This is an excellent question, and, I think, not fully answered by the replies received so far.
Originally Posted by mat
Craigb's reply is the received wisdom. Assuming you have a well-set-up table saw, cut to width +0.5mm then a 0.5mm pass on the jointer will do the trick. The jointer is not straightening the edge only replacing table saw residual marks with smoother, jointer blade marks.
The longer answer is this:
1) For many applications it doesn't matter. Glueing up table tops for example, the eye won't be able to tell if the individual boards are very slightly off parallel in width. Ditto for the glued-up table top.
2) Where the last cut isn't on a well-tuned TS (e.g. on a bandsaw, requiring more material to be taken off and/or some jointer straightening) and/or the jointer table lengths are short compared to the length of board, then there can be no guarantee of parallelism. (My case: I don't have a TS, and I only have a small 6" jointer).
If I particularly need the parallelism and exact width I run the last pass through the thicknesser. I run multiple boards at one time holding them against each other to inhibit twisting under feed-roller pressure. Usually this is where I am making a set of cabinets and drawers where I need lots of boards where I want to be confident any board I pick up will be the exact right width -- but in such cases the boards are only 65mm wide or so and they fit through the thicknesser.
For wider boards or glued-up panels or frame-and-panel doors that won't go through the thicknesser in vertical orientation (thankfully not real often) I hand plane to parallel and if necessary then do a final one pass over the jointer to ensure the edge is 90 deg to the faces (the jointer can make 90 deg faces easier than I can handplane faces at exactly 90 deg.) I have also been known to use the router with the fence on the wrong side but having got into trouble more than once with this technique it isn't something I like to mention in a forum such as this !
P.S. I know these are cumbersome work-arounds that would be solved with a good TS ... BUT .... it isn't the cost of the TS that is the problem with my shop, it is the cost of a new shop big enough to house the TS in!
All short sentences in economics are wrong.