This door is not the 'roll up' type. It is mounted in its own frame and when the RSD is opened, the frame (with the door) can be opened on its [the frame's] own hinges.
I thought about a Ring Latch but that brings me back to the situation where two people are approaching the door from opposite sides. One swiftly turns the Ring whilst the other person is reaching for the Ring on their side and experiences an 'Ouch!' moment. Not exactly an 'injury', but enough to leave the fingers tingling for a few minutes.
Let me try and explain it in the context of a domestic internal door with Lever handles. Two people approach from opposite sides. One person is about to push down the handle when it suddenly descends, having been pushed down momentarily before by the person the other side [of the door]. The first person, who 'thought' they were about to push down on the handle, just 'pushes down' on fresh air. No conflict of function and no [obvious] chance of injury - unless one gets the door opened into their face

- but that's a different matter.
What I am looking at to achieve a similar objective, but in a vertical plane [rather than rotational] so that if one person, on one side, bends their finger to lift the 'lever' [as per the image] then that lever 'knob' extends [through a vertical slot] to the other side whereby, should there be someone there carrying out the same action, their movements are not in conflict.
The door already has a Barrel Bolt, which will be retained as it has nibs and padlock holes. The problem with a[nother] Barrel Bolt, even if spring loaded, would be that there would be a reliance, on any person passing through the door, to then re-engage the Bolt whereas, with a Suffolk/Norfolk type latch [modified], once properly set, just swinging it closed engages the latch.
Thanks all.