Boost logo

Boost :

Subject: Re: [boost] encouraging review managers -- was Re: Review Request:Variadic Macro Data library
From: Gordon Woodhull (gordon_at_[hidden])
Date: 2011-02-21 00:55:39


On Feb 21, 2011, at 12:05 AM, Robert Ramey wrote:

> Gordon Woodhull wrote:
>> On Feb 20, 2011, at 8:37 PM, Edward Diener wrote:
>>
>> Fair enough. This doesn't worry me very much, but
>> - I think there should be a way to challenge a review afterward,
>> through the Wizards and/or list as appropriate.
>
> Actually there is already a mechanism for addressing this. I
> know this for a fact because I personally availed myself of it.
> The procedure can be summarized as follows:
>
> a) Understand and accept the objections raised the review which resulted
> in the library being rejected,
>
> b) redo the code/documentation
>
> c) re-submit.

Yes, this works great and I hope that I'm emphasizing that side enough.

If a library is rejected, the author can rewrite and seek a second review with (IMO?) a different review manager if necessary.

Ed's concern is with the far thornier issue of, what if you think a library was improperly *accepted*, if the review manager was too biased in favor of the library?

Aside from authors of competing libraries (who are free to submit theirs or try to get their features merged), and Luddites, have you ever seen someone severely angered by a result of acceptance?

I'm not saying it can't happen, and I want to be sure there is a mechanism to resolve conflicts in case of incompetent or over-favoring review managers.

If we admit that Review Wizards can't tell beforehand whether a Review Manager is going to do a good job, as Joachim and I are arguing, then people should just volunteer to manage reviews and not bother registering with the Wizards for a nonexistent queue.

But it is possible that these volunteers could be nincompoops or shills, and the Review Wizards and the whole community need the power to remove a bad manager or nullify the results.

The classic case might be, what if some bad company submitted a library that would lock everyone using C++ into their products (not sure how this would happen, but bear with me) and then "volunteered" one of their employees to manage the review? At that point I think someone should complain to the Wizards, who have the authority to reject managers. (I doubt this is exactly what Ed is worried about, just musing.)

Gordon


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