[boost] Re: Re: FORMAL review process not working well?

From: Paul A Bristow (pbristow_at_[hidden])
Date: 2004-10-27 05:51:38

| -----Original Message-----
| From: boost-bounces_at_[hidden] 
| [mailto:boost-bounces_at_[hidden]] On Behalf Of Peter Dimov
| Sent: 25 October 2004 11:14
| To: boost_at_[hidden]
| Subject: 
| 
| John Torjo wrote:
| > 
| > It's so funny, during reviews, everyone comes up with his own better
| > version of the reviewed library.

For the proposer who has put in much work, it is not so funny ;-(
 
| That's one of the reasons we have reviews, is it not?

Better ideas are Boost's strength, but

IMO review is the wrong time for radically better ideas -
for they are often shooting from the hip, half-baked and in need of much
more refinement.

It seems wrong to me to have work which has been discussed and refined and
tacitly used and accepted over a long time (necessarily waiting in the
review queue) to be suddenly turned upside down at review, often by people
who did not participate in the original discussions - a natural consquence
of the quite rapid Boost membership turnover leading to 'Not Invented Here'
syndrome.

My impression is that many much needed things like formatting, big_integer,
logging, units are not entering the library because of a search for
perfection - largely unattainable as the invaluable filesystem library
shows. (The Standard shows similar symptoms - there has been a desparate
need for a 'typeof' since templates use started, for example, but there are
no Standards, let alone implementations yet).  For many problems, the
Irishman's reply when asked for directions "Well I wouldn't start from here"
applies: but we are committed to start from C++, multiple character sets,
and a mishmash of incompatible operating systems - we have to find some one
way, even if none are ideal.

I agree with other would-be reviewers, that the time available and notice is
too short, unless all we are to do is to finally dot is and cross the ts,
and vote.

Would a longer period of formally 'working on a proposal' be better?
We would need to have a separate place for these 'under consideration'
proposals (not files, not sandbox).

So I foresee the process being:

1  Asking for interest.
2  Floating ideas and working on code.
3  Getting to a point where the submitter(s) feels that he(they) has
something worth considering.
4  A 'Working on a definite proposal' period - a month perhaps.
5  Digesting input and revising, re-grouping, or abandoning, or going back
to the drawing board.
5  Final formal review (a week) and vote.

Paul

Paul A Bristow
Prizet Farmhouse, Kendal, Cumbria UK LA8 8AB
[phone]  [phone]
mailto: pbristow_at_[hidden]