Yes, of course, there are bigger issues, but technical performance and scalability is a significant one. It's not the lagginess that's the bigger concern, it's the scalability. Right now, to the end user, it's fine - there are only a few thousand active users, and you only need to have a few SP to be active. But imagine if it goes up to hundreds of millions levels of activity - it'll cost a pretty penny just to make votes and comments. Not to mention, even after MIRA and other improvements, and despite very limited activity only 3.5 years into this, it still takes an entire week to replay a bare minimum Steem blockchain node. Who knows how many weeks it'll take for a full node. You can see how scalability is a very real challenge that makes social networks ill-suited to blockchain.
You are viewing a single comment's thread from:
There are protocols that make replaying the entirety of the blockchain unnecessary while still enabling confirmation of the data, gossip about gossip is one, and recently another method of hashing has arisen. By using such protocols, very low latency nodes are potential, and greatly enable scalability.
I'm not a dev, so my grasp of the cutting edge may not be relied on. I do note that technical advance eventuates, and expect it in this field as it ubiquitously impacts all. These are not the limiting factor anyway, as there's no need to scale up unless gaining users actually happens.
As long as automobiles were hand cranked failure prone devices limited to <60 mph speeds there was no need for highways. Until the social platform potentiates demand, scaling will not be demanded. Regardless of chickens and eggs, both problems require solution for Steem to survive.
I'm sure there are ideas, but it takes years to develop. Steemit Inc spent most of 2017 and 2018 focusing on scalability, which is why there were no major updates in the last 2 years. A hardfork with major protocol updates used to be every 3 months before, then it stopped while they focused on a broad range of scalability solutions - RocksDB, RC, MIRA etc. Scalability always comes before adoption, this is absolutely not a "chicken & egg" problem.
And to what end? Reddit, Twitter, Medium etc can already handle a million times greater activity. Why even bother? I wrote about this before: https://steemit.com/blockchain/@liberosist/why-blockchain-social
Not when configured as recommended for a minimum (witness/consensus) node. It takes about 5-8 hours on a low end system except with 64 GB of RAM (a memory requirement that is designed to increase relatively slowly even under heavier use).
MIRA allows running in less RAM but with much worse replay time, an optional tradeoff.
Full nodes are slower but still only a day or so with good hardware. Again, tradeoffs between cost and time exist but are optional.
So far it is still workable, and could still scale somewhat while remaining workable, while the developers still plod along at making improvements. How their pace of progress will go from here on out I don't know.