Lorenzo Nicora created FLINK-40282:
--------------------------------------
Summary: Expand and correct the concepts Glossary
Key: FLINK-40282
URL: https://issues.apache.org/jira/browse/FLINK-40282
Project: Flink
Issue Type: Improvement
Components: Documentation
Reporter: Lorenzo Nicora
The Flink Glossary page contains imprecise and ambiguous definitions, and has
gaps for core terms like Checkpoint, Watermark, or Parallelism.
It also focus on DataStream API only.
1. _JobGraph_ documented as a synonym for _Logical Graph_, collapsing a real
stage out of Flink's translation pipeline. Flink actually does _StreamGraph_
--> _JobGraph_ --> _ExecutionGraph_.
2. _Task_ and _Sub-Task_ are conflated
3. Core terms missing: _Checkpoint_, _Savepoint_, _Watermark_, _Parallelism_,
_Channel_, _Source Split_.
4. Three dead cross links to _Function_ and _Record_.
5. No Table API / SQL coverage. The glossary was DataStream-only; a SQL user
had no entry point.
6. _Partition_ conflates three things: logical partitioning, physical
partitioning, and external-system partitions (Kafka topic partitions).
7. Capitalization of defined terms is inconsistent
8. Entries are not in alphabetical order.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
