from Hacker News

Do you need a blockchain?

by kkleindev on 2/6/18, 10:43 AM with 183 comments

  • by jmull on 2/6/18, 2:45 PM

    It seems to me the key here is "Can you use an always online trusted third party?"

    I think generally, a TTP is going to be more efficient than a system without a TTP. "Trust" at its essence is a shortcut. Without it you have to create a general enforcement mechanism which is going to cost something. E.g., the cost of incenting miners. You can slice it difference ways, but in the end you need the distributed network of nodes to exist and the people who control them to be invested in the continued existence and integrity of the system.

    I say "generally" for a reason, because I think it's certainly possible that there are cases where the cost of the third-party is too high or the third-party may not be reliable enough, etc. But generally, it seems to me that blockchain is an inherently inefficient mechanism.

    On top of that, the only effective use so far has been to create an investment bubble. Once enough money has been lost on that I think there's a good chance blockchain will get a black eye which will cause people to steer clear even if they may have a use case that's a good fit.

  • by chvid on 2/6/18, 1:07 PM

    This is a reasonable paper discussing an obivous question given just how much hype is surrounding this technology.

    To me "true blockchain technology" in the bitcoin sense of the design is open, permission-less and only has a few real applications. The others: Private, permissioned are really just variations of a central database which is replicated multiple places.

  • by seanalltogether on 2/6/18, 11:44 AM

    Do you need cryptographically signed, P2P distributed SQL statements that define order by referencing a hash of the previous statement block.

  • by ForHackernews on 2/6/18, 2:36 PM

  • by ghosttie on 2/6/18, 5:56 PM

    Honestly expected a page that just said "No"

  • by tabeth on 2/6/18, 2:09 PM

    This paper is similar to an idea I'm working on finalizing from a conceptual point of view.

    I call of "Proof of Help" (as opposed to Work or Stake). To summarize, the idea is that in a market for tutoring, people can tutor each other, however there would be some sort of way to confirm that tutoring actually took place.

    From there, you may receive help in exchange for the quantity of time you've helped others, represented by the Help that you have accumulated by helping. Unlike Bitcoin, this Help can never be converted for fiat and is only useful for receiving tutoring.

    Unlike Bitcoin, this wouldn't be decentralized though. It would effectively be a public ledger run by ideally a nonprofit. I think blockchain is a good use for this. Mainly because the quantities of Help should not be able to be manipulated by the centralized authority. Their role here is simply to confirm identities and serve as a sort of access control to the blockchain itself (e.g. They can ban people, but not change how much Help someone has).

  • by philipodonnell on 2/6/18, 1:51 PM

    One question I would like to see discussion on is about the process of capturing value by the builder of a blockchain.

    Web 1.0 was about selling stuff. Web 2.0 was about hoarding data and using it to sell stuff. If blockchains are Web 3.0, what's the model?

    Bitcoin was released free and the designer got nothing beyond the initial stake, which only works for coins. Ethereum is run by a foundation.

    Say I want to launch a blockchain startup and really hit the sweet spot according to this paper, how do I build a business on it?

  • by endorphone on 2/6/18, 1:01 PM

    The definition of "trusted" is the foundation of the conclusions of this article (and of the snarky "no" pages that are so popular), and it really deserves a long and detailed exposition of its own.

    You may trust those third parties, but as always you also want the facility to easily verify (because your trust can be undermined by a single employee, a single malfunction, a single hack, and so on). When you add the notion of verification the whole evaluation flips.

  • by rodolphoarruda on 2/6/18, 11:40 AM

    Any justification or argumentation below the "NO" would be very welcome. I did download the article though, if that was the takeaway behind the website idea.

  • by aiCeivi9 on 2/6/18, 11:41 AM

    Lack one question: "Is your software/infrastructure in dire need of refactoring/upgrade but management isn't willing to pay for it"

  • by xHopen on 2/6/18, 1:25 PM

  • by adultSwim on 2/8/18, 8:49 PM

    No

  • by Fletch137 on 2/6/18, 11:42 AM

    I might be a bit grumpy, but I'm getting really sick of these type of pages. There's no information, just a smart-arse "NO". It doesn't help anyone and it's so overdone I don't think it's funny anymore.

  • by rb808 on 2/6/18, 1:07 PM

    I prefer this one.

    How to decide if blockchain is right for your project?

    https://twitter.com/MalwareTechBlog/status/93264913325659750...

  • by _wmd on 2/6/18, 11:54 AM

    mods: I think it is a deeply embarrassing indicator of the state of our industry that a balanced, well-referenced, academically framed discussion of the tradeoffs of a buzz-laden technology should be allowed to be flagged like this.

    At the least, replace the site with a link to the PDF. It seems the average user here is incapable of reading as far as the second link on the page.

    > The choice between a permissionless, permissioned or centralized database is not trivial. While this question has been discussed before [15], to the best of our knowledge, we provide in this article the first structured methodology to decide which technological solution is the most appropriate depending on which application scenario

  • by emirozer on 2/6/18, 11:40 AM

    Sorry but this is a bit dumb. If you say YES to this question: "Can you use an always online trusted third party?" Regardless of your other answers, it always says NO as a result.

    Can someone give me an example of an always online trusted third party, and also let me know that how can you see into the future so that this third party keeps its guarantees throughout life?

    Edit: flagging this post as it seems like a troll attempt to me

  • by philfrasty on 2/6/18, 11:41 AM

    I get the joke „do you really need a blockchain?“ and sure there is a lot of „AirBnB on a Blockchain“ types out there that do not offer a real advantage over a centralized solution.

    But why not embrace all the money (and talent) flowing into this space through the hype and let natural selection run its course to leave us with some great tech that really leverages blockchain technology?

  • by tlrobinson on 2/6/18, 11:37 AM

    A less condescending version of this that actually explained why you do or don't need a blockchain would be nice.

  • by NoGravitas on 2/6/18, 1:23 PM

    Obviously, Betteridge's law of headlines applies here.

  • by nukeop on 2/6/18, 12:50 PM

    In other words: do you want to see your stock triple in a week?

  • by tanilama on 2/6/18, 12:58 PM

    True

  • by mjfl on 2/6/18, 12:32 PM

    do you need online payments? Blockchain is nice.

  • by elmar on 2/6/18, 11:40 AM

    The system you are building needs to be decentralized and censorship-resistant? yes, then you need a blockchain.

  • by onion-soup on 2/6/18, 12:58 PM

    Do you need an internet? Ok, let's have a look at this diagram:

    Do you want to have social life? ---- Yes: you don't need internet ---- No: you need permissionless internet

    Clearly, you don't need an internet. It lacks coherence and autonomous smart contract decentralization cloud computing chainblocks. Case closed.