On 16/09/2009, at 5:32 PM, Joseph Wright wrote:

> I've made the changes outlined above. At present, I've retained .set:N
> and added
> - .dim_(g)set:N
> - .int_(g)set:N
> - .skip_(g)set:N
> - .tl_(g)set:N
> - .tl_(g)set_x:N
> Is everyone OK with retaining .set:N in this way, or would others  
> prefer
> only to have the explicit setting properties? I'd like to update the
> CTAN version soon to reflect these changes (which I hope are the final
> ones for l3keys), and so want to get this right.

If you're unsure about .set:N, perhaps it's better to drop it for now.  
I mean, *I* like it, but others have had mixed reactions. (You've been  
looking at l3keys recently, Manuel? Any thoughts?)

Not sure if this helps you to decide, though :)