I know it's trite on HN, but: isn't this literally just version control? Particular integrity choices aside, Git has all of the properties you state above.
Is it distributed, or are you hosting the whole instance yourself?
Does it use any other academic component of the bitcoin suite besides linked logs? https://queue.acm.org/detail.cfm?id=3136559 I ask because a hashed merkle tree probably isnt what comes to everyone's mind when they hear blockchain, even if it is the literal definition. The author of the article is using it to be a distributed ledger between untrusted parties. If your product isnt that, its sort of an apples to oranges comparison to say you have a use case for something different than what he is dismissing.
The entire tech is open source and can be used on-premise, in the cloud, or in any kind of hybrid. Today, we are offering Nebula which is an integrated engine and viewer with no outside dependencies, but everything is there for anyone to deploy it on a server and separate the parts. It's designed for that.
Coming very shortly are MSL Notebook and MSL.js which are UI components and a Typescript library for working with MSL and adding it to your own apps, wherever they live. We also have a "headless" MSL engine that lives in the cloud which will soon be public so you can play with it through MSL Notebook or use it to test and develop your own apps.
We are not using Bitcoin or any component of it. Our idea is based on the old skool meaning of "blockchain" in that it is a chain of records whose order can be proved by hashing. That is the only component of the blockchain concept that applies here. We use it to hash individual edits within a document, as opposed to hashing entire documents.
In a simple scenario, every edit you make to a document is recorded individually and hashed with the hash of the previous edit. This forms the chain. The chain itself can be stored anywhere. It's useless if you don't have the document. But if you do have the document or any fragment, you can instantly tell what version you have and if it has any differences from the original -- without seeing the original.
What is the use case for this system? How will this actually benefit me? What problem is this solving?
The TL;DR is documents are wrong and often out of date. People rely on incorrect information and that screws up the work and the workflow. We are fixing this with a language designed to track and correct text as facts change.
Here's our new pitch deck with the 10,000 ft overview: http://invest.mimix.io
In a simple scenario, every edit you make to a document is recorded individually and hashed with the hash of the previous edit. This forms the chain. The chain itself can be stored anywhere. It's useless if you don't have the document. But if you do have the document or any fragment, you can instantly tell what version you have and if it has any differences from the original -- without seeing the original.
You've just described a degenerate Merkle tree.
> The chain itself can be stored anywhere.
Then it's not a blockchain. A blockchain involves combining a Merkle tree with a distributed, trustless consensus algorithm.
Otherwise Git would be described as a blockchain and I don't think anyone refers to it that way. At least, not with a straight face.
In this scenario, however, we're going for the Wikipedia or original meaning of blockchain:
"A blockchain is a growing list of records, called blocks, that are linked together using cryptography."
The only use we have for this technology is the growing list of records linked by cryptographic hashes. We are not interested in consensus algorithms and they don't form part of our system.
The blockchain aspect of our product isn't any pixie dust. It isn't even necessary. Think of it like power windows. We don't sell the car because it has power windows.
Many users of our software (the powerful, pixie dust part being our MSL language) won't give a whit about blockchain and we don't sell or emphasize it on our website.
So I see where you're coming from, but in this case, we're coming from a little bit different place.
If you read past the summary to the actual meat of the article you'll find that:
> A blockchain is a decentralized, distributed, and oftentimes public, digital ledger consisting of records called blocks that is used to record transactions across many computers so that any involved block cannot be altered retroactively, without the alteration of all subsequent blocks.
You've just admitted your solution is neither decentralized nor distributed. By no common definition does it meet the description of a "blockchain" and I genuinely wonder why you persist in using the term, as it's at best misleading.
It sure appears to be in service of pure marketing. Which, hey, marketing is necessary! I get that.
But, in the context of this conversation, citing your project as proof that blockchains have practical applications when it doesn't appear to use a blockchain in any meaningful sense strikes me as pretty disingenuous.
Caveat emptor, for sure. Like I said, the proof is always in the product itself much more than in anything I say.
The Wikipedia article goes on to (incorrectly) attribute the blockchain to Nakamoto when the idea was well known in comp sci before Bitcoin showed up.
Our use is, indeed, both decentralized and distributed. I mentioned earlier that the record of hashes is often kept offsite and can be used by a non-trusted third party (who doesn't have access to the documents) to perform the kind of document verification I'm talking about.
Apparently, misquoting Twain is popular itself[0]. I only mentioned it in humor so I hope this wasn't taken as a slam in any way. Twain is an ancestor of mine and was famous for pithiness.
[0] https://www.mentalfloss.com/article/562400/reports-mark-twai...
What does this even mean? Are you Disrupting™ the Text Editing Industry? In what substantial way is it different - and better - than existing version control?
These are just a few of the aspects of the language and the tools we've built to implement it. If anyone is interested in this stuff, there's a ton at our website! http://nebula.mimix.io And I'm always open for questions and feedback.
I believe you're wrong, git commits are diffs, not snapshots.
(I don't necessarily agree with the article's messaging around "Haha you were only confused by git because you thought commits are diffs now that you know they're actually snapshots you have no more confusion", because, as the article itself admits, most of the commands like cherry-pick do work on dynamically-computed diffs anyway.)
Whether it's represeented as a diff internally seems completely irrelevant to me, even to understand the examples in the article. I would like to be able to "de-serialize" a real commit's data to see exactly what's stored, but this article doesn't seem to show how to do that... do you have a suggestion on how to do that?