If I quit WikiTree, the culprit will be LNAB

+27 votes

The LNAB was the main reason why I was a long time reluctant to join WikiTree. I eventually got over it, and am happy I did, because I've learnt a lot here. WikiTree has other issues, but I can live with all of them. They are the issues of all online collaborative endeavours, which did not compromise the growth and success of Wikipedia, for example. But LNAB keeps getting in my way, and on my nerves. And it will be the one and only reason why I quit, if I do.
And it could unfortunately be the major stumbling block on WikiTree's path towards achieving its universal mission.

I could go for hours listing the dire consequences of this basic feature which is in fact a serious congenital bug. And I'm on the verge of going away because of it. The problem being, where to? WikiTree is technically flawed, but there is nowhere to go with a similar spirit and community. So let me sum up once for all why LNAB is a bug, even if it's useless because I know the answer : c'est comme ça et on n'y peut rien!

1. Using a name as a basis for identification in a non-starter, and the more so in modern information systems. Even librarians, who have a looong tradition of maintaining "authorities" with "preferred names" and "alternate names", and the like, who have munched ad nauseam over the issues linked to naming and indexing for centuries, have eventually understood what LaoZi wrote in six characters very long ago, 名可名,非常名. Librarians came out with shared identifiers (dumb numbers) around which they gathered all the variety of names for a person, or for that matter a place, a book, or any subject a library catalogue wishes to reference. Each great national library did it, and then they sit together to create VIAF, a ring to rule them all. I'll be back with an example below.

2. Why "last name" and "at birth"? Looking at some recent and less recent questions in this forum, the mandatory LNAB to create a profile leaves to ridiculous debates and waste of time and energy. And what if there is no birth record available? And what if the birth record had it wrong, and was corrected afterwards? What if the siblings in the birth records have different family name? And what if the person belongs to a country, a language, a culture, a period, where the very notion of "last name" does not make sense? Those things happen all the time, and I read in a recent debate here that people just do not create profiles because they can't figure the LNAB.

3. Adding to the above, if you make a mistake when creating the profile, even a simple typo, of course you can change it, but it means creating a new profile and merging, and we are said not to do that too much, because it creates a lot of redirects in the data base, which is an overload on the system. A problem perfectly technically understandable, but from a end-user perspective, is seen as a big bug.

4. From a search point of view, LNAB is often not the most frequent name under which the person is known, and people searching in WikiTree, directly through the internal search engine, or from an external one, are likely either to miss it, or have hard time recognizing that they have indeed found the person they were looking for. I had this experience a lot of time searching famous people in France history, for whom I had to try several names before finding them out, or go through some member of their family I know was there etc.

When you put all that in the context, say, of French aristocracy, where the syntax and semantics of names are both very codified and of course suffering many exceptions to its own codes, it's bound to be a total failure. The adaptation rules chosen to make those convoluted names enter the data base structure are just pathetic, and will send any serious french genealogist de bonne famille to roll on the floor laughing, or cry in despair, but by no means will he take the whole thing as serious, and never ever join us. If any true French aristo with seize quartiers de noblesse is around here, could (s)he please step forward?

And moreover most of those profiles have been created by people with only vague notions of those naming subtleties, but who happen to be far descendants of an aristo, or think they are, they have it wrong. Which should not be a problem, because the Honor Code says We know mistakes are inevitable We don't want to be afraid to make them. But if mistakes relate to LNAB, you are not welcome to correct them because of point #3 above.

I stumbled on one of those, in reference, and I put a comment on the profile that Henri de La Tour d'Auvergne means, as far as I understand it, a given name of "Henri" to a member of the "La Tour" family, of whom fief was "Auvergne" at some point (a province of France). It was before "La Tour d'Olliergues", Olliergues being a city in Auvergne. You see there that the family has gained power. More at https://fr.wikipedia.org/wiki/Maison_de_La_Tour_d%27Auvergne

So, all people in this family, if we were to apply at best the silly LNAB rule, should certainly have as LNAB "La Tour", aka "d'Auvergne" for our Henri, best known in Histoire de France as "Turenne".

