Question: Match the `seq` modifier?

It seems that applyrule cannot handle variable numbers of arguments, and I cannot use something like 

applyrule(f(u::anything, v::seq(anything)) = g(v, u), [f(x, y, z), f(x, y, z, t)]);
 = 
                  [f(x, y, z), f(x, y, z, t)]

Strangely, Maple does support the identical patterns in parameter declarations: 

eval([f(x, y, z), f(x, y, z, t)], f = ((u::anything, v::seq(anything)) → g(v, u)));
 = 
                  [g(y, z, x), g(y, z, t, x)]

So the two designs do not appear coherent. Should this be regarded as a "bug" in a sense? 

Of course there is no need to use the seq modifier; here it is enough to use 

evalindets([f(x,y,z),f(x,y,z,t)],'specfunc'(anything,f),w->g(op(2..(),w),op(1,w))):

use f = MakeFunction('g(args[2 .. ], args[1])') in [f(x, y, z), f(x, y, z, t)] end:

use f = unapply('g(_rest, _w)', [_w::anything]) in [f(x, y, z), f(x, y, z, t)] end:

But the problem is, why is there such inconsistency described above?