© Charles Chandler
Programmers have to work a lot with names, and parsing them can be as much of a pain as date math. To make matters worse, now people are concatenating their last names when they get married, making surnames a lot longer than they used to be. Thus "Joyce Ann Hollingsworth-Macintosh" might not fit in the number of boxes given on the paper form, or in the database. And what's going to happen when the children get married — will they combine their two (hyphenated) last names to make one containing four surnames (e.g., Rebeka Magdelan Hollingsworth-Macintosh-Frederickson-Washington, daughter of Joyce Ann Hollingsworth-Macintosh and Andrew Harold Frederickson-Washington)?
Sooner or later, we have to start dropping surnames.
The better way is for everybody to not bother changing their names when the get married, because that's just confusing. Still, the children's names should represent the union of husband & wife, getting their middle names from the mother's family, and their last names from the father's family. Thus if the name is "Rebeka Hollingsworth Washington," the mother is a Hollingsworth, and the father is a Washington. The personal name (Rebeka) then distinguishes between her and her siblings.
People might even give their children middle names after ancestral maternal surnames, to re-establish the direct mother-to-daughter lineages through all of those generations, if it is known. Right now, most women have their fathers' family surnames, which don't represent their maternal lineages at all.
If this became the standard, all of the existing code & protocols would still work. The last name would still be the father's family, which is used formally as the only name, or slightly less ambiguously, along with a first & middle initial (e.g., "R. H. Washington"). To be more verbose, we would spell out both the maternal & paternal names (e.g., "R. Hollingsworth Washington"). Informally, just the first name is used. If that's ambiguous, we should use the first name, along with the middle & last initials (e.g., "Rebeka H. W."). Or spell out the middle name (e.g., Rebeka Hollingsworth), thereby putting more emphasis on the maternal name in informal usages.
Rebeka H. W.
Rebeka Hollingsworth W.
Rebeka Hollingsworth Washington
R. Hollingsworth Washington
R. H. Washington
If just one name is given, it's either a personal name, or the father's family name, which might be ambiguous. Likewise, if just two elements are given, it might be the personal name and the mother's family name, or that and the father's. But if three elements are specified, they are definitely the first, maternal, & paternal names. So any database that supports name searches has to require all three names (either spelled out, or just the initial).
For a universal name with a lot of meaning, uniqueness could be guaranteed with a 19-digit numeral, in 4 groups:
||mother's family name
||father's family name
|year of birth
||year of birth
||month/day of birth (3 for week/day)
||location of birth, hundredths of a degree (−90.00° to +90.00°)
||location of birth, hundredths of a degree (−180.00° to +180.00°)
So her universal name would be "Rebeka Hollingsworth Washington 1985-0721-3945-7538". For the sake of privacy, she might never use her universal name, and this might appear only on the most formal documents, such as her birth certificate. But it will uniquely identify her. Specifying the lat/lon to hundredths of a degree gets it to within 1 km. Then people just have to make sure that no two babies with the same first, middle, and last names are born on the same day in any given square kilometer, and the names will all be universally unique. Specifying all of this will be useful, since knowing the date of birth makes it so much easier to locate archives, and knowing the location of birth means that you know where to look for records.
It would also be nice to support a secondary personal name, which would be chosen by the individual.