The answer by Isabelle, after we had a private exchange about it, reads : yes it's a mess, but what could we do, because of all above listed reasons? My answer is : get rid of LNAB, rethink the whole naming issue on a really universal basis as librarians have done at some point, to allow any local naming convention, save the data, and migrate to a new data base. WikiTree will have to do that at some point, or die out, or stay forever a permanent frustration for all those here who are dreaming its dream. If change has to be done, the sooner the better. Migrating 20 million+ entries is not easy. When we are at 50 or 100 million, it will be worse.

Going back to VIAF : https://viaf.org/viaf/59190793/ is our man, see how different libraries have chosen different "preferred" ways to name him ... and the 218 (to date) "other forms of the name", just missing the one chosen by WikiTree.

[edited to get rid of rude language, thanks Julie for pointing at itwink]

WikiTree profile: Henri de la Tour d'Auvergne
in The Tree House by Bernard Vatant G2G6 Mach 2 (27.2k points)
retagged by Bernard Vatant
thanks for adding him

Wonderful read, thanks Bernard. Well put, and recognizable in sketching the frustrations of working within the confines of WikiTree (also because the many benefits WikiTree has over it's competitors). As a librarian by vocation I understand your description so well, For me personally the it is also hugely frustrating that as in the world of libraries and information systems, it is near always comes down to a standoff between the 'tech' guys - the programmers - who only seems to understand code and bugs - and the library-type of persons, who also understand the way information is used, gathered, harvested. Both are correct, but more often than not in disagreement. In some libraries the facilitating technical systems are state of the art. If WikiTree could be seen (and see itself) as a giant library (which it also is), perhaps there is a future for it. Because I too have seen that "WikiTree will have to do that at some point, or die out, or stay forever a permanent frustration for all those here who are dreaming its dream." - the latter unfortunately meaning that I personally along with all the others have wasted years of labour for it's cause. Having read the responses following below, I also realise that I should get out of the way with my opinions and let those who understand the jargon get started with creating the solutions.

It's a question for librarians, because they've been shelving books alphabetically for years.  Mostly by author's name.  But not exactly as printed on the title page.  They subconsciously use a transformed version of the name.

So if Henri de la Tour d'Auvergne, vicomte de Turenne wrote a book, where do you shelve it?

Question 2.  Your library is very international.  Most of your books are in foreign languages and alphabets.  How do you organize the shelves?

Question 3.  You have a nightmare in which you're running a library for United Nations delegates.  On investigating international libraries elsewhere, you find that they all have different answers to Question 2, to suit the habits of their local readership.  But now you don't have a local readership.

I think we find that internationalization means different things.  To bring the whole world together into one system, you have to ride roughshod over the conventions and customs of different countries.  Respecting different cultures means fragmentation.
Try the Library of Congress - If I remember correctly all books ever published should at least have one copy in that library, organised by the Dewey Decimal System. But there are also there are different ways of organization a library - genre, subject, author, alphabetically, sytematically (for example UDC). But this has all been hashed out in the past here on WikiTree. The difference between say the world 30 years ago and now is the amount of digitized content on the internet, which is harvested and indexed in new ways. So this asks for new solutions. I for one are too old school to be part of those solutions except for sharing the end-user experience.

The Library of Congress has a Confusing amount of numbers on their Catalog records. And that is a quote from the Frequently asked questions section at https://catalog.loc.gov/vwebv/ui/en_US/htdocs/help/faqs.html. If they can't have one system then what hope does Wikitree have to have one system that will respond to every different way people are named?  

There is multiple current ways that people are named today that don't fit a First Name, Middle Name, Last Name version and that doesn't even take into account previous ways of naming people. Coding the Wikitree Database to reflect every way every culture may have used is very unlikely as even small changes result in unintended circumstances. I believe one recent change resulted in European profiles with multiple given names throwing up error messages on every edit even though for those profiles it was not an error. Not using LNAB would be a major change and thus very very unlikely. 

