499444 - Figure out the right name dictionary for spell checker

Status: RESOLVED WORKSFORME

(Core :: Spelling checker, defect)

Description

[Matt Caywood]

Currently the Firefox en-US dictionary is carrying around a large number of first names, many in obscure or variant spellings, e.g. 

Brittaney/M
Brittani/M
Brittany/SM
Britte/M
Britteny/M
Brittne/M
Brittney/M
Brittni/M

The last names in the dictionary are mostly well-known or common and seem less problematic.

We should figure out 

(1) where these names came from -- it's not clear to me when this diverged from MySpell/Hunspell trunk, and it would be easier to maintain the dictionary if we were closer to trunk.

(2) whether this is the name list we want to use, or whether we should use an updated version of the SCOWL name dictionary (taken from census data, apparently: http://wordlist.sourceforge.net/scowl-readme). There are also huge databases of more recent names in e.g. NameVoyager, but that is probably not OK for licensing

I think we probably want to be more permissive than the standard Hunspell dictionary, because people are using Firefox spell check for email and other tasks involving large numbers of personal names. 

If there is an upstream provider we can adopt, that would certainly simplify the situation, and it could be used as another source list in Bug 479334.

Comment 1

[Kevin Atkinson]

These are likely form the Moby names lists, which I effectively removed after SCOWL 6 since they contain too many errors and rare names.  I recommended you drop them.  SCOWL already includes most common names.  Thus, I recommended you look at what already included.  It may be missing some newish, yet now common, brand names, however.  I welcome suggestions for names to add, but if it a name relating to technology I need to be able to at least recognize it.

Upstream Maintainer.

Comment 2

[Kagami Rosylight [:saschanaz] (they/them)]

Britte/Britteny/Brittne/Brittni are now rejected, so I guess some improvement happened here.