db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Susan L. Cline" <home4...@pacbell.net>
Subject Re: New contributors?
Date Thu, 07 Sep 2006 00:53:58 GMT
--- Daniel John Debrunner <djd@apache.org> wrote:

> Susan L. Cline wrote:
> 
> > Hello Derby-dev, 
> > 
> > I'm writing to point out an experience which discouraged me about growing our community
with
> new
> > contributors. 
> > 
> > Recently I submitted a patch for a JUnit test.   I have contributed very few patches
in the
> form
> > of code to Derby (I've had just a couple for the Eclipse plug-ins.)  I am not a
Java developer
> by
> > trade or training, but I have contributed to the Derby community by writing 
> > articles, sample applications and some wiki entries. 
> > 
> > My JUnit contribution received comments by two different folks.  I believe that
some of the
> > comments were valid and helpful to me to improve my programming skills and also
to help insure
> > that the test I wrote was readable and understandable, as well as to fit into the
constraints
> that
> > JUnit imposes as well as the Derby test harness. 
> > 
> > However, I feel that some of the comments were more style-related or were nits that
the
> individual
> > reviewer had.  This is where I feel discouraged. 
> 
> As one of the reviewers I'm sorry you felt discouraged, that was never
> my intention and I'm sure it wasn't Kristian's (as the other reviewer).
> 
> Looking back at the set of comments I can see comments about various
> style issues, but to my reading none of them seem to say "you must do it
> my way", in fact they include phrases like "Maybe a style issue, but I
> prefer not to have constants like", and "2 and 3 are nits, use your own
> judgement. ". I personally could use help to know exactly what
> discouraged you, so that any review comments in the future would not
> discourage contributors.
> 

I felt discouraged because I think if it is a style issue or it is a nit, then why
does it need to be mentioned?  I'm not sure if others feel this way, but I can say
personally, when submitting a patch (which I mentioned I have rarely done) I already feel
a little bit nervous about it being coded correctly, without regard to style.

Maybe another example in an area I feel more confident in will help to illustrate the point.
If someone takes the time to contribute to the Derby wiki, or to write an article about Derby
I would hope the community would provide feedback about the technical accuracy of the article
without regard to a specific style.  We all have different styles - for someone who writes

frequently and fluently that style may be seen as more preferred, but for someone who does
not
trying to achieve this same style may be forced and artifical.
I would hope that we as a community would allow these differences.

I'm not sure if I am making my point more clearly to help you understand, but I guess it is
an attempt at explaining if you feel like you are contributing something and you are already
worried about the technical accuracy, you may not be feel very encouraged when given feedback

about the style of something.

> > One of the biggest goals I think we have is to
> > embrace contributors (and therefore their contributions) and to guide them 
> > where it is necessary, but also to accept contributions that represent "incremental
work." 
> Also,
> > I think when someone makes a contribution it has something to do with that individuals
"itch".
>  In
> > this case I thought it would be fun to learn JUnit and become more familiar with
the 
> > Derby test harness, and that was my itch.
> > 
> > I am fortunate in that I am paid to work on Derby, but I think of other individuals
who are
> not
> > paid to work on Derby.  If they make a contribution which is not accepted, partially
due to
> style
> > issues, how likely are we to get new contributors?   Again, I'm not saying that
I wrote the
> > perfect test and others are not welcome to comment on my code, but on the other
hand if we
> make
> > the barrier to contribute (and it was just a test!) this high, how likely are we
to get
> > contributions from individuals who have a lesser stake in Derby than their own livelihood?

> 
> I agree it would be very bad if the Derby community did not accept
> patches that practiced incremental development, or based upon style
> issues. I don't think that has ever happened.
> 
> I guess I'm a little confused by your last two paragraphs, the community
> should "guide them" in the first paragraph, which is what I thought the
> review comments were, but in the second paragraph it seems you view the
> same review comments as a barrier.

I believe I differentiated between types of comments.  Those that are based on
technical issues and comments and those that are style or "nit" comments.  I believe
that the community should guide them via the technical comments, but if the comments
become based upon style or personal nits then yes, they are a barrier to submission,
or at least to iteratively submitting a patch.  

> 
> Since Apache is a peer based community, maybe it should be made clearer
> that reviewers' comments are just "their comments", the contributor
> doesn't have to follow all, or even any, of the advice. Any committer
> can commit any patch that they "have a high degree of confidence in".
> 
> In my opinion a committer may gain confidence in a patch by requesting
> certain changes in it through review comments. If those changes don't
> happen or partially happen, then committer can re-evaluate their
> confidence level in the patch, possibly leading to committing the patch.
> If the committer decides not to commit the patch, then it's possible
> they may modify it to have confidence in it and then submit it, though
> this might occur much later when the committer (or anyone else) has time.

This makes sense to me, and in a way is part of the reason I am confused
by the process.  After your gave me your comments I believe I addressed all of
them, even the "style" comments.  However, then I received a second set
of comments from another person which were different from yours. 
 
This is a somewhat rhetorical question, but should it be the responsibility of the 
contributor to address all the comments of different committers? I would imagine this
boils down to the desire the contributor has to getting his or her patch committed,
which again, I see as a barrier to contribution.

 
> Thanks,
> Dan.
> 
> 
> 


Mime
View raw message