That's not so strange. Few libraries have "1 system" - What is meant here for example (as explanation to the uninformed) - an ISBN is not the same as an ISSN, yet they (these internationally recogized standard numbers) both denote different types of material, in this case books (ISBN) versus periodicals (ISSN). And many libraries also receive whole collections from elsewhere with a different organizing principle.

I do not want to distract from the question of Bernard above but if you scroll down further you'll see the following (quoting) "The LC Catalog uses Unicode (UTF-8 encoding) for searching, displaying, printing, and downloading records. In 2004, all Catalog records were converted to the Unicode standard for MARC 21. In addition to Roman scripts, the Catalog contains records in the following languages/scripts: Chinese, Japanese, Korean; Cyrillic-based scripts; Greek; Hebrew, Yiddish; and Perso-Arabic script (e.g., Arabic, Persian, Pushto, Sindhi, Urdu). For more information, see Searching/Displaying Non-Roman Characters. In general, your preferred web browser settings will not need to be changed to correctly view records that contain diacritics, special characters, or non-Roman characters. If you have problems, however, you may need to reconfigure your browser opens in a new window. When searching the LC Catalog, most marks of punctuation in your query are converted to spaces. Some punctuation and diacritic marks are removed: apostrophes, alifs, ayns, middle dots, primes and double primes. A few special characters are retained in searches: ampersands (&), plus signs (+), at signs (@), number signs (#), and musical flat (♭) and natural (♮) signs (musical sharps are converted to spaces). Special characters are generally converted to their nearest alphabetic equivalent (for example, search an æ diagraph as ae or a þ thorn as th). Bibliographic and holdings records can be downloaded in either MARC UTF-8 (Unicode) or MARC-8 (non-Unicode). The Library's Cataloging Distribution Service (CDS) also provides MARC 21 bibliographic and authority records in both MARC-8 and UTF-8."

I guess what I'm stating is that as with WikiTree, many libraries struggle with similar issues, and (as WikiTree with records ) also aim to encompass all of their materials in a coding system that does justice to both the content and the search for information and materials. You state below: "How to prevent profiles being created because someone thinks it isn't on Wikitree as their preferred spelling isn't used rather than the Project approved spelling is something that could be a good thing to look into. Using every possible spelling or version of the name only creates duplicates which is against the basic mission principle of One person, One Profile." Creating a parallel index (to the current LNAB) with standardised names might be one answer (for a few projects at least), but I doubt if it would be suitable for all of WikiTree and would be equally feasible.

In response to your last suggestion of a Parallel index and tying that back to the linked profile on this question there are 

the 218 (to date) "other forms of the name", just missing the one chosen by WikiTree

That is an insurmountable amount of names for any project to list for just one person and if every Aristocrat has a similar amount of names it would take up an immense amount of time to find and record every name. Never mind any Notables that have "Stage Names" and multiple other options. 

Then there is the fact that the linked profile is most known as Turenne apparently which is part of his title of Viscomte de Turenne. (I only say apparently as until I read and responded to this question I never knew of this person). That would be like me expecting to find the profile for Prince Harry, Duke of Sussex by only using Sussex (Of course if it was visible). A title isn't part of someones name and can be gained and lost at any time. For example Prince Charles, The Prince of Wales wasn't born with that title and some day will lose that title when he becomes King. 

I for one wouldn't want to work on a genealogy site that could have 218 versions of the same person because people couldn't agree on any one part of the name to base the profile off. If we get rid of LNAB as mentioned in Bernard's point 2  then it becomes a free for all and multiple profiles will result as a consequence of that decision. It happens even know with LNAB required but it is greatly reduced and any duplicates are easier to find when combined with the Parent Child links. 

