docs(cip-19): add NamespaceRangeID along with NamespaceRangeContainer #197
Conversation
vgonkivs
force-pushed
the
docs_cip19_range_container
branch
from
1e96e3e
to
1ffad29
Compare
cips/cip-19.md
Outdated
| @@ -311,6 +311,49 @@ Namespace data may span over multiple rows, in which case all the data is encaps | |||
| containers. This enables parallelization of namespace data retrieval and certain [compositions](#protocol-compositions) | |||
| may get advantage of that by requesting containers of a single namespace from multiple servers simultaneously. | |||
|
|
|||
| #### NamespaceRangeID | |||
There was a problem hiding this comment.
namespace range may suggest it's a range of namespaces, like 0x00000001 up to 0x00000004. Maybe just NamespaceData would be better? just dropping Row prefix from RowNamespaceData, also better showing correlation between the two?
There was a problem hiding this comment.
We have:
RowNamespaceData - namespace data within a row
NamespaceData - all the namespace data that can span over multiple rows
To avoid namespace confusion I propose to do:
NamespaceRange -> RangeNamespaceData - namespace data range that can span over multiple rows
There was a problem hiding this comment.
Agree with @Wondertan. The container itself and request namings should contain Range
|
|
||
| ```text | ||
| NamespaceRangeID { | ||
| SampleID; |
There was a problem hiding this comment.
what is the purpose of sample here? does it mean take Length shares from given SampleID? I think it needs better clarification below. Also why not just using EdsId + Namespace providing whole namespace from given block? Is there a usecase for a user to request his shares but eg. skipping the first and last one?
There was a problem hiding this comment.
does it mean take Length shares from given SampleID?
Yes
There was a problem hiding this comment.
Also why not just using EdsId + Namespace providing whole namespace from given block? Is there a usecase for a user to request his shares but eg. skipping the first and last one?
This means we will retrieve the whole eds from the requested namespace. For example, blob.GetProof needs a specific range of shares rather than the whole namespace.
There was a problem hiding this comment.
should we then specify here what should happen if cid is pointing to shares outside of it's namespace? eg. what if the sample id is not in correct namespace or the length goes beyond the namespace?
There was a problem hiding this comment.
also I'm not exactly sure what is the usecase here, user would need to already know the blob's position in eds and it's length in shares to even use that, however shwap itself doesn't give any of such abilities
| SampleID; | ||
| Namespace; | ||
| Length: u16; | ||
| OmitData: bool; |
There was a problem hiding this comment.
should OmitData also be present in RowNamespaceId?
There was a problem hiding this comment.
It may also be part of SampleID, i.e., one can request just the proof for the share/sample without the data itself. However, we are unaware of any use cases for the Row and Sample containers that can be requested without data. We could add that to be consistent with OmitData usage across types, but our approach is to specify and implement only what's known to be helpful in practice.
Co-authored-by: Rootul P <[email protected]>
|
Moving this PR to draft until we have a working prototype to avoid any changes in the future. |
Overview
Added NamespaceRangeID that forms a request to get a range of namespace shares along with their inclusion proofs. The shares MAY span across multiple rows in the DataSquare.