Zoom Logo

Secure Data Storage - WG - Shared screen with speaker view
Juan Caballero
02:36
https://github.com/decentralized-identity/edv-spec/issues
Juan Caballero
06:25
Manu was valiantly screensharing and issue-wrangling all of the hour before this
Manu Sporny
08:20
q+
Dave Longley
08:55
q+
Andreas Freund
09:23
q+
Dave Longley
11:25
i'd rather keep things separate by default but i'm not necessarily opposed
Derek Trider
11:37
q+
Dave Longley
12:12
it may end up being fine -- but it may also create some undesirable constraints/implementation challenges on implementations
Manu Sporny
12:38
q+
Dave Longley
12:48
i think it's `/documents` btw -- (not `/docs`, could be wrong)
Dmitri Zagidulin
13:06
oh, that’s right. there’s a separate issue for /documents vs /docs
Dave Longley
13:37
q+
Derek Trider
13:44
/docs or /docs/batch?
Derek Trider
14:05
*oops I mean /batch or /docs/batch
Derek Trider
15:14
These are good points - seems like separate endpoint is the way to go!
Manu Sporny
15:21
Yeah, +1 to /docs/batch <-- I feel much better about that.
Dave Longley
15:23
`/edvs/<edvId/batch` would give the most flexibility
Derek Trider
15:30
Sounds good to me!
Dave Longley
15:46
q+
Dmitri Zagidulin
17:12
+1, my vote is for `/edvs/<id>/batch` also
Dave Longley
17:28
i still don't know what you can send to the endpoint
Derek Trider
17:39
q+ to answer Dave's question
Dave Longley
17:46
if you send a doc that hasn't been created before does it accept that doc?
Dave Longley
18:58
given that we do updates via tombstoning, you could batch delete as well, i'd think
Dave Longley
19:49
+1 ... don't like the inconsistency
Dave Longley
19:50
q+
Manu Sporny
21:36
q+ merge PR, open issue
Dave Longley
22:11
+1 to being explicit
Dmitri Zagidulin
22:14
+1, agreed, we want to be explicit with the operation
Dave Longley
22:17
about the operation
Derek Trider
22:38
So I will
Derek Trider
23:08
*So I will 1. change endpoint to edvs/vaultID/batch and 2. make an issue and point to it regarding the operation syntax
Derek Trider
23:44
Sounds good. Thanks everyone! I'll update the PR soon.
Dmitri Zagidulin
24:00
https://github.com/decentralized-identity/edv-spec/issues/20
Dave Longley
24:16
q+
Dmitri Zagidulin
25:09
q+
Dave Longley
27:41
that's still "write" permission :) ... but ok.
Manu Sporny
28:48
q+
Derek Trider
30:05
q+
Dave Longley
30:55
mostly, an editor takes the consensus positions of the group in issues and creates PRs for them
Dave Longley
31:07
and reviews/merges PRs from others
Dave Longley
31:17
(once consensus from the group is achieved)
Dave Longley
31:45
does anyone disagree with anything i said? :)
Manu Sporny
32:13
q+ can tombstones be revived?
Dave Longley
32:32
short answer: you'd just do another write with the next sequence number using the previous content
Manu Sporny
32:50
no fun answering in chat, man. :P
Dave Longley
32:57
i'll q+ to answer you
Dave Longley
33:04
q+
Juan Caballero
33:13
you can sprinkle holy water on them and baptize them
Michael Shea
33:18
resurrection?
Manu Sporny
33:24
can you batch resurrections?
Manu Sporny
33:37
q+
Dave Longley
34:01
q+
Dmitri Zagidulin
35:55
q+ to ask about concrete tombstone syntax
Dave Longley
36:07
q+
Manu Sporny
36:45
re: EDV baptism... born again file systems? Are EDVs unbaptized if they're tombstoned? :P
Dmitri Zagidulin
38:01
https://github.com/decentralized-identity/edv-spec/issues/21
Manu Sporny
38:30
q+
Dmitri Zagidulin
39:47
q+
Dave Longley
39:50
q+
Manu Sporny
41:44
q+ to say I made it up
Dave Longley
43:18
q+ to say, btw, tombstoning and having a sequence number always count up is a good reason NOT to change the name to "revision" ... as people think about revisions as being something you can roll back to.
Dmitri Zagidulin
43:39
+1 to terminology point
Dmitri Zagidulin
45:16
https://github.com/decentralized-identity/edv-spec/issues/4
Derek Trider
45:32
q+ to explain
Manu Sporny
46:08
q+ to ask how we settle this in the long term
Dave Longley
50:53
q+ to mention a relevant development
Derek Trider
51:48
q+
Manu Sporny
54:03
q+
Dmitri Zagidulin
54:06
https://github.com/decentralized-identity/edv-spec/issues/5
Dave Longley
55:19
transferring 33% more over the wire is also painful, so not just about storage at rest (and storage at rest may be easier for servers to optimize in some custom way)
Dmitri Zagidulin
55:35
q+ to mention DAG-CBOR for the Hub task force
Dave Longley
56:22
q+ to mention transfer cost as well
Dave Longley
56:38
q-
Manu Sporny
56:40
/me tears his hair out over DAG-CBOR :P
Dave Longley
56:50
(captured in dmitri's comment already)
Juan Caballero
57:06
by 3box
Juan Caballero
57:44
https://github.com/ipld/js-ipld-dag-cbor
Dave Longley
57:48
q+ to say ... DAG-CBOR not standardized, right? may create a problem with w3c
Manu Sporny
58:19
q+
Juan Caballero
58:22
https://github.com/ipld/specs/blob/master/block-layer/codecs/dag-cbor.md
Juan Caballero
58:54
(Did Ian Preston present to this group or just interact on GH? I thought he was attending regularly for a while)
Dmitri Zagidulin
59:04
I think Ian presented, yeah
Manu Sporny
59:26
q-