Jump drive works by opening a wormhole connecting the start point's frame of reference with the end point's frame of reference, then projecting a ship (cocooned in a bubble of normal spacetime) through it. That wormhole is a quantum incursion into jumpspace, so although the ship exists in its own bubble of normal spacetime and thus normal physics apply to it, the wormhole itself is governed by the physics of jumpspace. The shortest temporal event possible is 1 chronon, which in our spacetime is 2x10^-23 seconds but in jumpspace is 6x10^5 seconds (about 7 days). [Note that a chronon is not the same as Planck time, an abstract concept used in physics.]
The end point of a jump wormhole is displaced by X number of parsecs spatially and by approximately 1 week temporally. There are three temporal interfaces with jumpspace: between normal spacetime at the wormhole entrance, between the normal spacetime in the jump bubble, and between normal spacetime at the wormhole exit. Since when jump is initiated the wormhole is created, the wormhole entrance interface can be assumed to be locked in step. (This may not be the case in catastrophic misjumps.) This leaves the other two interfaces some degree of variability. Because misalignments cause turbulence and thus drain energy from the system there is a tendency for these other two interfaces to be closely synchronised. However, even in a non-misjump situation, it is possible for the end-point to vary by upto +/- a day, and the duration for the travelling ship to vary independently by upto +/- a day. The degree of variance in either is a factor in how violent the exit from jumpspace is.
Since there are 2 disparate frames of reference used the issue of a world moving due to a slight delay in exit doesn't exist. Also, since the exit point has a frame of reference based on the nearest large mass (usually a planet) an emerging ship has a similar momentum, in terms of direction and magnitude, as that large mass.
As the jump wormhole doesn't exist in our universe except at the ends no jump masking is applied except at the ends: an end point's initial position is translated directly away from any offending mass (both spatially and temporally) until the 100d limit is reached (leading to violent exit); the start point's position is distorted (increasing the chance of a misjump).
The 100d rule comes from the 'stressing' of spacetime by mass objects. An analysis of this stressing can be found here. The underlying assumption is that this 'stressing' is against a general galactic background that changes little across 'known space'. However, the closer you get to the galactic core you will find you need to get further and further away from a planet before you can jump because the galactic background gravity is more. In fact, the central stars of the galaxy are unreachable by jump for this reason. Meanwhile, travelling a significant distance away from the galactic core has the opposite consequence. Put another way: while confined to the relatively small region of known space the effects of the galaxy on jump has already been factored in.
The membrane between the normal spacetime bubble and the jumpspace wormhole is porous and allows both heat dissipation and mass transfer. However, the physical properties of an object in jumpspace are described by complex numbers rather than real numbers. This means that conventional physics is inadequate to describe interactions between such objects. Fortunately, nothing has been found in jumpspace other than gas and dust equivalents. Even time works in unusual ways: in addition to a different value for a chronon generally entropy is vastly accelerated ... but local temporal stasis pockets can readily form.
The other issue raised relates to computing power. IIRC, Traveller scientists admit they do not fully understand jump: jumps greater than 6 parsecs are theoretically impossible yet do occasionally occur in misjumps. So the formulae used to calculate jumps must themselves be imperfect approximations of the actual mechanics. In other words, no matter how powerful your computers, the results they compute will never be exact.
A misjump is a malformed wormhole ... the exit point may distorted into a new point in space and time, the interior physics of the wormhole may be corrupted resulting in a ship getting ripped apart, there may not be an exit point at all (the jumping ship travels forever), the membrane my dissolve and the contents of the bubble subjected to jumpspace's incompatible physics (temporally fragmented and non-linear).