Subject: Re: [boost] Formalising the review process into a well specified workflow
From: Niall Douglas (s_sourceforge_at_[hidden])
Date: 2015-06-03 11:28:58
On 3 Jun 2015 at 16:06, Phil Endecott wrote:
> > 5. Ideally in the future a review manager would have their own form
> > with boxes to tick, and part of their form would be to check the form
> > the library author filled in i.e. the forms themselves feed into one
> > another as part of the programmed workflow.
> No no no no no no no.
> What we want is (a) common sense to be applied, and (b) domain
> experts to actually do reviews, and (c) in the absence of too
> many of those, non-experts to at least try the library and
> express their opinions or drive the discussion by asking questions.
I was thinking of form questions for review managers such as:
"Before beginning the review have you verified the library to be
reviewed compiles on at least two compilers?" Tick yes/no.
"Have you included a link to the root of the library documentation in
the review announcement email?" Tick yes/no.
It's not rocket science. But it makes the exact process and workflow
crystal clear for the review manager involved instead of trying to
interpret words scattered over at least three web pages, and then
finding some reviewers have very different expectations to others
because you didn't use conventional idioms prespecified by the
This is all about saving everybody time, hassle and confusion. Not
about prespecifying how reviews are written, understood, or
-- ned Productions Limited Consulting http://www.nedproductions.biz/ http://ie.linkedin.com/in/nialldouglas/
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk