Re: [Wikitech-l] Reply: [GSoC 2016] Query about the ideas' project time

2016-03-21 Thread James Salsman
>
>
>
 Hi Linxuan,

Thank you for your question:

 >... What does the "reputation score" in the description refer to?

I've asked Priyanka to reply with her current design, but here is some
of the advice I gave her:

"Each reviewer needs, at a minimum, data indicating the number and
proportion of reviewers who have agreed with them. However, the third level
of tie-breaking review introduces an extra bit for each disagreement which
determines whether agreement or disagreement should be counted in their
favor. So, even if a given reviewer only agrees with 50% of the other
reviewers, the determination of the tie breaker in each case of
disagreement controls whether their reputation score ranges from 0% to
100%. (As too does the agreement proportion, which is unlikely to be
exactly 50%.)

"Do you want the reviewers to know their agreement ratios and reputation
scores? How might their behavior change if they are and aren't told those?
Could there ever be a case when you might want to withhold them? Would
there ever be a benefit from distorting them? How about displaying them as
a range instead of distorting or withholding them? That last possibility
seems superior to me. You might want to do that when you are unsure that
the precision of the mathematical values is near the accuracy of the
knowledge they represent. Do you want to be able to tell each reviewer the
responses which have contributed to defects in their reputation scores,
i.e., do you want them to know which disagreements were tie-broken against
their favor?"

Her reply at the time was:

"In case of two reviewers agreeing, we add a +1 to the reputation. In case
of disagreement, we seek the opinion of a 3rd reviewer. If A says Yes, B
says No and C says Yes to an edit, A and C will have an agreement ratio of
50% and reputation of 100%, whereas B  will have an agreement ratio of 0
and reputation 0%? This would of course change as more edits are reviewed
by them."

I believe that is still an accurate description of the current design.

Finally, I regret that the GSoC program doesn't allow more than one student
per

Best regards,
Jim Salsman
___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

Re: [Wikitech-l] Reply: [GSoC 2016] Query about the ideas' project time

2016-03-21 Thread Andre Klapper
On Tue, 2016-03-22 at 02:14 +0800, DJAKN wrote:
> Thank you for your reply. I have read the instruction and the blog of
> this project and have some ideas now. But there appears another
> question: What does the "reputation score" in the description refer
> to? Should we design it and make some metrics, or just use the
> existed scoring standard?

That seems to be a specific question about the proposed task. 
Maybe the potential mentor (jsalsman) will reply here, but you might
have more luck asking task-related questions [1] on the task itself:
https://phabricator.wikimedia.org/T89416

Cheers,
andre

[1] Also see 
https://www.mediawiki.org/wiki/How_to_become_a_MediaWiki_hacker#Feedback.2C_questions_and_support
 

-- 
Andre Klapper | Wikimedia Bugwrangler
http://blogs.gnome.org/aklapper/



___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

[Wikitech-l] Reply?? [GSoC 2016] Query about the ideas' project time

2016-03-21 Thread DJAKN
Thank you for your reply. I have read the instruction and the blog of this 
project and have some ideas now. But there appears another question: What does 
the "reputation score" in the description refer to? Should we design it and 
make some metrics, or just use the existed scoring standard?


Sincerely,
Li Linxuan




-- 原始邮件 --
发件人: "Niharika Kohli";;
发送时间: 2016年3月21日(星期一) 晚上11:54
收件人: "Wikimedia developers"; 

主题: Re: [Wikitech-l] [GSoC 2016] Query about the ideas' project time



Hi Li,

On Mon, 2016-03-21 at 20:58 +0800, DJAKN wrote:
> > I am interested in the "Accuracy Review" project, but there is a note
> > saying the estimated time for a senior contributor is 3 weeks. Other
> > projects in the idea list have the similar estimated time. So should
> > we complete more than one project during the three-month internship
> > or just one?


A "senior contributor" here refers to someone who's been actively working
with MediaWiki for a quite long time now. Typically, mentors for a project
are the senior contributors. It's a yardstick we use to measure whether a
project is big enough for GSoC.

Niharika
Software Engineer (International)
Wikimedia Foundation
[User:NiharikaKohli]
___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
___
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l