Using an example from my watchlist if you found Elizabeth Kirkwood, Elizabeth Kellett, Elizabeth Condie, Elizabeth Gourlick and Elizabeth Long listed as profiles It would be very unlikely that anyone would know that these five names represent the same person. Yet they are and having one profile based on her LNAB helps to keep the information in one place. She had multiple profiles on Family Search where people can enter whatever they decide on. And to counter Bernard's point 4 in the original question;  All five options of name when searched for bring up the profile for my Elizabeth on Wikitree. So forcing a LNAB does not prevent profiles being found by the name they may have been more known by. What forcing a LNAB does do is help reduce duplicates. 

The fact remains that having it with no part of the name as the Database identifier would only result in Duplicates galore. A LNAB required Database identifier reduces duplicates and encourages cooperation and collaboration more than any other method. Any Full name or Usual Name system opens itself to duplication of profiles as people always have the ability to change their name or not use part of their name whereas a Name at Birth provides a Fixed point that is the first name someone is known by. Any random letters or random letters+numbers system opens itself up for duplication as well and would be more likely to have mistakes done by people entering links to other profiles. I base my statements on my experiences on other Genealogy sites and programs. 

I do not dispute this. In agreement.
As someone who is a 'tech' person AND a librarian, I sympathize with the LNAB frustration. Librarians imagined 'access points' which were names filed alphabetically on cards, and this approach was transformed to electronic bibliographic (and authority) records. In modern practice the standard number or identifier was preferred over the name or label, and that was preferred when we moved to URLs/URIs. Wikibase imagines a similar style where an item identifier is created (Q#).

The problem with LNAB is that instead of using an identifier number or a disambiguation scheme a profile is created and uniquely identified with a last name. Ultimately, this doesn't work across various naming 'eras' along with non-Anglo-Saxon naming practices. I think the data model would need to be redone - rearchitecting Wikitree to use Wikibase item identifiers instead of LNABs probably is a good option long term.

I think everyone who has done Wikitree pre-1500 genealogy has recognized you can consider LNAB as a collocating term which is manipulated for that purpose rather than as a surname. The bio can have all the details you want, albeit not very machine-actionable. The various brief and search views are incorrect, but you have a resource which is well-cited. We continue to ask for additional fields, along with search and indexing enhancements but in the past few years nothing has changed. But this site is free, has a good community and the quality is relatively high.

I know I would appreciate it if you'd stick with WT for a while - I think building non-English language trees is very helpful even though it is hard.
Well said Kirk.

11 Answers

+14 votes
Bernard, I understand why the LNAB does not work well, to say the least, for European nobility, etc., and maybe there ought to be a special rule for them.  But--sorry if I'm missing your point--for the vast majority of profiles on WikiTree, what would be an equally convenient means of indexing and finding them?
by Julie Kelts G2G6 Pilot (282k points)
Bernard, if I may chime in here one last time, I would like to offer a suggestion.  You do seem to understand the problem well enough, and you clearly have a fair number of supporters who agree with your complaint.  Rather than just let the whole discussion drift into oblivion, could you perhaps consider creating a free-space page (or a white paper, if you prefer) that summarizes all of the features and capabilities that would be needed to support different naming conventions that the site does not provide now?  You could enlist some of your supporters to contribute their own ideas and inputs, and perhaps achieve some sort of consensus about what capabilities are needed.  If appropriate, it could be subdivided into things that are essential or nice to have, that would support Language A but not Culture B, and so forth.  I'm suggesting you could just articulate the requirement, not the solution.  If you are not ready to make a specific proposal or request that's fine, but I could see such a document being very useful if/when the powers that be do decide to get serious about providing better international support.  Management will certainly need some help from those who live with the other conventions in order to get it right.
Good suggestion, Dennis!
Dennis, that's a very constructive proposal, thanks a lot. I will try to do that, although it will maybe first on my personal web, in the form of PDF slides, I'm not familiar at all with the free-space pages.

Bernard, free-space pages are quite simple to set up, and you have a lot of freedom as to their construction, etc.  Use this link and then just click on Creating Free-Space Profile..."Simply use this form."

