[r6rs-discuss] library versions (was: Rationale issues)

From: Matthew Flatt <mflatt>
Date: Fri, 29 Jun 2007 06:20:22 +0800

At Thu, 28 Jun 2007 18:08:11 -0400, William D Clinger wrote:
> Matthew, not speaking for the editors, wrote:
>
> > But when it comes to distributing a library to users who don't have
> > access to your SVN repository, or who are combining libraries from many
> > different sources (managed in different repositories), then it makes
> > sense to put a version number at packaging time, at least on modules
> > that are intended for external use.
>
> This makes sense if you are the only one who is
> distributing a library of that name, but it does
> not make sense when several different vendors are
> distributing incompatible libraries that have the
> same name *and* version number.
>
> By requiring implementors of Scheme to distribute
> incompatible libraries that have exactly the same
> name and version number, e.g. (rnrs base (6)), the
> recent drafts have given us an example of how the
> proposed versioning facility should *not* be used.

I don't see it that way. To me, the 6 version in the case of `rnrs'
corresponds to its specification. It's not the same use of versions as
some other possible uses that also fit into the system, but I don't see
it as a misuse.

The version number may turn out to be useful if the specification of
version 7 merely extends version 6. Or it may turn out that it's better
to use a different name for a future specification of the library.


By my estimate, we lose little if we have versions on `rnrs' libraries
and they turn out not to be useful. We lose much more if we leave out
library versioning altogether and (as my past experience suggests) it
turns out to be useful.

As someone looking for a compromise that lets us proceed, that seems
like the right trade-off to me. Others will estimate the costs and
benefits differently, and arrive at different conclusions.

Matthew
Received on Thu Jun 28 2007 - 18:20:22 UTC