Re: [boost] Review: Boost Coding Guidelines

From: Greg Colvin (gcolvin_at_[hidden])
Date: 2001-07-20 12:45:06

I think this is a very reasonable approach.

From: Beman Dawes <bdawes_at_[hidden]>
> At 12:40 PM 7/20/2001, Darin Adler wrote:
> 
>  >On Friday, July 20, 2001, at 09:18  AM, David Abrahams wrote:
>  >
>  >> From: "Beman Dawes" <bdawes_at_[hidden]>
>  >>
>  >>> I think that Boost Coding Guidelines should be accepted, but subject 
> to
>  >>> refactoring into separate "Requirements" and "Guidelines" documents. 
> The
>  >>> refactoring should include material from the current "Library
>  >>> Requirements
>  >>> and Guidelines" page, http://www.boost.org/more/lib_guide.htm, and
>  >>> perhaps
>  >>> one or two other current pages.
>  >>
>  >> How will we decide which are requirements and which are guidelines?
>  >
>  >My two cents:
>  >
>  >By default make everything a guideline.
> 
> That's an interesting approach.  But there should be at least one 
> exception:  Anything which is already a requirement in 
> http://www.boost.org/more/lib_guide.htm stays a requirement.  There are 
> only four coding requirements currently, all having to do with portability.
> 
> It should be possible to promote a few or the guidelines to be requirements 
> if they are non-controversial.  For example, I'd like to see "ISO Standard 
> C++" elevated to a requirement, although even for that we would want to add 
> a caveat allowing non-standard extensions be used in alternate 
> implementations.
> 
>  > Convert the entire document into
>  >the guidelines document. Then as time allows, promote things to be
>  >requirements. You and Beman can decide whether formal review is required
>  >or not for your initial requirements documents with your best judgement
>  >and depending on how many requirements you end up with and whether they
>  >are things that were controversial in the guidelines review process.
>  >
>  >The requirements will presumably be a small set compared with the
>  >extensive set of guidelines. We can always downgrade a requirement to a
>  >guideline after the fact, if we decide it was a mistake.
>  >
>  >The issues are not the same as with code, where making a change after the 
> 
>  >library is accepted can easily result in incompatibilities for people
>  >upgrading to a new version of the Boost libraries.
> 
> --Beman
> 
> 
> Info: http://www.boost.org  Unsubscribe: <mailto:boost-unsubscribe_at_[hidden]> 
> 
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ 
> 
> 
>