CFV: New JDK8 Committer: Deven

John Rose john.r.rose at oracle.com
Mon Jul 22 16:35:03 PDT 2013


First, I want to thank Deven for working on our shared code base.  This is how it should work.

Unfortunately, I also have to question the timeliness of this call for votes.

Quoting http://openjdk.java.net/projects/#committer-vote :
> Nominating a Contributor or Author to be a Committer
> 
> Any Committer may nominate a Contributor to be a new Committer. Becoming an Author may be helpful in establishing credibility to become a Committer, but it is not required.
> 
> As a rough guide, a Contributor should make at least eight significant contributions to that Project before being nominated.


About six of Deven's change sets are what I would characterize as point fixes, and it is hard to see them (either individually or collectively) as fulfilling the requirement of making "significant contributions to the Project".

I appreciate the carefully built regression test code that comes with each of Deven's change sets, including those I regard as "point fixes".  This is definitely what we want to see in our engineering.  And "point fixes" are also what we want to see as well as large changes.  A flow of such fixes and updates are very definitely significant and important to the success of this project.

But as a whole, I think this set of contributions is light in weight compared to the "rough guide" quoted above.  In cases like this I would prefer to see a some larger changes, or a longer list of small changes.

Looked at another way, the total volume of these change sets (with metadata and context lines) is about 800 lines of text.  In many cases, a *single* significant contribution to our code base is of that scale.  The average (exported, non-merge) changeset size of the last 100 commits to jdk8/jdk is about 500 lines.  (There is of course great variation; a geomean comes to 135 lines; Deven's geomean is 76.)  We would expect the required minimum of 8 commits, on average, to be much larger than 800 lines, even if it is somewhat less than the average size of 4000 lines.

Therefore, I have some doubt that, as a whole body of work, this collection of nine changesets matches the "rough guide" for a call for votes.

Of course, every single contribution to our community is significant, but if the words of the "rough guide" are to mean anything, a set of very small fixes is not sufficient to make up the initial set of contributions.

I understand that we need committers, and so I hesitate to throw this cold water on our intake process.  And I will not interfere with Deven's progress here.

But, I encourage those who call for votes to ponder the meaning of the word "significant" in the guide, and be ready to answer queries like mine.  If candidate has authored small change sets, there should either be something unusually functional about them, or there should be significantly more than the minimum number (8) of them.

I am abstaining from this vote.

Best wishes,
— John

P.S.  The patch size extraction command I used is:
  (for rev in $(hg log -l100 -M --template '{node|short}\n'); do echo $rev $(hg export ${rev}|wc -l); done)
The repo is an updated copy of http://hg.openjdk.java.net/jdk8/jdk8/jdk .
I spot-checked the largest 10% of the contributions and found they did not look unusual.

On Jul 22, 2013, at 2:26 AM, Jonathan Lu <luchsh at linux.vnet.ibm.com> wrote:

> I hereby nominate Deven to JDK8 Committer.
> 
> Deven (youdwei) has contributed 9 patches to the JDK 8 repositories.
> 
> http://hg.openjdk.java.net/jdk8/jdk8/jdk/log?rev=youdwei
> 
> Votes are due by 02:20:00 PDT,  August 5, 2013.
> 
> Only current jdk8 Committers [1] are eligible to vote on this nomination.
> 
> For Lazy Consensus voting instructions, see [2].
> 
> - Jonathan
> 
> [1]http://openjdk.java.net/census
> [2]http://openjdk.java.net/projects/#committer-vote
> 



More information about the jdk8-dev mailing list