The situation with respect to two-stroke transfer ports is like pizza night when no one has the twelve bucks. Everyone is furiously running around the house, searching coat pockets and sofa cushions for stray change, busting into piggie banks -- anything to make up the twelve.
It's especially bad in 250 because there is the specter of the Aprilia, its peak revs rising into the stratosphere beyond 13,500, where reed engines become notably enfeebled. Two years ago both 250 and 500 engines were having their port-to-port septums cut down to nothing. Now attention is being lavished at last on the exact shape of transfer ducts.
Anyone who has used a velocity probe on transfer windows of a cylinder sitting on a blower has seen that each window has its own delivery map. Its usual for the high velocity to be concentrated at the tops of the windows, and sometimes to be crowded to rear corners. Also seen is outright reverse flow down at the bottoms of windows - a region in which a kind of "roller" exists. This is the probably reason why Yamaha's flat-topped pistons have a chamfer around them, not very wide, of about 8 degrees or so. This may serve as an invitation to the flow to "come on down" and adhere to the piston, rather than let go in classic style to leave a no-flow dead-zone right over the piston crown.
Here is another point. It's all very well to shape a duct most cleverly, as four-stroke porting specialists do, but when the flow reaches the duct's end, what happens? That carefully-delivered flow suddenly expands into nothingness -- making a very poor bargain of trading in velocity against pressure. The Superflow people (makers of flowbenches and dynos) estimate that 40% of the flow loss in a four-stroke intake port is diffuser loss at the point of entering the cylinder. If that flow can be attached to at least one surface, the outflow from the port window will be under some degree of expansion control, and the process efficiency may rise.
Another tidbit from the old Superflow book is this; that the highest flow per square inch of port area that they had seen (the book is old, and the situation may have changed by now) belonged to a modified Norton twin. Flow in that case was flat -- meaning nearly at right angles to the cylinder axis, rather than fashionably at steep downdraft. Flow would then exit from the far side of the valve and attach to the inner surface of the cylinder head -- not as in more modern engines, freely jetting out into empty space. This kind of attached diffuser flow is hardly discussed, but I think it's potentially very important -- two-stroke or four.