[
https://issues.apache.org/jira/browse/KAFKA-7315?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16586709#comment-16586709
]
Matthias J. Sax commented on KAFKA-7315:
----------------------------------------
`Serde` are a Streams concept – it's an abstraction that wraps a serializer and
a deserializer in a single class. We can of course add serialization and
deserialization docs for producer and consumer respectively.
No sure, if a centralized place that explains all three things for the 3 APIs
would make sense. If we document serialization/deserialization within Consumer
and Producer docs (or maybe have a section "data types" for both), the `Serde`
datatype section in Streams can just link to both, and only explain how to
create corresponding Serdes using the plain Serializer/Deserializer (similar to
what we have for JSON already).
Adding other formats than AVRO makes sense to me in principle – however, we
should be careful to not blow up the docs – there are many hundreds of data
formats out, and we don't want to document all of them. As you pointed out,
Kafka uses byte[] format internally and there is no good reason to add
unnecessary docs on "how to (de)serialize datatype X".
The question we should answer is, do we add value adding those docs, or do we
duplicate things that are documented somewhere else?
So don't get me wrong. I am always in favor of improving/extending the docs in
general. We should just think/discuss about what makes sense and what not.
Looking forward to the opinion of others.
> Streams serialization docs contain a broken link for Avro
> ---------------------------------------------------------
>
> Key: KAFKA-7315
> URL: https://issues.apache.org/jira/browse/KAFKA-7315
> Project: Kafka
> Issue Type: Improvement
> Components: streams
> Reporter: John Roesler
> Priority: Major
> Labels: docuentation, newbie
>
> https://kafka.apache.org/documentation/streams/developer-guide/datatypes.html#avro
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
