I disagree! - we will discover nothing we did not know already. Viz: if you edit an observation there is a good chance that your pictures will vanish irrecoverably and you will have to upload them because the iSpot programmers dont care two sh!ts about fixing the errors of bad programming. (the pictures must still exist: it is surely only the links that have vanished).
We know that already. All that the tracking will do is allow one to confirm that the user made an edit, and therefore that the pictures did not just evaporate on their own: but we know that already.
I process 50-500 observations per day. There is no ways I will detect if an "edit" corrupts data from my Changes Tracker: I am only interested in New IDs and Comments. And these will be hidden in my own 50-500 changes (done by me) so I wont even know about it except by accident.
In your example above we dont even know if it was the 22 or 27 edit which lost the data (presumably the latter otherwise Tyrole would have squealed - unless she is so demoralized that she no longer cares).
What would be nice is if we can figure out why they disappear, but it appears to be random. It is not the type of edit (some are gone by editing habitat, others by editing tags and some by adding pictures). Might it be the page that one leaves from? We know that some characters cause problems (but this has allegedly been fixed). As you say, it happens to unsuspecting users, not to those of us who could intuit a possible reason based on what we did.