Irakli Gozalishvili

So here is rough proposal what we could do to address this problem (I'm making a lot of assumptions here, so please call out when those are incorrect). - Canonical...

> `{“/“: stringCID}` is actually taken from that old canonical JSON representation, so to the extent that there is a generic solution, that’s it ;) Are you suggesting that if...

> ’m pretty sure this will break a lot of stuff. I’d need some time to work through all the consequences of isCID accepting this form but my initial reaction...

> Frankly, this is a big step backwards from IPLD’s point of view. Link being a top level type in the Data Model unlocked a lot of stuff for us...

> I would assume that these worker processes aren’t just responding to every random message anyone decides to post. The API will actually be something along the lines of It...

So here is a more concrete example based on input I got here and discussion with @mikeal on a different sync channel. I understand that having a canonical top level...

I would like to also note that this is not just about passing things around worker threads. It's bit more general web platform thing, e.g storing things in IndexedDB has...

One thing that came out in conversation with @mikeal > because it means that you can’t represent data that wanted to use `{'/':someOtherMeaninfulString}` **Acyclic** nature of DAG could be exploited...

> @Gozala can you explain this last comment a bit more? I don't understand how this is relevant to the discussion here. So as I understand main concern with not...

Few note here: 1. `CID.isCID` is flawed pattern, which I've attempted to outline and propose alternative in #111, which may also address the issue I have being having here. 2....