As Julie says, it's easy as could be, and the pages have privacy settings just as people profiles do.  You could make it Open and invite others to add their own thoughts and comments, or you could make public and add only those you wish to the trusted list.  A site leader suggested a similar thing to me for trying to clean up a lot of inconsistencies in a big family with about six or eight profile managers involved.
OK, thanks, I will have a look.
I will be happy to have a look as well.

Darren Kellett wrote:

Granted the Last part of the acronym isn't the best( I think Family name at Birth or FNAB might be better) but it becomes a matter of terminology which can have multiple meanings even within the same culture never mind other cultures. A FNAB provides a fixed point that won't change whereas people can change names thru marriage or stage names or other means.

Then you forget the fact that many cultures didn't generally have family names at all. Where I come from, common people didn't have family names in a legal sense until 1923. Most got by with their patronym. A typical family could consist of Hans Olsen married to Anne Pedersdatter, with the children Peder Hansen and Kari Hansdatter.

Common practice with respect to Scandinavian patronymic culture here on WT is to enter the patronym as LNAB. Other genealogy sites have chosen to use the concept of "Farm names" (which really are addresses) as family names, thus imposing a set of conventions which are totally incompatible with the historical facts.

There is nothing intuitive about "Patronym as LNAB" though. Therefore, we find Norwegians all over the place who have been registered with Farm name as LNAB, patronym as middle name, or patronym + farm name in the LNAB field. Plus a whole slew of other methods. That's one of the results of having a data model that doesn't account for other naming practices than 20th century American.

People often build their trees on other sites.  The gedcom arrives with a "surname" marked off with slashes and WikiTree follows that.  The issue needed to be addressed somewhere else.

There's no need to get hung up on labels.  The label on the box is only a mnemonic, not a definition.  The use of the box is defined by what the software does with it.

Historical facts aren't an issue.  Question is, if users need to be able to see whether a certain name appears in a long list or not, how should we organize the list?

I am directly descended from people born in the 1550's that  are on Wikitree and have baptism records with identifiable LNAB/Surname/Family name whatever anyone wants to call it. Having a Family name (My preferred method of calling it) isn't a recent American thing but it is something that most cultures have adapted into having at various stages. Just because one culture took until 1923 (To use Scandanavian culture) to have LNAB's does not mean Wikitree can't use LNAB's as the webpage naming system. 

There is nothing that requires the LNAB field to be the same from parent to child at all. What this field does however is have one field that is used for the Public display that is unlikely to be subject to change. People can marry and change their name, People can change their name due to their profession, People can change their name due to moving to a different location. The one thing that won't change is the name they were born to.Therefore that field is used for the Profile name in Wikitree. There are other fields to reflect the other names adopted during their life. Some affect the display name in lists and others don't. The Wikitree search engine usually finds them. 

If Wikitree was to go away from using the LNAB as the primary method of displaying the difference of profiles publicly what would the replacement be? For example for a Royal person, Queen Elizabeth II has at minimum 18 different ways that she could currently be known as. Would you prefer that 18 plus profiles be created for the one person or have one method that can be agreed on which results in https://www.wikitree.com/wiki/Windsor-1 being the profile for her. Or would you rather search by random number and/or letter combination that most people would never guess relates to a person? Examples related to how Queen Elizabeth II are recorded in other databases are 20471592, 327145858093723022004, a11257799, b8nqqnnv2ndsn1j, 34452, Q9682 etc. Two of these specific examples are for the same database as well. So there is no one way that everyone can agree on to ensure that every reference to the same person is linked to that person in every database. 

A LNAB provides a way for Wikitree to identify one profile per person. It doesn't completely prevent duplicates but it does provide a way for duplicates to be easily identified when combined with the Parent Child links. For example I think there may be two duplicate profiles for Queen Elizabeth II on Wikitree both created by relatively newbies (Less than 300 contributions) and one duplicate for her father and mother respectively made by one of those newbies. The term LNAB isn't right for a lot of people (For instance those cultures where the family name is the first part of their name for example) and is why i personally would like it labeled as Family Name at birth and with a bit of coding the display could be adjusted to reflect a Family name first naming system. Similar coding could reflect a name was Patronymic or a Farm name or a House name and change the label to reflect the situation. Something similar to the recent coding done for Middle name/ Multiple First names that I saw mention by some Dutch Wikitreers a couple of months or so ago. 

However that coding does't affect the database coding system whereas a change to using something other than a LNAB would massively effect the Database. Would any of the people commenting on this thread volunteer to change every Orphan Profile to reflect whatever the new system could be or would they end up being a relic of the LNAB naming system?  Then what about those profiles that are managed by Wikitreers who aren't regular users? They would most likely not bother seen as there are managed profiles from 2011 that haven't been edited since then. 

Wikitree does allow names that people were not born with such as Anon, Unknown, D, G etc or Project approved names like I suspect The British Royal family names were decided on by the British Royals and Aristocrats project. How to prevent profiles being created because someone thinks it isn't on Wikitree as their preferred spelling isn't used rather than the Project approved spelling is something that could be a good thing to look into. Using every possible spelling or version of the name only creates duplicates which is against the basic mission principle of One person, One Profile. 

+10 votes
I agree the last name at birth doesn't work well particularly with the Euroaristo.  It is rare to see an actual birth certificate as the best information is found in the house records.

Not all families are like ours where there is a genannt and where the noble name is heroic rather than geographical.

As the Baltic Nobles also were a cadet nobility which spread into many other countries, there is no consistent spelling.

And a noble may be known by a spelling other than his birth one.
by Lloyd de Vere Hunt G2G6 Mach 1 (18.9k points)
+11 votes
The complication is that bios are stored as MediaWiki pages,which are like Wikipedia pages.

The MediaWiki people did some cute hacking here  They use pagenames as unique IDs for pages, with the idea that in Wikipedia the pagename is basically the article title.  The wins are

1- a wikilink can often be created (in Wikipedia) just by putting square brackets round a word or phrase, without having to enter a page ID

2- in listings of pages, eg category members, the pagename will do as a display name without having to retrieve further information

3- listings can be sorted by pagename (and the pagename sort is just bytewise, or maybe with a Unicode collating map at worst).

So the WikiTree ID was invented to serve as a page ID.  But it's not easy to devise pagenames to meet those objectives.  Category listings used to be sorted by WikiTree ID, but now they get re-sorted.

But the sort key still starts with the LNAB.  If we didn't have the LNAB as we know it, the issues wouldn't evaporate, they would resurface in constructing sort keys.

But I don't think the MediaWiki people would pretend that their cute scheme would work for a multi-alphabet multi-language pagename space, or with random non-significant spelling variation.

And I don't think librarians have solved those problems.  They often transliterate.  They probably sort names using a scheme designed for the customs of their own country and makeshift in dealing with a minority of foreigners.
by RJ Horace G2G6 Pilot (550k points)
That is why there is wikidata now, to handle with the massive flaw of a title name as the title URL identifier.
@RJ All good points, and Michel took the answer from my mouth. Except that Wikidata is now in a spiral on uncontrolled growth without quality check, which raises other problems.

I'd rather point at VIAF, as said at the end of the post. Librarians did it perfectly right there. You can't solve the naming issues by using more names. URIs and opaque IDs are the way out of the naming mess.

And BTW, I think we should start adding VIAF IDs to profiles, but I will post ASAP a separate thread for this proposal, but please have a look at https://www.oclc.org/en/viaf.html
Wikidata is still supposed to have a notability criterion, of sorts.  At least, they should be notable enough to have published genealogy.  It's not intended for massive numbers of ordinary people. Eg if you imported complete censuses, you'd have massive duplication across decades that could only be resolved by doing all the genealogy.
It used to be, but notability criteria for Wikidata are more and more elastic. You can add just about anybody without anybody complaining. VIAF has an objective criteria of notability, which is the presence in at least one library catalogue, often many more, meaning the person is either an author of some book/publication, or is quoted in them. Either way, having a VIAF ID means there are documents stored somewhere in a library shelf by or about you. Which is, in the information world, the cleanest definition of notability you can find. Moreover, through this ID, you are likely to find references to the said documents, aka sources.
Another aspect is that Wikipedia is an encyclopedia, not a relational database. Print encyclopedias and dictionaries have headwords/titles and cross-references as their lookup method, with no index or even table of contents; the vagaries and nuances of database structures are simply not applicable to them. The transfer to HTML and its hyper-cross-referencing capabilities doesn't fundamentally change that.

WikiTree, on the other hand, very much _is_ a relational database. Genealogy software is all a subset of database management software. WT, being cobbled together mostly based on text-management software and on unconscious American assumptions, is sometimes rather singularly unsuited to its purpose. Its handling of names is the prime example of this.
@J Palotay : what I begin to understand, thanks to various above comments of @RJ Horace and @Darren, is that a part of the problem is that the Web layer of WikiTree is based on MediaWiki technology, which needs identification of pages based on "linked names" (the genial idea of Wiki inventor Ward Cunningham). Hence the invisibility of primary keys which indeed exist in the data base, like in any data base, but are not used as IDs for the Web pages, because it conflates with the whole Wiki idea. OK so far.

That said, remains the choice of mandatory "LNAB", among all possible choice of names, which has no technical justification. Wikipedia, based on the same Wiki technology, has no constraint whatsoever on the structure of a name, and has set pragmatic solutions to the homonymy and synonymy problems, without problems for search.
Are those listed benefits really important? The comparison to Wikipedia also reminds us of a big difference. On Wikipedia article names are changed constantly. No one cares about that or thinks that using Wikipedia depends on article title searching.

Wikipedia has no capability for sorting lists of names.  It has a lot of lists of names, but they all have to be sorted manually or not at all.  Results can be poor


There's no consistency between lists.

Reader-sortable tables are a feature.  If a table has a name column, you should be able to sort the names in alphabetical order.  You can't.

In genealogy, the ability to generate ad-hoc lists of names and sort them on the fly is essential.

Looking at the VIAF page, most libraries would sort Henri under Turenne, his title.  The same libraries would probably sort "Diana, Princess of Wales" under D.  This sort of thing is fine for libraries but it won't do for genealogy.

Computers can't sort free-form names by some kind of intuition.

I personally don't find that so essential. I think google style searching has changed the way people find things. Sorted lists are often just ways of finding things.

The only time we need to still rely on alphabetical order now is for smaller printed lists that people need to go through "by eye", which we can often just rejig manually?

But in any case, surely the only way to ever really fix this (or make a least worst solution) is with a separate manual sorting key like Wikipedia?
But then we're back where we came in.  We do away with the LNAB box, we add a Sort Key box, and the problem of what to put in the LNAB box becomes the problem of what to put in the Sort Key box.  The only thing that's changed is the label.

If the labeling and the naming rules made it clearer that it's all about sorting really, then the issue would mostly go away.  Or at least, we'd understand the question.  Should Henri appear under La, or Tour, or Auvergne, or Turenne?  The computer needs an answer from us, and takes one, even if we were fooled by labels into thinking we were answering a different question.
That does not seem like a crazy direction to think in: a "sorting only" field.

It perhaps means another field, although perhaps it can be done with some re-purposing.
+10 votes
I just want to state that I totally agree. There is no technical reason, except for the 'old fashioned way that wikis used pagenames and searching' to use the LNAB as the identifier. Any modern wiki uses real IDs. See Wikidata, see musicbrainz, see all documentation on linked open data. Even the data doctors use ids...

These are all technical reasons and will not change the way the 'wiki' part works. You can still use templates, links etc. It's just the programming that needs to change.
by Michel Vorenhout G2G6 (9.3k points)
+6 votes
I would like to defend the WikiTree system and offer a few suggestions:

1) Last Name at Birth should stand. It is useful information.

