[
https://issues.apache.org/jira/browse/HADOOP-12065?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Kai Zheng updated HADOOP-12065:
-------------------------------
Status: Patch Available (was: Open)
> Using more meaningful keys in EC schema
> ---------------------------------------
>
> Key: HADOOP-12065
> URL: https://issues.apache.org/jira/browse/HADOOP-12065
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: io
> Reporter: Kai Zheng
> Assignee: Kai Zheng
> Attachments: HADOOP-12065-HDFS-7285-v1.patch
>
>
> Commented by [~szetszwo] in HADOOP-12029:
> bq. We should also change "k" and "m" in the following to something like
> "numDataUnits" and "numParityUnits".
> {code}
> public static final String NUM_DATA_UNITS_KEY = "k";
> public static final String NUM_PARITY_UNITS_KEY = "m";
> {code}
> This sounds good to me, as in fact in many erasure coding related papers,
> discussions, and implementations, though {{k}} and {{m}} are broadly used,
> the meanings are not consistent. For example, {{m}} can also mean
> {{numDataUnits + numParityUnits}}, all the rows in the coding matrix. The
> suggested names are more clear and explicit.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
