re concerned about.

Not much there (and most of what is there is miscategorized, and doesn’t belong there), but when I think of un-censorable repositories, the pirate bay immediately comes to mind. They have a “physibles” section created for this express purpose: http://thepiratebay.se/browse/605

It’s laid out as a repository nearly as well as thingiverse or anything I would consider a potential replacement for it, but it’s good enough as a place to link to for things that have been banned everywhere else.

Good idea. Not in my destiny, but it’ll be a great addition if someone does it.

Thanks for the pointers @Laura_Thomas & @Whosa_whatsis , I’ll keep you in the loop @Victor_Didra as things progress.

Git is one option I’m looking hard at but so far I haven’t found a way to make it work in a multi-master arrangement that doesn’t require centralized control. That said the version control is excellent and integration options abound so it might be a good place to start. I’m also looking closely at bittorrent to provide an efficient storage replication layer.

I need to learn more about TOR. One key aspect of this is that we don’t loose the accessibility and convenience of Thingiverse,

I’m less concerned about raw performance than convenience. I think between a transparent replication scheme and a shared, distributed index we’d be pretty close to the target. If it was simple to search across all available sources then the index could provide a means of resolving the location source of the data (and handle failed lookups if cleverly implemented).

Thanks for the ideas, pointers and encouragement. I may be able to piggy-back this research on some other projects I have in-flight that have simular needs.