Subject: Re: [boost] Interest in a "concepts lite" library?
From: Robert Ramey (ramey_at_[hidden])
Date: 2016-04-22 14:47:00
On 4/22/16 10:09 AM, Paul Fultz II wrote:
> Also, Robert Ramey has done a lot of writing about type requirements, I would
> like to see if I can utilize some of that for introductory documentation
> perhaps, or maybe Robert could help with modifying some of the writings to
> focus it on the Tick library with the purpose of having it included in the
> Tick documentation.
LOL - If I had nothing else to do I'd re-write huge swaths of Boost
Documentation. The funny thing is, I don't consider myself all that
great a writer. But I think I've done a pretty good job recently on
things like the safe numerics library. At least it seems that those
interested in criticizing it are not doing so due to misunderstanding
about what it does and on one has criticized the documentation.
My initiation on the subject was when I made the serialization library.
The library was pretty popular so the documentation was subjected to
much criticism and scrutiny. I had a really, really hard time figuring
out how to do this. It was the subject of a long, contentious, threads
about how to do it. I was subjected to the didactic style of boosts
legendary "Mr. Personality" David Abrahams who sadly has moved to the
dark side (working for Apple) to support his family. Without going into
more details, you'll find it more pleasant to get suggestions from me.
Basically I'm come down to a "formula" - almost a form based model.
This I've detailed in the boost library incubator website under
"requirements/documentation" and under "simple tools".
I've banging on this particular drum for over a year now. I ranted on
the subject at C++Now last year. I'm very gratified to find that it
might be having some effect. I see submissions to the boost incubator
are improved in this area. I see reviewers are more carefully
scrutinizing documentation of submitted libraries and are insisting on
improvements and that many of the criticisms seems better focused.
I'd still like to do more here. One thing I'd still like to do is
reconcile DOxygen with the documentation requirements as described in
the incubator. I failed on my first attempt to do this, but now I'm
thinking that it might be possible to do after all.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk