Will this replace the function or some requirements of full nodes?
I will see this as a separate activity, a different type of node. Requirements will be very different so probably will be talking about high redundancy environments like AWS
Would there need to be 20 nodes + backups like the witnesses?
That's TBD, I would see many more content hosts than witnesses. Probably 100+
Would the rewards draw from the 10% witness pool, or would it carve out a new portion of the pool?
I would see a new reward pool designed, to include this as well. It's a new part of the business.
Will this allow for improved content control- such as making particularly reprehensible content inaccessible?
I don't think I understand the question, but this feature shouldn't allow censorship, on the contrary, it should increase the availability of content.
Do you think it's possible for full nodes to be rewarded- or at least incentivized in the near future? And do you think this restructuring would ease they resource and cost burden that just a few full nodes currently endure?
I also believe it would be beneficial to support a greater number of hosts, even with different capabilities.
This may allow people to choose a faster service with smaller storage limitations- and vice versa, as well as reducing barriers to entry by only supporting the highest cost services.
It makes sense to carve out a distinct portion solely for this purpose.
With no form of censorship or control- an unfortunate situation may arise where a party ends up hosting illegal content- to which they could potentially be held responsible.
There is something in the blockchain that prevents that - I don't remember exactly what it is, but when some content is considered illegal or a request to remove it for privacy reasons is received, that content can be removed, there is a transaction for that.
As for incentivizing full nodes, that would certainly be great. It's becoming almost impossible to keep up with the growth. SSD storage needed for a full node is nearing 200 GB, and we're only having less than 40k people hanging around.
It all depends on what Steemit INC considers relevant to implement at this moment. Alternatively, a group of devs could fork the codebase, start developing this and then submit a PR. Or just keep playing with the new codebase, by starting a new blockchain.
It's good to know that there is a mechanism in place to protect people from that.
Is a full fork the only way to sufficiently manage changes that aren't directly requested by steemit inc?