[
https://issues.apache.org/jira/browse/LUCENE-9855?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17314413#comment-17314413
]
Julie Tibshirani commented on LUCENE-9855:
------------------------------------------
Thank you [~tomoko] for continuing to move this forward! I hope this doesn't
distract from our focus, just wanted to clarify an aspect. I think we may be
discussing two separate questions at once:
1. What should be the name of the single new format interface we're
introducing (currently called {{VectorFormat}})? Based on LUCENE-9322 and other
discussions, there was consensus for making the API is general enough to allow
different approaches to NN 'under the hood'. So it seems good to not mention a
specific NN algorithm in the interface name. For me this is the focus of this
current issue, finding a better name for the format interface.
2. What should be our strategy to support multiple NN implementations? The
current design is to have a single format implementation like
{{Lucene90VectorsFormat}} that can handle multiple approaches. But perhaps each
approach should have its own format implementation instead, like
{{Lucene90HnswVectorsFormat}} or {{Lucene90ClusterVectorsFormat}}, etc. This
would affect the name of the current HNSW format implementation (but not the
interface). Based on recent emails, there doesn't seem to be consensus on this
question. Maybe we could spin off a new issue to discuss further?
> Reconsider codec name VectorFormat
> ----------------------------------
>
> Key: LUCENE-9855
> URL: https://issues.apache.org/jira/browse/LUCENE-9855
> Project: Lucene - Core
> Issue Type: Improvement
> Components: core/codecs
> Affects Versions: main (9.0)
> Reporter: Tomoko Uchida
> Assignee: Tomoko Uchida
> Priority: Blocker
>
> There is some discussion about the codec name for ann search.
> https://lists.apache.org/thread.html/r3a6fa29810a1e85779de72562169e72d927d5a5dd2f9ea97705b8b2e%40%3Cdev.lucene.apache.org%3E
> Main points here are 1) use plural form for consistency, and 2) use more
> specific name for ann search (second point could be optional).
> A few alternatives were proposed:
> - VectorsFormat
> - VectorValuesFormat
> - NeighborsFormat
> - DenseVectorsFormat
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
