Boost logo

Boost :

From: David Abrahams (david.abrahams_at_[hidden])
Date: 2002-04-11 16:47:34

----- Original Message -----
From: "Braden McDaniel" <braden.n.mcdaniel_at_[hidden]>
Newsgroups: gmane.comp.lib.boost.devel
To: <boost_at_[hidden]>
Sent: Thursday, April 11, 2002 4:30 PM
Subject: [boost] Re: Re: WWW-site: proposed patchset for autotools as

> On Thu, 11 Apr 2002 16:28:38 -0400, David Abrahams wrote:
> > ----- Original Message -----
> > From: "Braden McDaniel" <braden.n.mcdaniel_at_[hidden]>
> >
> >
> >> As I see it, the
> >> difference to users is simply that an autotools build offers a
> >> lacking in the Jam build: deployment. This is practical with the
> >> autotools because for POSIX systems, there are better understood
> >> notions of how software should be installed; and as a result of
> >> the autotools provide support for deploying software per those
> >> conventions. But deployment is a perfectly reasonable feature to
> >> regardless of whether you're developing software or just installing
> > it.
> >
> > I agree with you on that point. However, there's an additional
> > that we haven't decided what deployment should mean for boost. It's
> > unlike other software deployed for POSIX in several non-trivial
> > respects,
> Such as?
> As you might guess, I don't agree at all. There is a tremendous
> of software available for POSIX systems; in the history of the
> a whole lot of different requirements have been addressed. I don't see
> what's novel about Boost in this respect.

I confess that I'm only trivially familiar with the domain, but these
things seem different from most POSIX packages:

1. It actually consists of a wide variety of mostly-independent library
2. Little or no attention is paid to binary compatibility between
3. Strictly speaking, source compatibility is broken frequently

I'd love it if there was precedent for all of this, and we could just
adopt well-known practices. Is there? Can we?


Boost list run by bdawes at, gregod at, cpdaniel at, john at