I believe it does. Let's take it point by point.

Shared data store: yes. One of the biggest problems with Bittorrent is that you can only use it if you already know the infohash of the content you want. There's no index of the available content. There are lots of third-party search engines that try to monitor the network and help you find stuff. They are all incomplete and using them is a bad experience.

Does more than one entity need to contribute: yes. Anyone and everyone can publish to LBRY or pay for LBRY content. If only one entity had write access, that would be a point of censorship.

Data is never deleted: yes. Once you publish something to a decentralized network, its out there. You can't put the toothpaste back in the tube. We actually do have a mechanism to "update" the metadata (description of the content that's stored in the blockchain) by connecting a diff with updated data to the original blockchain transaction.

Sensitive information will not be written: yes.

No single trusted entity: yes. That's how we can make a strong promises about censorship resistance and you don't have to trust that we won't change our mind later.

Tamperproof log of all writes: I'm not sure why this one is strictly necessary, and why we should not use a blockchain if we don't care about a tamperproof log. Please help me understand.

Blockchain requires a tremendous amount of electricity, contributed by all those participating in the cryptographic portion of the network. If you're not using blockchain to create a tamper-proof log, you shouldn't be wasting the planet's resources on a blockchain project.

Even if you can roll back the requirement to "tamper-resistant" you can get away from the necessity of blockchain: use a hash chain instead.

Since you don't care about tampering, I request that you avoid blockchain.

Check out this simpler one (even hosted on LBRY!!): https://spee.ch/doyouneedablockchain lbry://doyouneedablockchain

I have a simpler one https://cdn-images-1.medium.com/max/1600/1*5aEzSYyVZ9pJk5Y5m...

And now it's on spee.ch! https://spee.ch/4/blockchain-flow

Oh nice, that's simpler to follow! But the other one is helpful in suggesting the specific alternative.

No. It fails at the second step because ownership is generally not contested and a single trustworthy organization can be responsible for all writes. In the event that a blockchain based project is financially successful, you should expect that a more efficient, single trustworthy competitor will arrive with lower costs.

What single company do you suggest can be trusted to maintain a listing of digital content that anyone can contribute to without censoring or interfering with the listing in some way?

We already know we cannot trust Google, Facebook, Apple, or Amazon.

And even if there is a company that we could trust, what happens when a censorious government wants something removed?

IP licensing can't be enforced outside of governments... Either you have enforcement and possible censorship, or you have no enforcement. The stated goal was to provide some form of payment system for content, but in a truly distributed system any valuable content will be paid for once and then reposted without payment required.

The ability to repost content without payment required is already true of every existing digital media distribution system. We do not claim to improve this situation.

LBRY does not provide IP enforcement. The primary reasons people pay for content when they could get all content for free are: 1) their personal morality and 2) fear of legal repercussions

The same will be true on LBRY.

We have Wikipedia, internet archive, library system that could be counted here as well no?

I'm fairly certain that Wikipedia does not allow anyone to upload a video or image and to my knowledge (though I'm less certain) the Internet Archive does not either. My second point would apply to both of those entities, regardless.

This is a good chart. At a first glance - yes, their use case does pass this test.

Could you summarize why?

I see two failure points - I don't think you need a tamperproof log of all the writes to the datastore, and you might want to update records for where things are located over time.

Could you explain why a blockchain should not be use if a tamperproof log is not necessary? If the log is just a byproduct but the rest of the properties are necessary, is that still a valid use case?