Update bit mapping in README.md

[susan-pgedge]

This issues was reported by Asif N - the bit mapping in the readme is off... in this patch I've updated it to match the values he has shared.

[ashwinstar]

(no specific to this PR) I am curious to learn on the design choice -

• why timestamp was introduced? ideally node-id + counter should give unique sequence values to each node as well
• If timestamp with millisecond precision is used then why is there a need to WAL log the updates to this sequence type?

[luss]

I'll attempt to answer question #1; and leave it to a Spock internals developer to answer #2. The reason to use a monotomically increasing timestamp, instead of a simple counter, is that the valuable information about the system clock time of the node is embedded in the Snowflake Sequence and can be use for sorting, filtering, custom-conflict resolution, ....

…

On Fri, Apr 25, 2025 at 2:51 PM Ashwin Agrawal ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In README.md <#18 (comment)>: > @@ -7,8 +7,8 @@ Internally, `snowflakes` are 64 bit integers represented externally as `bigint` ``` bit 63 - unused (sign of int8) bits 22-62 - timestamp with millisecond precision (no specific to this PR) I am curious to learn on the design choice - - why timestamp was introduced? ideally node-id + counter should give unique sequence values to each node as well - If timestamp with millisecond precision is used then why is there a need to WAL log the updates to this sequence type? — Reply to this email directly, view it on GitHub <#18 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAMWOHSZYZGATJMHTE4KY7T23J7ZTAVCNFSM6AAAAAB2575UTKVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMZDOOJVGAYDQMZVGY> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>