Boost logo

Boost :

From: Beman Dawes (bdawes_at_[hidden])
Date: 2002-01-24 12:31:30


At 06:32 PM 1/23/2002, bill_kempf wrote:

>This brings up a question. For changes such as this do we need a
>formal review period before integrating into the library? It may not
>seem so for this small change, but when I add the new concepts, such
>as read/write locks, I'm not so sure the answer is the same.

We need procedures for a "Mini-Review" - less formal than a full review,
but gives others a chance to review important changes to already reviewed
libraries.

Sometimes this would cut it when a new library is accepted subject to so
many changes a mini-review is necessary.

Other times because a stable library is about to get major changes.

Might mini-reviews work something like this:

* Announced at will by the developer. [to cut administrivia]

* Announcement subject always to include Mini-Review Request" [to alert
both human readers and email filters]

* To run at least 7 days. [to ensure time for comments]

* Developer responsible for rattling people's cages to get at least one
response from other than the developer.

* Developer responsible for responding to comments, and deciding if they
are serious enough to delay acceptance.

Thoughts?

--Beman


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