Boost logo

Boost :

Subject: Re: [boost] encouraging review managers -- was Re: Review Request: Variadic Macro Data library
From: Gordon Woodhull (gordon_at_[hidden])
Date: 2011-02-20 19:58:40


On Feb 20, 2011, at 2:24 PM, Edward Diener wrote:
> On 2/20/2011 12:12 PM, Joachim Faulhaber wrote:
>> 2011/2/19 Edward Diener<eldiener_at_[hidden]>:
>>> On 2/19/2011 3:57 PM, Gordon Woodhull wrote:
>>>> Ed, you volunteered to be a review manager about a month ago. Have you
>>>> approached any of the authors of prospective libraries on
>>>> http://www.boost.org/community/review_schedule.html
>>>> who are listed as needing review managers?
>>>
>>> I was told to contact the review wizards and offer up my services to review
>>> libraries. I did so, and mentioned the libraries I felt I could review. As I
>>> understand it, it was then up to the review wizards to determine whether I
>>> was qualified to review a library and to contact me about doing so if they
>>> thought that I was. I was not subsequently contacted.
>>
>> !!
>> This is at least delicate and definitely "not amusing".
>> (1) Ed volunteered to be Review Manager for a couple of libs
>> (2) Ed is around in the Boost community for quite some time including
>> BoostCon
>
> I have to correct that. I have never been to BoostCon.

I object to the idea that going to BoostCon makes one a better member of the community. While BoostCon is truly amazing and I'm glad to have seen and heard people, proving oneself helpful and knowledgeable enough to manage a review happens right here on the list, and in the code.

>> and is going to be first time library contributor. As an
>> active Boost member he should qualify to be an RM, shouldn't he. At
>> least he deserves a respose!
>> (3) Many libs still don't have an RM
>> AND
>> (4) he was not subsequently contacted!
>
> Thanks for the boost on Boost, but I am fine with the decision.

You thought that they decided they shouldn't assign you a review to manage because they didn't get back to you? I sincerely doubt that; every reply I saw was positive. The "review queue algorithm" is broken.

> I do believe that having so many libraries waiting for reviews, which puts the potential, at least, for a library being added to Boost as some time pretty far relatively in the future, is not a great thing.

Right.

> I made the suggestion in the past that more than one review going on at a time, and a longer review process for each library, be allowed in order to get libraries reviewed more quickly.

Yes!

> I honestly do not see that the slowness of the review process has much to do with libraries having review managers or not. Usually a library which is scheduled for review fairly soon will get a review manager somehow.

More like, a library that has a review manager gets scheduled quickly.

> At the same time the process for a library submitter finding a review manager for his library seems very odd to me. One posts a message on this mailing list and hopes someone responds saying that they are willing to be the review manager. if no one responds, what does one do then ?

You're right, it would be nice if the Wizards could help out newcomers. But really, the best way for this to happen is for one of the thousands of people who read this list to step forward and express interest.

> If someone responds, and the library implementer does not know that person from the mailing list, how does one choose whether or not that person is acceptable or not ?

I think you just have to trust that the person is competent and not hostile, otherwise why would they have volunteered? A couple of times I have seen someone volunteer to manage a review on the list, and then someone else who was obviously more competent jumped in and said "no, actually I should." And the first person defers.

I think RMAs (although I object to that name - more on that later) can really help out here, because sometimes the really really experienced people don't have much time to do a totally thorough write-up. I have never doubted the decisions and every report is a fascinating read, but I like when the review report takes the time to summarize every single issue that made any sense, and how it was resolved.

> In a real way I would rather a review wizard go through a list of people which he knows are knowledgable and experienced enough to be a review manager and contact each of those people until he finds one to be a review manager for a library. It would be much easier than placing the burden of finding a review manager for a library on the library submitter.

I agree it would be nice, especially if the Wizards knew all of the potential review managers so well that they could just decide the right person. Infinitely wise and perfectly tuned in to the competencies of the entire community...

I'm not being sarcastic, that would be really nice!

> It would also almost assuredly mean that the review manager would have little personal bias approving or not approving a library for inclusion into Boost at the end of the review process.

You seem to be hoping that the system can ensure objectivity. But you are talking about a group of very passionate, brilliant programmers! Of course the review manager is going to have opinions.

[Case in point: Christophe Henry, who is using my library MPL.Graph, has volunteered to manage the review. Of course he is going to be biased toward acceptance, but I don't doubt that he'll be objective enough to take any No votes or Conditions on Acceptance seriously.]

But I've never seen someone maliciously volunteer to manage a review because they wanted to reject the library, although I've seen review managers reluctantly vote against the library. Generally if a library is rejected the author is encouraged to rewrite and resubmit, and they could certainly request a different review manager the next time.

Cheer up, this is a friendly place, if strongly opinionated. Everyone wants you to do your best work.

Gordon


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk