Hi Valerie,
> Is it me, or is the mail-archive not showing all mails in a timely manner?
I have not yet seen the post by Jasper to which you referred so maybe the mail is slow.
I think I am slowly coming to see what you mean.  Just another couple of questions about F1 transformations to clarify things for myself.
> The tiles behave like objects, so you can transform them with the F1 transform
> tool, and you can snap them to other guides.
Does this mean that the guides for the base tile are visible at all times (in order to be selectable)? I was assuming that the guides for the base tile would only be visible in Tile Edit mode since they are not rendered in the final image – that is why I was concerned with overloading the interaction in Tile Edit mode.
Can you select the guides and modify them independent of the group of objects to which they are associated?  That is, can the base tile guides and the “Original“ object (using the terminology of the current Clone menu) be selected and modified independently?  For example, can I change the size of the Original object after I have associated it with a base tile without affecting the base tile?
Or, alternatively,  do you need to select the Original object first and this will cause the tile guides to be displayed with their own transformation handles (i.e. there will be two sets of transform handles visible – one for the Original object and one for the base tile).  If you are thinking of this kind of interaction, then there may be some confusion as to which “object” the keyboard accelerators are sent.
>> Even worse, take a p2 tiling and put a 180-degree rotation
>> center on its "translational" edges (I'm referring to the diagram drawn
>> by veronika(?)), what you end up with is another p2 tiling with a
>> different base tile...
I think what Jasper is referring to is the fact if you only show a minimum set of operations to create a tile group, this is not a unique representation.  As a result, changing some guide behaviour will not necessarily change the wallpaper group.  Below is an image to illustrate his point.  It would appear that changing the vertical edges of the base tile from translations to rotations has created a new kind of wallpaper group – the result on the Original object is definitely different when applied.  However, closer inspection reveals that we can choose a different base tile and the resulting wallpaper group is still p2 – that is, the wallpaper group has not changed but the base tile has.
The point of showing how the tile will be applied to an object is still relevant.  Even though it is not a unique representation, as long as it is a valid representation, the information required by the designer is conveyed.
The idea about changing the behaviour of the base tile by modifying the behaviour of the different edges or corners, however, is a bit tricky.  Only some modifications will result in new or even valid wallpaper tilings.  The allowed changes would need to be constrained to only valid ones.  Perhaps multiple representations of the same wallpaper group could be allowed – this would need to be reflected in the behaviour of the code when applying the base tile since as shown in the image above, the different representations do not yield exactly the same result.