Basic token implementation

[robinmonjo]

Token system implementation. This PR includes:

• new Token app that uses the existing blockchain app
• fully working token proof-of-concept
• new READMEs

This is mostly complete, even if some tests are missing. TODOs before merging:

• change repository name to coinstack
• ensure Travis still run on new repository name
• read over READMEs for typos and english mistakes

[yordis]

Those three

 # https://en.bitcoin.it/wiki/Proof_of_work
  def perform_proof_of_work(%Block{} = b) do
    {hash, nounce} = proof_of_work(b)
    %{b | hash: hash, nounce: nounce}
  end

  defp proof_of_work(%Block{} = block, nounce \\ 0) do
    b = %{block | nounce: nounce}
    hash = compute_hash(b)
    case verify_proof_of_work(hash) do
      true -> {hash, nounce}
      _ -> proof_of_work(block, nounce + 1)
    end
  end

  def verify_proof_of_work(hash) do
    difficulty = Application.fetch_env!(:blockchain, :pow_difficulty)
    prefix = Enum.reduce 1..difficulty, "", fn(_, acc) -> "0#{acc}" end
    String.starts_with?(hash, prefix)
  end

move it out to some module call ProofOrWork or whatever will be and that module implements some protocol probably

[yordis]

If that pow_difficulty is actually related to ProofOfWork which I think it is no?!, then do

config :blockchain, Blockchain.ProofOfWork,
  pow_difficulty: 10_000

Application.fetch_env!(:blockchain, Blockchain.ProofOfWork, :pow_difficulty)

Dont over use the :blockchain scope, this way you will help people to understand what is going on with that config.

[yordis]

Wouldn't be better to make this a config that you could swap? If not then this implementation force you to use PoW which it is fine but in my head is probably better to create some config, even we can plug more than one computation (but that's too fancy for now)

[robinmonjo]

@yordis found some time tonight, what I did:

• moved PoW logic in its own module as you suggested
• make PoW module pluggable from the config and created Blockchain.Test.ProofOfWork for testing so tests run faster
• modified the config as you suggested for pow_difficulty
• added some tests on the Transaction.Verify module

I haven't scoped everything under the Coinstack domain. We don't need it for now.

I'm pretty happy with the state of this PR, you made really interesting feedbacks ! I think I'm gonna go over the READMEs one more times and merge it. What do you think ?

Once merged, if you go over the code and find things that could be improved open an issue, I would be happy to handle it !

[yordis]

Don't do this

if you comment out this line

#     import_config "#{Mix.env}.exs"

then you need to create the respected file for the environment. Because it will load that file to

What you could do is to put the default configuration on config/config.exs and then use config/test.exs for override the values.

Side note, the override is using a deep overwriting

[robinmonjo]

ok took care of that

[yordis]

I have question because you are 100% more involve than me.

What would happen if we want to support something like PoW + PoT together, IF I am not mistaking Dash coin does that, it split it up into 50% 50% work.

Also, let say tomorrow I want to create ProofOfSharing (PoSH) how would that look like?

When I told you to move it into some module I was actually thinking on create some config with the module (not called proof_of_work because defeat the idea 😄 ) where we can setup either a list or just a module and you will call the compute function on it, which is why I told you about the protocol.

So that way you can have as many implementations of how you reward or do any computation of this kind.

Something like

config :blockchain,
  some_better_name_key: [ProofOfWork, ProofOfStake, ProofOfShares]

# or

config :blockchain,
  some_better_name_key: [
    pow: ProofOfWork, # <- for the sack of the demo whatever key value pair
    pos: ProofOfStake,
    posh: ProofOfShares
  ]

the last example could be useful when you want to control that you HAVE to use specific module for X reason. I would even stick with the array for now 😄 we need study what other coins do differently and more important where the crypto world is moving to

[robinmonjo]

I understand what's your point here. However for now I only know proof-of-work, I would have to dig more into other protocols that replace PoW to have a better idea on how this could be implemented and how we could go to what you suggest here. If that's ok with you, I'll move your really good comment to an issue and spend times on it later

[robinmonjo]

issue here: #8

[yordis]

|> Mining.mine()?

[robinmonjo]

indeed bad, name here, I changed it to MiningPool.add

[yordis]

if it's working then go and merge to master so we do not keep looking at big commits to review 😄 I am more than happy to keep giving you feedback. Sometimes just asking you questions or debating about the implementation we both learn.

So far so good 💟

P.S: @robinmonjo did you read my latest comments? Because I think I didn't hit the comment button and I am not sure now. You have to probably expand the comments

[robinmonjo]

@yordis, again, thank you for your review !! I'm doing Elixir on my own and that's really interesting to have feedbacks from a more experienced Elixir dev. I think we can merge this PR, but please, keep "hitting" me with feedbacks that's the best way for me to improve 😊.

Will go over everything one last time tomorrow and merge.