If your scope has parallax, I think I would be tempted to zero on 10x/pa 35, then change pa to 23 yards and shoot groups at all ranges to make sure everything is as expected.
If your scope doesn't have easily adjusted parallax - I'd be tempted to parallax at 23 and zero at 35 (again on 10x mag).
The problem is, that ultimately you want to reduce as many variables as possible and doing what you're doing you're always going to have a variable in the mixture somewhere.
For starters, if you zero on one magnification and then change to another - there's a chance that your zero will shift (depending on your scope).
If you zero on one parallax and then change it - again there's a chance your zero will change depending on scope. If you don't change parallax, there's a chance you'll get some parallax error.
So.. for those reasons, the course of action to take that would eliminate as many variables as you can would be to zero at 23 yards on a 23 yard parallax at 10x mag. ( Which is the one of the reasons I recommend doing that in the HFT Novice to Ninja articles - www.anstonftc.co.uk/hft-scope-setup-guide