Thanks for all the thoughtful stuff in this thread, and it's great to see considerably less than the usual unhelpful fussing and fighting between proponents of convenience and proponents of efficacy.
As regards correcting erroneous chart info, it strikes me that this needs a bit more discussion and thought.
Those who favour a paperless nav station typically promote multiple redundancy (carrying multiple devices, using multiple types of electronic charts) in consideration of the volatility of electronic data.
This is also a favoured remedy to address various shortcomings of individual datasets, and styles of portrayal. (I personally feel uneasy about this, but that's another topic.)
I think it's worth reflecting on several ways this multiple-source nod in the direction of efficacy can extract a cost in convenience, if we are not careful.
One question is this: which dataset do you correct when you find an uncharted or mis-charted danger
? And to what extent is that correction itself volatile?
A prudent mariner would correct all onboard data sources. In a perfect world each volatile data source would also be backed up an offsite location. Not to do the first of these things involves the risk that you might find yourself using the wrong dataset at some future date.
But that is starting to look like a serious amount of work, digging out all the various devices and data backups, and reacquainting oneself with how to make a permanent correction on each particular platform.
If you always use the same dataset, there is a risk attached to your resulting unfamiliarity with the alternatives, if access to that primary dataset is lost
(which is clearly a strong possibility, otherwise why would anyone carry multiple alternatives?)
I reckon it is generally better to use a poor tool you're familiar with and whose shortcomings you can work around, than a good tool you don't fully understand, and whose more esoteric capabilities you are unable to command.
If I was running a paperless navigation
station, I think (on superficial reflection) I would actually correct only one dataset, the one most commonly used, but I would also note all corrections in a durable notebook with waterproof pages. (Purists could instance the plastic nature of the pages to rebut accusations that their operation was no longer strictly 'paperless')
I would refer to this when using any but the 'master' dataset, and treat it as the durable, lifetime record
with the dark side. I would keep it with my passport, and cherish it for all time.