Please don't mix A and Ä in search results!

Question

Hello,

please, is it possible to not mix search results "Gräf" and "Graf"? My mother's ancestry has the name "Gräf", but when entering as a tag, I get thousands of entries with "Graf", which I'm not interested in at all!

Comments

A possible fix is to filter to show exact spelling as entered in search box?? A question for the tech team :)

---

I've added this to the list of things to consider when we next update the search.

---

yes, i know this point is over a year old.

if someone searches for "Graf" it is unknown if they have the capability to enter "Gräf".  if they type in "Gräf", it can be assumed they could type in "Graf" if they preferred.  it is generally true that all European/American keyboards have "A" but not all have "Ä".  the default should be to search both if "Graf" is used but search only for "Gräf" if that is used.  if the user has used "Ä" or "ä" elsewhere that could be a valid indication they can enter that.  perhaps it could be useful to collect "known Unicode entries" for each user to readily determine what their searches should do.

---

I think it would suffice if "ä" would be differentiated in the sort order, so that all "Gräf" entries would be listed together and then all the "Graf" and "Graef" entries

Answer 1

With something as difficult to define as family names over time and geographic distribution developing a search algorithm that satisfies everybody is nearly impossible. Too narrow criteria and you miss a lot and too wide you annoy those feeling to have to wade through lots of junk. Perhaps the best solution would be a toggle between narrow search and a loose search option.

Answer 2

Martin, we all deal with this to some degree - for me it is Poole, Pool, Pole, Paul etc.

Thankfully there is the ability to refine the search. Also when families move around the spelling can alter and you might miss someone who should be connected to your family

Comments

But if I go to your profile and click on Poole in your name at the top, I only get Pooles.

The complaint here is that clicking on Gräf produces a list of Grafs.

---

Not only that, but in German (and also Swedish) those are two different letters.

---

Then there is the case of Löwenstein and Loewenstein, and sometimes Lowenstein. I adopted an entire abandoned gedcom of over 1,000 profiles, in which that name appeared hundreds of times, in all variants.  It was painstaking to determine which spelling was the correct LNAB for each of them.  Some changed when emigrating but others didn't.  Some of the next (or later) generations changed, while their siblings didn't.  The result is that the variants need to show up in searches in order to include the right people.

Please note - I'm not trying to argue for one way or the other - just illustrating the case for including variants.  What I would like to see is options for degree of fuzziness when we search (to satisfy all needs), but that would be way too complex to ask for.

---

The German version of it is Löwenstein (with the points). The international transcription of it is Loewenstein. For Lowenstein I suppose that they either didn't know that transcription or it wasn't yet the usual case for it. But true, if I really wanted to look for all my Eckstädts, I would have to look for Eckstaedts and Eckstadts as well.

Answer 3

Was meant to be an answer grrr

A possible fix is to filter to show exact spelling as entered in search box?? A question for the tech team :)

Answer 4

As Gaile suggested it would be nice to have an option. The problem is that for as many people who would prefer to have the letters distinct, there are as many who want them to sort together in searches. We see that with other languages as well. We see it with French as well.

Answer 5

Names in Swedish,Norwegian, and Danish get changed when the immigrate to a place where the diacritical marks on the letters don''t exist.  Two different A's become one. Two different O's become one.  And so forth.

At wikitree, we have some advantages dealing with a worldwide genealogy site, but there are disadvantages also.  Some of us have to use special keystrokes,which we don't always remember, to spell the names properly as our key boards don't have multiples A's or O's, etc.

Answer 6

All searching and sorting depends on a collating scheme, which defines what order to put things in and what things are equivalent.

Trouble is, conventions vary from one language to the next.

But WikiTree can't do that, it's "one size fits all".

Universal schemes exist, but of course they aren't very good.  Inevitably they'll do things that are sometimes "wrong".

And names are the worst thing to get wrong, because people are so precious about them.

https://en.wikipedia.org/wiki/Collation

https://en.wikipedia.org/wiki/Alphabetical_order