Skip to content

Commit d1dff7c

Browse files
authored
Merge pull request #396 from clbarnes/rfc9-c3
RFC 9 comment 3
2 parents bea79a1 + 8e8b905 commit d1dff7c

File tree

1 file changed

+39
-0
lines changed

1 file changed

+39
-0
lines changed

rfc/9/comments/3/index.md

Lines changed: 39 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,39 @@
1+
# RFC-9: Comment 3
2+
3+
(rfcs:rfc9:comment2)=
4+
5+
## Comment authors
6+
7+
This comment was written by: Chris Barnes, German BioImaging
8+
9+
## Conflicts of interest (optional)
10+
11+
None
12+
13+
## Summary
14+
15+
I have implemented a [simple rust-based CLI tool](https://github.com/clbarnes/ozx) for writing a filesystem OME-Zarr hierarchy into a .ozx.
16+
17+
## Minor comments and questions
18+
19+
### Use case documentation
20+
21+
The limitations of the ZIP format are presented well in the RFC;
22+
it will be important for the eventual spec changes to make clear
23+
the use cases for which .ozx is and is not appropriate.
24+
e.g. as a write-once read-many format for small datasets for archival or transport purposes.
25+
26+
### Beyond OME-Zarr
27+
28+
The recommendations made by the RFC (zarr.json at the root, sorted metadata files)
29+
are valuable to all single-file Zarr users,
30+
and the only OME-Zarr specific elements are arbitrary markers
31+
(the ZIP comment and file extension).
32+
While no readers are currently optimised for sorted metadata reads,
33+
exposing a larger community to that access pattern makes it more likely that one will eventually be written.
34+
Would it be possible to submit this meta-format as e.g. a [zarr convention](https://zarr.dev/conventions/), with more generic markers?
35+
36+
## Recommendation
37+
38+
I recommend exploring whether this RFC can be generalised to a broader Zarr context,
39+
although if doing so would cause significant delays, accepting it anyway.

0 commit comments

Comments
 (0)