Byteball Know-it-all bot weekly update (4)
Solving a real world problem : Participation and Gratitude
A test to found out whether incentivize Q&A as a concept can bring participation to a whole new level.
Simple frictionless flow that works now
- Question is broadcast to random selection of users
- Incentive can be attached to the question as a bounty by anyone
- These users can send in answers and/or vote for an answer
- When half the votes are in or time is up, users who voted for best answer split the bounty
Still debugging the bot a bit because of variable payment on the DAG, but think this is the core of it. Luckily the contest allows buggy / crashing code, it is here.
This bot could even be a good way to distribute bytes, it gets harder to game the reward as users number increase, because not only are voters selected randomly without any bias like balance size or history, but voters need to do work answering and voting within a racing deadline. Best of all voters are actual participants on the network, like a Captcha that gives useful answer.
Aggregated data from questions and answers could provide insight on where users have frequent questions, like Stackoverflow.
Adding to the Byteball ecosystem
My hope is that more conversation related to Byteball will move to my bot, so we don't have to spread our attention on 2 telegram groups, multiple slack channels and etc. Basically to consolidate short form communication to the bot, and so contributors taking questions can also get reward.
Since this bot deals with micro-transactions, the more transactions this bot can do, the faster the confirmation time for the entire network too.
Best of all, it kinda fun to use, changing votes and guessing which answer is going to win.
Costs for the operator
I was about to run this bot on an iCore 2 machine with 4Gb RAM and 40Gb SSD, each question only takes about 1K to store depending on the length of the question, and another 1k block to store user states. Since I am storing everything as files using node-persist, I have a few advantages which I have only seen with my bot has so far:
- since data is store as individual files, I can easily archive old questions using file tools and not worry about dealing with database
- since data is stored in a key / value store, in the future I think I should be able to migrate the data to the DAG to achieve better decentralization
- A separate bounty asset from bytes in order to have network value extraction rate control, similar to steem / steem dollar system
- Promote command which increase range of random selection
- Store user states in byteball DAG for more decentralization
- Aggregated data analysis available for purchase and then redistribute as general bounty