I looked at my mill and thought I would add an idea to the discussion. The trick to making it work will be knowing exactly how far one revolution of the handle down feeds the quill (on inch machines there is a gap in the gradations since one revolution doesn't equal anything practical in inch-land).
On the non-operator side of the head where the quill return spring lives, you could attach something to capture rotary motion in place of the retaining screw washer. Perhaps the rotary encoder Ross suggested above(I think this is how the Fehlmann Picomax Z-axis readout works). But I am thinking something cheaper, jankier and something that is completely reversible.
The outline of the idea is you mount a pulley on the quill return spring housing in lieu of the washer. The OD of the pulley is sized to correspond to length quill extension per revolution. A cable wraps around the pulley as the quill is extended. The actual measurement device is attached at the front of the y-axis stop slot. You would use a cheap Bridgeport style quill digital scale, the ends of the scale would be mounted onto your bracket. The travelling part of the scale would be mounted to a bushing that travels on a rod that is mounted parallel to the scale. The bushing would have a hole that the cable would go through and a screw perpendicular to that hole to lock the bushing travel to that of the cable. After going through the bushing, the cable would go over a pulley at the top of the bracket and on the back side of the bracket you would have a spring to keep the cable tensioned throughout its travel.
Now for the fine points. I am thinking the pulley should have a helical groove so it doesn't wrap over itself and change the diameter (this would also be a good project for the spiral milling attachment). By adding an additional pulley at the base of the measurement device, you can orient your scale vertically so it is easier to read. Possibly you could also make the pulley diameter 2x the quill travel, have a pulley attached to the bushing and then have the cable go to the tension spring.
I am sure there are problems with my idea, but this is a interesting problem unique to Deckel mills. Maybe my idea will inspire an even better one!
Regards,
Nigel