proposed and wrote a physics-based jumpjet system a while ago. I was against it because it did not seem to reliable enough to hold up half a factory. He repurposed the system to implement amph floating physics, where it works because floating has less room for error than jumping. The impulse jumpjet gadget still exists and even has a disabled modoption. The gadget probably needs a few API updates, but I would expect it to work otherwise.
My issue with impulse jumpjets is the jank. For jumping units to be at all balancable they need to function fairly reliably. An impulse jump could fail if:
The unit takes impulse from any source (Ripper etc) and veers off course.
A clump of jumping units are told to jump together, and they bump into each other en route.
A jumping units lands on something, and bumps off it into the abyss.
A jumping unit clips the terrain or a map feature and bumps off course.
A 0.1% chance of your Pyros falling down a cliff or into a hole would be far too high, and with all of these factors the chance has to be higher than that. If JJ is to be a reasonabley startable factory then Pyro/Constable/Jack need to not be randomly lost or delayed by a bad jump. While there are ways to lose units, most of them feel more like the player's fault, and existing rare frustrating moments doesn't justify the addition of new ones.
In short, I value Pyro reaching its destination 100% of the time over a bit of extra physics. I don's see much value in the additional physics either, as the physics would mainly add a bunch of counters to jumping - counters it doesn't need.
Lobster is more a support unit than a core part of the factory, and tends to appear later in the game. People don't tend to fire multiple Lobsters at the same thing from different angles, as is common with jumpjets, and I think there is more of an understanding that Lobsters are a bit of a gamble.