fix(person): Improvements to Dutch name generation, in particular regarding affixes

[amarillion]

Addresses #1777, plus some minor improvements to Dutch name generation.
I've based this on PR #1637 (fullname name patterns) assuming that will be approved as well.

[codecov]

Codecov Report

Merging #1778 (aa41222) into next (d5b78c7) will decrease coverage by 0.02%.
The diff coverage is 100.00%.

Additional details and impacted files

@@            Coverage Diff             @@
##             next    #1778      +/-   ##
==========================================
- Coverage   99.64%   99.63%   -0.02%     
==========================================
  Files        2347     2346       -1     
  Lines      235657   235735      +78     
  Branches     1145     1142       -3     
==========================================
+ Hits       234811   234863      +52     
- Misses        824      850      +26     
  Partials       22       22              

Impacted Files	Coverage Δ
src/locales/nl/person/index.ts	100.00% <ø> (ø)
src/locales/nl/person/female_prefix.ts	100.00% <100.00%> (ø)
src/locales/nl/person/last_name.ts	100.00% <100.00%> (ø)
src/locales/nl/person/male_prefix.ts	100.00% <100.00%> (ø)
src/locales/nl/person/suffix.ts	100.00% <100.00%> (ø)
src/modules/internet/user-agent.ts	89.34% <0.00%> (-7.11%)	⬇️
src/modules/location/index.ts	98.70% <0.00%> (-0.22%)	⬇️

[ST-DDT]

#1637 has been merged. Please rebase.

[amarillion]

Rebased onto next

[import-brain]

Please put the last name list in alphabetical order

[amarillion]

I've sorted the list of names ASCIIbetically. NB this is not the Dutch sorting order.

[import-brain]

I've sorted the list of names ASCIIbetically. NB this is not the Dutch sorting order.

Thank you! @faker-js/maintainers Should we sort the names ASCIIbetically or the Dutch way?

[amarillion]

Oh, and I misspoke - it's not ASCIIbetical (because that would imply case-sensitive sorting - Z before a). It's currently sorted with regular English case-insensitive sorting.

So, in a nice example of "be careful what you wish for" we have the following sorting options:

1. case sensitive sorting (ASCIIbetical): Kok < Visser < de Leeuw
2. case insensitive sorting (English): de Leeuw < Kok < Visser
3. case insensitive sorting (Dutch): Kok < de Leeuw < Visser

The one that is implemented with the latest commit is #2. So, which sorting method should I use? Personally, I don't mind either way.

[ST-DDT]

One of the reasons why we want the sorted lists is to aid in the review process to detect similar or identical entries faster.
So I think a case intensive sort is definitely required.
For the start, we should use (2) English locale. We just have to make sure that (our tests/scripts) this also works for non-English users.
@amarillion The "Dutch case insensitive sorting" applies only to names or is the alphabet itself sorted differently from the English one? Everything that is not easily applicable to the entire locale, is a non-goal.

[amarillion]

I've added two more sources for the family names.

Since the purpose of sorting is detecting duplicates, I think the current English sorting is fine and we can leave it at that.

list is now sorted

[ST-DDT]

I've added two more sources for the family names.

Well, you added more links, but you didn't add any new names, so it is basically still the same list from the first source in the list.
(And thus doesn't solve the legal implications)

ff4dee0 (#1778)

[amarillion]

I don't get what you mean by legal implications. It's quite obvious that wikipedia doesn't own this list, and that the original source is in fact from https://www.meertens.knaw.nl/nfb/documenten/top100.pdf, as stated in the article.

A list of "top 100 most occurring family names in 2007" are facts, and facts are in fact not copyrightable.

[amarillion]

However, if it helps I've made a small modification to the list so it's no longer exactly the same as the source.

[ST-DDT]

A list of "top 100 most occurring family names in 2007" are facts, and facts are in fact not copyrightable.

IANAL: The facts itself might not be, but the work to compile it may be or at least, some claim it to be: https://blog.polco.us/copyright-apply-survey

Thanks for your efford to address this, but 1% differnce might not be enough.

[matthewmayer]

Maybe go back to the original nl last_name list and just make some changes to add some more tussenvoegsel names, using your own knowledge/judgement rather than relying on a single source?

[amarillion]

All right, I made a major modification to the list. It's now a fully original compilation of names, a random selection of various sources created by me personally. Any resemblance with any other list you might find is purely coincidental, of course lists of common names are bound to have overlap.

Maybe go back to the original nl last_name list and just make some changes to add some more tussenvoegsel names, using your own knowledge/judgement rather than relying on a single source?

Funnily enough, after adding the appropriate tussenvoegsels, nearly every name on the original list is also in the top 100 list. I only detected two differences: 'Klein', which I added, and 'Stichting', which isn't a family name.

[amarillion]

Nice work guys, I appreciate your attention to detail!