I would always side with accuracy over tradition when it comes to dating a set.
Use the US Caramel set as an example. It has long been listed as a 1932 issue, but it can be shown to be a 1933 issue. I realize that changing the accepted date on the set will make an awful lot of slabs inaccurate. But if new information comes to light, the "facts" as we all thought we knew them should change also.
The subsequent data bases should also be changed, and if that means double listing, I'm OK with that.
What about the old slabs? I don't care.
|