With both @neatonk and @sweetmantech I had sessions around explaining how the neume strategies work and it was genuinely difficult even for me to explain how this works. Here are some of my observations
Users don't get why a strategy has to return {write:..., messages} and what those properties are for
Users don't know that write writes to the disk eventually
Users don't know the order in which function init and function update and executed and it may not make sense that e.g. function init is executed only once but then function update is executed continuously
If users aren't already familiar with extraction worker and the general concept of neume, intuitively they want to sequentially run a for-loop that just fetches all results within a strategy
But when I explain that the extraction-worker has a lot of useful benefits, they're OK with using it.
Sometimes we also have cases where it'd be nice to just do a synchronous blocking call like e.g. getting the totalSupply of an NFT contract, but then this has to be built with this weird mixture of first returning that task in function init and then picking it up in update and that's annoying. Maybe we should allow the user to just make such calls
{write:..., messages}and what those properties are forwritewrites to the disk eventuallyfunction initandfunction updateand executed and it may not make sense that e.g.function initis executed only once but thenfunction updateis executed continuouslytotalSupplyof an NFT contract, but then this has to be built with this weird mixture of first returning that task infunction initand then picking it up in update and that's annoying. Maybe we should allow the user to just make such calls