From: John Phillips (phillips_at_[hidden])
Date: 2007-06-07 10:23:59
David Abrahams wrote:
> on Wed Jun 06 2007, "Tom Brinkman" <reportbase-AT-gmail.com> wrote:
>>>>Posts a notice of the review schedule on the regular boost mailing list,
>>>>the boost-users mailing list, and the boost-announce mailing list.
>>>>>The submitter can do that.
>>No. The submitter should not need to do that.
> Yes, I think another thing that's being overlooked is the role played
> by a review manager of helping the submitter (for lack of better
> words) to be comfortable. Announcing on his or her behalf is just one
> of those things that makes the submitter feel as though he or she is
> not "doing this entirely alone."
> Much as I wish it were otherwise, submitting a library to our process
> can be daunting, and having a review manager can help with that, if
> only a little.
This point may deserve a little expansion.
When I ran the review of the units library, one of the first things I
did after volunteering was to go back and look at the discussion of Andy
Little's quantitative units submission. I was a part of this
conversation, but I wanted to be sure I recalled it clearly. One of the
things that struck me was that there were really two separate
discussions tangled together. One was about the quality of
implementation and documentation of the library Andy made, and the other
was about whether his design goals were the right choice. Unfortunately,
in much of the review the posts weren't clear about which of the two
issues they were concerned about. I don't think any of us had considered
it that way while we were doing the review.
Before the review of the units library, I wrote to Matthias and
Steven and suggested that we try to keep the two discussions clearly
separated. Mathias had also participated in the review of Andy's
library, and the discussions in the review drove many of his design
decisions. He made some choices in design that were very different from
Andy's and he understood that not everyone would agree.
In my opinion, trying to make that distinction clear was an important
part of making the review as productive as it was. Also, when the
question came up after the library was accepted about what makes this
library different from Andy's, it was easy to answer and support.
This is the sort of process that someone who has been part of the
boost developer list for years and who has participated in reviews will
know to do, but some very good libraries are submitted by people who
don't have such a long history with boost. We could say that it is tough
for them, and they should do their homework, but I prefer providing a
little more support than that.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk