How do you handle surnames in Gramps, especially when you do have many variants which also change over generations. Currently I do add all variants as alternative names and this is quite time consuming, but needed for filtering. I wondered if someone has found a more efficient way.
I only add alternate name records when there is evidence that the person was listed, or known to use, the variation. I am not concerned (or have not yet been concerned) about filters because for the most part I tend not to use “names” as a filtering criteria. Most filters I use are “Ancestors of…” or “Descendants of…” filters which are not name dependent.
I will use the Group As feature to bring the various spellings together in grouped lists. As an example my surname is often seen absent the “i” so for both surnames “Scheipers” and “Schepers” they each get grouped together as “Scheipers (Schepers)”. @emyoulation was kind enough to put this into the wiki.
2 Likes
I also try to avoid name variations as much as possible. Most people where probably illiterate farmers, so only the cleric wrote down their name in churchbook records. Unfortunately this happened always in the way the name sounds, so there are many variations since there was no standardised spelling.
Many people of my database lived in the same area, so when entering new person I have to check if they are already in my database to avoid doubles. If I only add one variation some people don’t show up in search/filter results.
Unfortunately this feature is only useful for person tree view and does not solve the problem of people not showing up in search results e.g. if you group Schepers/Scheipers and don’t add the alternative names you don’t find people named “Schepers” if you search for “Scheipers”.
A possible solutions I came up would be using filtering for names using regex. I already use that for the more common ones which I search a lot, but I would probably need 50+ custom filters for the names in my database if I stop using alternative names. 
1 Like
I have a large database (that I am slowly cleaning) so I too will always check to see if I had already added someone. And this is done using Grouped People. But for this to be practical, the sort order needs to display people as “Given Surname, Suffix”. This puts “Petrus Scheipers” right next to “Petrus Schepers”. It does create some funky sorts… “Petrus J Scheipers” will sort before “Petrus Scheipers”.
But to complete the picture, I have my default display name as “Title Given Surname, Suffix”. For those people with no “Title”, their sort is as I want them. For those that I have their preferred name with a Title, I actually override the default sort; “Given Surname, Suffix, Title”… If I do not do this, the Title sorts before the name taking the person out of order.
For those wanting/needing more information… see Gramps’ Name Editor.
I have not used the “Use regular expressions” often. And never for portions of a name so cannot offer any advice.
In case others are following the thread, here is Gramps’ Regular Expression wiki page.
I usually filter persons by their (sur)name and sort the list by birth date. That avoids funky sorts, but it requires that every person has a birth date. For people without birth record I just use the “estimated” date qualifier.
I would expect that, sooner or later, we’ll start automatically generating & storing SoundEx, Metaphone or Double Metaphone equivalents as an attribute for names. Then filter on those Attributes to make a fast handler for phonetically similar names. It is too slow to continue generating them on the fly.
There is currently a SoundEx Gramplet and the Find Possible Duplicate People tool has a SoundEx option
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.