Boost logo

Boost :

Subject: Re: [boost] [metaparse] Review Manager
From: Robert Ramey (ramey_at_[hidden])
Date: 2015-03-19 15:05:20

christophe.j.henry wrote
>>If the first couple of reviews highlight some show stopping problem which
>>forces the author either to retract his review request or accede to
>>significant re-design - the review resources have also been wasted.
> I thought this was the point of having to find a review manager and
> organize
> a review. There is always a risk that the library fails to pass the
> review.
> Otherwise, why have one?

Of course, but why pass up the opportunity to get some advance feedback?
Such feedback will make the library less likely to fail. It's not either/or
it's about making the review process easier and more effective.

> I can only praise your efforts and it is one more tool to get quality
> libraries in Boost. But the "standard" one works sometimes too. We can
> have
> both. The bar is usually so high to find a review manager, if you find
> one,
> chances are that there is interest. Sure, maybe only from a few "geeks",
> but
> these geeks are usually the ones who use these tools to make new
> libraries.

The review process DOES work. I'm a huge fan of it. But we've had concerns
for many years that it's hard to find review managers and that there aren't
reviews. And the review process takes a lot of work. Imagine how much
the the review process would be if there were already 5 reviews posted?
How much easier would be be to get a review manager? As a thought
experiment, consider what would happen if we already had 5 reviews.
a) all reviews are wildly positive - review management is pretty simple
b) all reviews are strongly negative - the review process would end there
c) reviews are mixed - the most common case. The library author would
be strongly motivated to address the complaints as soon as possible.
The final review would more likely to move toward a) above. (BTW, I wonder
if the
improvement in documentation you refer to above is motivated by some
feed back to the author).

So you see, I'm not proposing an alternative to the review process. I'm
trying to address the things which have made it less effective in the past
without in anyway diminishing it's value - or even altering the prodedure
itself. I think I found a good improvement with no downsides.

So, I would encourage anyone who expects to participate in this review
to post their review on the incubator at their earliest convenience. You'll
have it done already when the review actually comes up, and you'll give
the author the option to act on your feedback earlier and maximize his
chances of getting his library accepted.

Robert Ramey

View this message in context:
Sent from the Boost - Dev mailing list archive at

Boost list run by bdawes at, gregod at, cpdaniel at, john at