I'm not sure how useful a critique of the GEDCOM format in this area is, but there are some oddities. Consider the (lightly twisted) name and parts thereof:
1 NAME James Tiberius /de Kirke/
2 NPFX Captain
2 SPFX de
2 NSFX III
Note that only
some of the NAME has been further analysed into components. There's no given name and no surname. So far as I can see from the GEDCOM 5.5.1 soecification, this is entirely legal. But it might make a designer of a consistency check think a bit.
Secondly, and I'm not sure this is a problem, note that the values given by NPFX and NSFX (prefix and suffix to the name as a whole) do
not appear in the NAME, but the SPFX (prefix to the suname)
does appear in the NAME. It may not be a problem but it does feel a little inconsistent. Why not have:
1 NAME Captain James Tiberius /de Kirke/ III
2 NPFX Captain
2 SPFX de
2 NSFX III
Why not? Well, one good reason is that it is "significantly in conflict with concepts already in use" as David puts it.
By the way, the 5.5.1 says that if you have a list of Given Names for one person, they should be separated by commas. What?
Why? Indeed when it comes to NAME_PIECE_SURNAME_PREFIX then it says
Different surname articles are separated by a comma, for example in the name "de la Cruz", this value would be "de, la"
Tamura Jones in the Annotated 5.5.1 is absolutely against this and says:
FamilySearch's instruction to insert commas in prefixes is not just unmotivated, it is wrong.
You should never modify names. ...
Best practice
Record the prefix as is to ensure it will transfer unmodified.
Modifying the prefix by inserting one or more commas is almost sure to create problems. Most
genealogy applications do not expect commas in any name part.
While Tamura is not the arbiter of the standard, I think that he has a point. Why
any commas that aren't in the actual name?
By the way Name Prefix and Name Suffix (abbreviating their names) are described as non-indexing but NAME_PIECE_SURNAME_PREFIX doesn't mention non-indexing, so it's not clear whether FamilySearch intend the Surname Prefix to affect "indexing" or not. I can't actually see any use for it if it doesn't.
I'm also not mad keen on this vague term "indexing". I fear circularity - the GEDCOM defines the data in terms of "indexing", which is a function of the software, but the software only does it because it's defined in the data definition? Fortunately, we can refer to real world "indexing" and collating / sorting such as is found in book indexes, directories and library catalogues. I'd much prefer a real-world explanation at this point. It then directs someone to consider
whether Vincent should be sorted in with "van Gogh" or with "Gogh" in their world-experience.
Apologies if you don't want to read this but I think it's worth thinking through - if you like that sort of thing.