From: Thorsten Ottosen (thorsten.ottosen_at_[hidden])
Date: 2007-11-17 14:09:37
Ronald Garcia skrev:
> Hi Thorsten
> On Nov 16, 2007, at 9:39 AM, Thorsten Ottosen wrote:
>> Also, some of the libs are very small, and should be made as fast-
>> Furthermore, I think very specialized reviews should be able to happen
>> in parallel with other reviews without any problems. The specialized
>> review would simply require that a number of reviews be found in
> A few unexpected turns of events with several reviews in the queue
> have caused delays recently. When you factor those in, getting
> review managers for libraries has definitely been more of a sticking
> I don't think there's any benefit right now to forcing reviews into
> smaller slots or running more in parallel. We are not short on open
> time slots in which perform library reviews.
> More often we are short
> on submitted library reviews. Many recent library review periods have
> been extended in order to get a reasonable number of full reviews
> from boost participants.
> Shortening library rno eviews to fast-tracks
> and running reviews in parallel will shrink the pool of people who
> can provide reviews within the time period and split the pool of
> reviewers, seeing as some reviewers will be interested in multiple
> libraries and but not have time to perform multiple reviews.
>> We could need some wiki or something were interested reviewers could
>> sign up (for being a potential reviewer) and record their
>> experience in
>> the area. This would be very useful for the review managers.
> What do you see as the benefit of such a wiki? It's not obvious to me.
Example: if I had to be review manager for "Floating Point Utilities",
I could see if there was a reasonable list of people with the right
experience so the review could commense.
Very often, there absolutely no overlap between libraries. Furthermore,
if the review manager gathers 5-10 people in advance with experience in
the area of floats, I see absolutely no reason that the review should
wait 10 months for an available time slot.
Gathering reviewers *after* the review starts is often too late, because
people need time to plan well ahead.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk