I read your write-up on EOSIOS. Thanks. It was very interesting…but.
Below are the 4 things that make clear there is little use for the technology. My points on these I put in brackets [ ]:
1. Scalable aBFT consensus — makes every chain UX friendly and industry ready
- 0.5s block time
- Approximately 10,000–20,000 transactions per second capacity depending on the type of usage*
** A transaction does not deduct money from a user’s account every transaction.
[The qualification of the TPS almost certainly means it's largely useless for 99% of TPS-sensitive applications. Block time and its application and time constraints will never be overcome. 0.05s is miserable and the non-deterministic nature of "transaction" being decoupled from block time assures this. I take it you are familiar with Bloxroute. They try to overcome the problem by effectively creating a private Level 0 infrastructure and charge the miners a fee. This is supposed to ultimately speed up the block rates, which it does, but think about this: already POW is a monopoly mostly centered in China and, from a rig standpoint, pretty much entirely centered in China. Add Bloxroute as another basically centralizing component and you have a de facto undermining (no pun intended) of the beloved feature of decentralization. Not to mention the requirement that all nodes (applications) must use a Bloxroute sidecar... free for now, among other potential non-technical flaws.]
2. Consensus governance agnostic (see below)
- Up to 125 block producers (BPs)
3. Delegated proof of stake governance where each token can vote for 30 block produces, the top 21 by vote count are then authorized to operate the blockchain.
[2 and 3 belong together. This is not dissimilar to the algo I worked on, in that delegated proof-of-stake-duration was used, instead of simple delegated POS. But still, the complex of problems with such a scheme is significant. I can go into depth somewhere else, but ultimately the complexity of the algorithm creates an enormous space for either implementation error or unexpected non-deterministic behavior. AKA software perfection]
4. Flexible resource accounting — allowing for free** transactions and for resources to be paid for by smart contracts instead of its user.
** A transaction does not deduct money from a user’s account every transaction.
[I get the "necessity" of this. But the simple fact is you have a chicken and egg problem. Just as all decentralized platforms have: Getting to a point where the platform can function as it should requires ultimately money. Unless you sponsor the first 20? 50? 100? nodes, there is no incentive to run one. You have to trust and believe like it was your religion, otherwise, and spend the money, time and effort to spin one up. The number of otherwise seemingly technically impressive projects that have gone nowhere or entirely disappeared is evidence of this.]
I point these out simply because they are fundamental flaws/shortcomings. All the other features that you list as advantageous contribute nothing to the fundamental viability of the model. To me, they are a form of plugging holes in a sinking ship.
The very first and general observations about your response is that you have not responded to what is central to my view of “blockchain” and decentralized distributed ledgers, DDLT:
What problem has been solved and what problem will be solved by this technology? My answer so far is nothing and nothng.
As a techie, I very much understand the attraction of the technology and “solving” its many shortcomings. I participated and enjoyed (initially) the challenge. And I was also financially very attracted. But I cannot identify any application that will in any signficant way benefit. And, in my own efforts to rally applications to use a platform and the effort to make that happen from a technical and business perspective made clear this was just really pointless and vanity. If for no other reason than that the financial underpinning of all these platforms is either mining and/or the cost/benefit analysis of running a node.
My prediction is that DDLT will subside forever into the shadows. It won’t go away, but will eventually accept that its main use cases will be either furthering control of (permissioned chains) corrupt autocratic states or supporting platforms that allow discovery-free illicit activity (see BTC, ETC, USDT, etc.) and the dark web eCommerce a la Silk Road.
