After giving it a long thinking through, I decided that I would give the side/face cutter a go at it, then if that didn't work, I would take the difficult route.
First the problem.
I don't have enough movement on the x axis to do the job in one hit, so two shots will need to be done. But because I am working off the feet bottoms as a datum, if I do a good cleanup between operations and make sure everything is clean, I should be able to keep within my self imposed tolerance of 0.001".
I have decided to go with a slightly smaller cutter to help keep everything a little more rigid.
I also decided, because the rear side had up to 0.011" variation, to actually put cut on of 0.015" (0.4mm) and hope the area can be cleaned up in one quick swipe.
Once the cut was set for height, both z axis' were locked up rigid, and the DRO set to zero. From now until the end of the job, all areas will be cut at this setting.
This pic shows the start of the exercise, and to my great delight, the underside wasn't hard skinned, and gave a nice steady, smooth cut.

A shot from the top on the steady creep along the casting.

And this is what it looks like from the underside whilst cutting, notice how nice the surface is after being given the treatment by the cutter.

You can just about make out the spindle speed, I had it running at 200 RPM. The power feed was set very low and took 25 minutes to travel the half casting distance, overall it took 1 hour to complete both cuts, that included the casting end change.
I could most probably have got away with a much higher speed and feed, but I didn't want to take any chances. You only get one go at it.
I just sat back with a fag and a cup of coffee, while the machine got on with the job.

Here is the casting swapped over for the second cut. I set the casting straight longtitudinally by eyeballing the back edge and lining it up with one of the t-slots. That should get it within a couple of thou of being true, and that is close enough for this job, as I do not need to take anything off using the outer edge of the cutter.

The finished run, and I have put on an arrow to show the end of the old cut and the start of the new.

Here is a close up of the join. It looks rather mismatched, but after measuring all over the area, no detectable mismatch could be found, so no problems on that score.
After checking the whole run length I was delighted to find that my 0.001" tolerance was well within limits.
Punch the air, kiss the dog and after getting all the loose hair out of my mouth, time for a fag and another coffee.

I restarted just after lunch, and set up the rear side of the casting, and without changing the height of the cut, the next two cuts were carried out just like the first side. Now starting to sweat, to see if my measurements and calculations were correct, as this was the side with the biggest measured sag error, 0.011".

I needn't have worried, it turned out exactly the same as the other side, except for the very extreme left hand end (1/2"), that had dropped to within 0.002" tolerance. That was accepted by me because that was the thinnest end of all the measurements. In fact, if I had put 1 more thou on the cut, that would also have been within my limits. But it doesn't matter anyway, the saddle should never be needed to go that far back.

I did do a slight deviation on the fourth cut. Under where the head sits there was a machined area, to enable a stable flat area for the clamp down bolts. It meant that I had to feed the cutter in to that depth before carrying on to the end.

So that is the casting machined up to a good standard, with an added bonus. When I took the clamps off after the final machining run, there was no rock on the casting. Whether it stays that way will only be known in the future.
But, this was only the first part of the job.
As I was discussing this machining exercise with Darren, an article appeared in Model Engineers Workshop, showing how to make and fit tapered jib strips for this very machine. So that will be the topic for the next bit of this post.
I can now stop sweating, and maybe get a bit of sleep tonight.
Bogs