I do also take corrections and log everything in XDrip+, which provides an approximation of the remaining inulin on board (IoB.) I take this IoB value into account when I need to bolus for food (or correct again, I happily stack corrections as I can see how much is remaining on board.)
Like
@Antje77 my bolusing is not an exact science (even though I am a research engineer), as it depends on current rate of change, what the food stuff is, how much exercise I've been doing and will do, etc., so very much experience based. I also habitually split my bolus doses for larger meals in order to avoid post-meal hypos, so the splitting and correcting merge together.
With that said, it's interesting with the CGM to see how BG responds - mine tends to produce similar behaviours over multiple day periods, probably longer, I should look at the data so it's not just from memory. E.g. sometimes my BG changes will be very spikey, constant upward gradient after eating + bolusing, then an immediate mirror downward gradient after 30min, 1h, etc - if I correct too early/aggressively in this case I quite often end up with the the correction becoming active just as I'm dropping anyway, other times, as you mention, BG can remain persistently high and I'll keep correcting with no apparent effect, and sometimes I get nice smooth gradient changes in response to food and bolus which means I can easily work out whether a correction might be needed. The problem, of course, is that it's hard to know which behaviour you'll see (or rather I know what is has been like but I'm never overly confident that it won't suddenly change and I'll be caught out.)
Constant intervention is both tiring and I think probably produces unwanted oscillations (where one then needs to correct a low by eating, then ends up high again, etc) so I can see the reasoning behind the advice to not bother correcting if a meal is on the near horizon. One half-way house approach I sometimes take if I really don't want to run low before eating (e.g. to not spoil my appetite, or because I will be travelling to go out for a meal, etc) is to correct but with say half (or less) of the correction dose - then at least your BG will start heading the right way and you likely have some IoB for the meal in question, which is one way to improve the effectiveness of the subsequent bolus.