From: Tobias Schwinger (tschwinger_at_[hidden])
Date: 2007-08-01 18:49:47
Paul A Bristow wrote:
>> -----Original Message-----
>> From: boost-bounces_at_[hidden]
>> [mailto:boost-bounces_at_[hidden]] On Behalf Of Tobias Schwinger
>> Sent: 30 July 2007 23:35
>> To: boost_at_[hidden]
>> Subject: [boost] [x-files] the code's out there...
>> There is lots of useful, Boost-worthy utility code out there:
>> Implementation details of existing / upcoming / (maybe even) proposed
>> and rejected libraries that could stand alone, code on private hard
>> disks never brought up for review, additions to existing
>> libraries, etc.
>> Fact is, that for small submissions the administrative overhead of
>> (even) a formal (fast-track) review can be too high for developers'
>> decency. Just imagine someone allocating five fast-track slots at once
>> just for factoring out some unrelated bits of code from an
>> existing library.
>> So, why not group a bunch of unrelated utilities into a set of "Boost
>> X-Files" and have them reviewed as a single library? It's completely
>> non-intrusive to the review process and could seriously increase its
>> efficiency. Like the idea?
> In a way we have this already, in an unstructured way - called the vault ;-)
:-) Well, not really. Some stuff is just highly experimental or
unfinished or not even intended to be brought up for formal review.
> But I have been advocating the idea of a 'Boost candidate' status for some time to:
> * get code of the type you discuss into use.
> * get code used 'in anger' to provide user experience and feedback.
> * to provide more incentive for developers to do the hard work getting stuff up to Boost release standard.
Theoretically brilliant. But does the community have enough free
resources to get it going?
> * judge what is good enough for the agreed high cost of a formal review (or a 'light-touch' review for tiny but useful items?).
And this one needs some kind of authority, a "pre-review manager" so to
say, doesn't it?
Still, a pre-review seems a good measure for framework-style submissions
or infrastructure libraries. For small submissions, however, your
suggestions seem to make things even more complicated AFAICS.
Given components such as SmartPtr, Function, Bind, Variant, Optional,
etc. are the majority of the most broadly useful components in Boost, we
should probably forward small submissions (note I'm looking at the size
of the interface - not the implementation) in days of a full review
queue and many framework-style submissions. The latter are often bound
to a particular problem domain (or sets of problem domains), take more
time to evaluate and thus there are fewer reviewers available. A bunch
of either broadly useful or real tiny utilities, OTOH, requires fewer
domain expertise and time, so there are more potential reviewers and a
collective review will free review slots and other reviews from time
pressure. It might make the whole process scale better to do justice to
a particular submission without even touching the process and just
changing its input.
Also just theory, so far :-).
> Perhaps we could ask for several nominations to promote from the vault to the 'candidate' collection, or we could have a straw poll,
> or ??
I believe it's fine if authors decide (not at least because other
alternatives require manpower). Of course, the usual guidelines apply
(to be checked during the pre-review inspection done by the Review
Manager). If there's an inferior part, the review will detect that it
either needs to be changed or dropped.
It might actually work...
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk