Joseph Wright writes:
 > On 05/01/2011 18:42, Frank Mittelbach wrote:
 > > Joseph did a cleaner reimplementation of that recently. This new
 > > implementation is working very nicely and it is available from the svn for
 > > people who want to play with it.
 > 
 > I have updated a snapshot to CTAN, as part of xpackages. You can 
 > therefore download xpackages.tds.zip for testing [once Robin has 
 > installed it, of course :-)].

ah, Joseph already announced this ... me too late as usual.

 > > Basically what xcoffins currently does is to provide a fairly natural way
 > > for specifying how boxes (called coffins for some reason) are to be
 > > aligned to each other including rotation and with Joseph's implementation
 > > also scaling.
 > 
 > Scaling here means 'stretching the content of the box', i.e. 
 > graphic(s/x)'s \scalebox and \resizebox.

I might add that the stretching and scaling is a part that I'm least convinced
off. Not that it is wrong or should be taken out, but that I doubt that this
particular functionality is often needed. There are use cases for it, I
guess. But in many cases a designer wouldn't want to loose control over font
sizes which will be the result of such type of scaling.

There is another type of scaling/sizing which stretches the inner glue of a
box and which isn't implemented. Again I don't think that there are many
applications for it, but I would expect equal number of use cases for this
type too. So perhaps we should consider this as well.

frank