Re: [boost] [process] [sorting] [singularity] [compute] [others] Who needs review manager?

Subject: Re: [boost] [process] [sorting] [singularity] [compute] [others] Who needs review manager?
From: Antony Polukhin (antoshkka_at_[hidden])
Date: 2014-09-14 06:24:16

2014-09-14 0:25 GMT+04:00 Robert Ramey <ramey_at_[hidden]>:

> Antony Polukhin wrote
> >
> > I'll take a closer look to the library and write a personal note with
> > comments/notes as soon as I get some free time.
> >
> > But first of all, I'll take care of the Compute review :-)
>
> I'm wondering if perhaps you've got confused about the differing roles
> of review manager vs reviewer.  IRC a review manager doesn't review
> libraries himself, he reviews the reviews of others, weighs them and
> arrives at a acceptance/rejection.  Think the US court system
>
> judge - review manager
> jury - reviewers.
>

I'm not going to write a full review of the library, I'm just trying to
satisfy the first bullet for Review manager:
http://www.boost.org/community/reviews.html#Review_Manager

In other words, there is a need to look through the sources, docs and build
system, make sure that installation notes are sufficient, tell the author
about missing/weak parts. For example: Compute library looks good, however
its docs have no performance section. This is important, because in most
cases users will use Compute library to speedup their code.

If a lot of notes appear during such "preliminary investigation", then
it's worth writing a review/comment to Boost Incubator. Compute library is
not such case.

-- 
Best regards,
Antony Polukhin