I don't think that anyone would disagree about the date/time handling in LiveCode needing to be updated.
I do think that the majority of formats you encounter in IT systems are probably covered by 'the internet date' and 'the standard date' (I don't really see the utility in using ISO8601 or RFC2822 to identify them - unless one just wants to try and sound clever, arcane and require people to have to google those terms when reading your code... 'internet' and 'standard' indicate well enough what they mean, and a reference in the docs would be more than sufficient as far as I can see). (Indeed a lot of the variance you see in renderings of date/time are probably down to variations allowed in ISO8601).
I'm all for allowing choice so a customizable date format would seem to be a good option to add. I'd perhaps suggest that http://userguide.icu-project.org/formatparse/datetime
is a more appropriate reference for a formatting string as it is a little more complete in terms of i18n. Also, we do need to look into taking into account non-Gregorian calendars if we are to have exemplary i18n credentials. (The current dateFormat is irksome - it probably falls under the remit of something we want to change as an 'anomaly').
By the way, whilst "Every time I turn around I find another time format I need to deal with, and inevitably it isn't supported in the engine." might be an amusing way to open a feature request, I do wonder if it is a bit of an exaggeration. If we want to make sure we have good date/time handling then having a list of as many things as possible that we are trying to model is an absolute requirement. So, @FourthWorld, please put your money where your mouth is and enumerate all these time/date formats you claim you keep encountering so we can do a proper analysis of the requirements