Oct 2022 Community Call

[expede]

Sign up: https://lu.ma/ucan

Suggested Agenda

In Attendance

Video

Chat Log

00:03:10	Philipp Krüger:	https://github.com/ucan-wg/spec/discussions/114
00:03:41	Chris "cdata" Joel @ Subconscious:	Hi, sorry for the late addition, but I added an agenda item for discussion on Github
00:04:03	Dmitri Zagidulin:	yeyyyy thank you for moving the call dates!
00:04:03	Philipp Krüger:	Well, this came up this morning, no problem!
00:10:45	Philipp Krüger:	best explanation of identity hashes in CIDs I could find: https://github.com/ipfs/kubo/issues/4697
00:19:35	Mikeal Rogers:	i think it’s prudent to swiftly ship something we’re comfortable calling 1.0 at which we are committed to backwards compat, and then ship 1.1 shorty after to set to expectation for what forward compatible versions will look like
00:20:15	Mikeal Rogers:	my typing is terrible cause my daughter is home sick and runny around me as i type
00:28:51	Mikeal Rogers:	btw, if there are questions about identify CID’s i can answer them :)
00:29:25	Philipp Krüger:	No Qs from my side, just wanted to give people not familiar with them a resource. But of course they can also ask in chat :)
00:30:48	Philipp Krüger:	FWIW I don't even think we need to make use of identity hash CIDs for compatibility between entities
00:33:55	James Walker (@walkah):	🦀
00:35:09	Chris "cdata" Joel @ Subconscious:	BTW: Ancient UCANs are kind of nice for uplifting to 0.9 because you can always just parse the token with a compat library and generate a new token (or tokens)
00:35:11	Mikeal Rogers:	ya, it’s common for people to think they need to support identity CID’s in places they actually don’t ;) so i often just show people that there’s less work than they think
00:48:44	Brooklyn Zelenka (@expede):	and/or Wasm
00:49:15	Chris "cdata" Joel @ Subconscious:	Do we need to get go-ucan up to speed?
00:52:18	Chris "cdata" Joel @ Subconscious:	UCAN interop hack on-site? :D
00:55:28	Chris "cdata" Joel @ Subconscious:	Thanks for the thoughtful compatibility discussion 🙏

Notes

• How far back should we support? v0.8 or v0.9
• General consensus that we should start to intentionally move towards a v1.0 release sometime early 2023.
• It is time for ucan-cacao!

[cdata]

@expede I would like to propose an agenda item for discussion:

UCAN compatibility across spec version and language implementations: Are you using rs-ucan today? Would you like to? If it maintained some degree of backwards compatibility with old versions of the UCAN spec, would that enable you to start using it, or make it easier to keep using it?

Spec authors: how and to what extent do you take forward migration into account when thinking about breaking changes to the spec?