Spike: Support Marlowe's spike

[ch1bo]

trafficstars

Why

• streamline, simplify, enable open-source contributions
• that for, a potential IO internal user: Marlowe will be a guest and evaluate Hydra as a potential puzzle piece in running Marlowe contracts more efficiently
• we want them to have experience and also ensure discovered things are to the benefit of future contributors

What

• As a first step the Marlowe will

  • explore the repository
  • run the demo
  • hackily try to run a marlowe script inside a hydra head (by extending the demo case or so)

• That for, the Hydra team will commit to respond on any questions on StackExchange, Discord or Slack within a 12 hour time frame.

  • Where it makes sense, StackExchange or Discord is prioritized for the benefit of the greater community (esp. SE could become a knowledge base)

• Also, we will ensure that this first Marlowe spike produces a

  • solution design for one Marlowe use case on Hydra
  • any existing features needed for that use case are identifed
  • any missing features are tracked as idea items on our roadmap

How

• [x] #404
• [ ] Reflect on what documents were helpful and which were missing during this interaction

[paluh]

Hi @ch1bo,

thanks a lot for the nice intro and for the above clear commitment ;-)

We are starting to scaffold something here: https://github.com/input-output-hk/marlowe-hydra-poc

This week we are going to migrate parts of this ticket to our project. The rest of the colab could be done using usual issue / PR flows and other communication channels as you suggested above. Should we close this ticket then?

[ch1bo]

Hi @paluh

great to see you getting started! I think we should keep this ticket open for the first of your spikes and do the "reflect what was helpful and/or missing" part on our end. I expect that you guys will be time boxing the exploration and hacking spike on your end? That would allow for a definition of done on this one and not have it be the never-ending story ;)

[ch1bo]

After @paluh and @yveshauser had the opportunity working with Hydra from Marlowe, they have shared what resources they found helpful, outdated or missing. I am listing them here non-prioritized and with some ideas on follow-up actions:

helpful

• The website is great and really helpful to do the first steps and run the demo application etc.
• Clear codebase with clear namespaces and sub packages (choices like Hydra.Cardano.Api and not Hydra.Utils.Cardano without clear substructure).
• Architecture decision records and how the decisions where taken (e.g. no effect systems) - keeping it simple. Would expose the ADRs even more. -> Make them a top-level navigation entry on the website?
• Examples / the TUI application

outdated

• Project seems to be in good shape, we could think about dropping the -poc suffix -> Let's do it

missing

• In order to understand and implement our demo application we mostly explored source code - hoogle and haddock would be preferable, but are missing -> Provide instructions how to setup repository using haskell.nix and withHoogle = true?

• Differences differences between the paper and the current implementation. -> Yes! #448

• Were not sure how "hydra channel isomorphism" handles policy scripts and minting (it is not apparent from the paper). -> Mentioned a bit in the FAQ and one feature; should write about it in the docs though!

• Hydra client in PureScript/TypeScript -> Sounds like a perfect future open-source contribution!