2) Current Last name could stand (or be changed to Last Name at Death.)

3) Please increase the number of Other Last Names fields. They could be called Other Last Name 1, Other Last Name 2, and Other Last Name(s) 3 and more. Many profiles would not need all of these fields, but this would help deal with aristocracy, people who were adopted, and woman with multiple marriages. I believe that this would allow for easier database searches without substantially changing Wikitree.

4) Please increase the number of Middle Name fields. They could be called Middle Name 1, and Other Middle Name(s) 2 and more.

5) The European nobility represents a small portion of world names. If we want to be more accessible and include people from countries such as China, Japan, South Korea, and so on, "Last Name at Birth" could become "Family Name" and First Name could become "Given Name".
by Anne Creighton G2G6 Pilot (133k points)
So names are very confusing in the Irish Landed Gentry.

So our name is de Vere Hunt.  But based on the pre-fire proof, the family was licensed to use the name de Vere for Ireland, having proven descent from a female member of the family of the Earls of Oxford.  When that occurred, a Phillips who had female de Vere Hunt blood, took the Hunt name over.  So the Philips name was open and it was taken by yet another family with a female descent from de Vere Hunt.  Then by a questionable will - there were males available, the property went to an O'Brien who had married a de Vere an then he became a de Vere.  

The above scenario also replayed in many other families, just a massive domino effect.
Some of us had an entertaining moment a while back about the so-called "American naming system" of WikiTree. Did you know that the Americans are supposed to use "Their conventions instead of ours" on WikiTree as we all do, but they don't. The phrase "Middle Name was first used in 1835 in a publication, Harvardina. Then you should allow for when states or the Federal Government actually started using the phrase in official documents which could be another 50 or 100 years later. It simply means that the American ancestors on WikiTree are not named correctly, at least definitely not before 1835.

Read for yourself here is the link Havardiana

They probably called it their 'second name'.

You asked what they might be called if not a 'middle' name.  I answered: they probably called them 'second name'.  That would be the same as 'second forename'.  They just weren't called 'middle name'.

Please, let's get back to the topic. This question was about the Last Name at Birth. There are several other threads available to discuss the "middle" name. Thank you....
Isabelle, you're right.  Sorry for posting.  I'm going to hide my previous comments.  Others can do the same if they wish.  Apologies to readers if they can't make sense of the last few posts.  

Note to mediators:  Please don't reshow (not that you would, but sometimes I've experienced that).

I hid my previous response because I unintentionally hijacked the topic to other parts of the name.  Mea culpa...  

So to turn it specifically and only to LNAB...   

I still respectfully disagree that this is the only name we should really query on or care about.   

Most of the pre-1500 and pre-1700 profiles have made up LNAB typically using either where they were born or if a dynasty or house is known using that.  This is why querying for more recognized names of Kings or famous people of olden times are not easily found.  

Sometimes decisions are made that make total sense at the time of that decision but databases change based on the new data being accumulated and this might cause reason to reconsider older decisions.  I think with WT's desire to branch out to other cultures this is a perfect reason to review and perhaps rethink old decisions.  It does not mean they were wrong, but it might mean there needs to be some modifications.  One thing I would think about adding is a field called commonly known as.  This would allow someone to drop in Charlemagne and not be taken here


but be taken here Carolingian-77.    It will also cover those instances in Germany where a man married a woman and took her LNAB because it as the farm name and the farm name was more important than th family name.  

And it would cover pseudonyms like Mark Twain.  

Laura, you arguments make sense, but I wonder whether changes will be made.

After I saw your post, I Googled "Charlemagne on WikiTree" and the very first listing was the right one.  I wonder if as, at least a temporary measure, we could just refer people to Google?  After all, it is there, it works...

That might not cover ordinary people, but at the least the notables could be found more easily.
I did not mention the "middle name" to take away the discussion about "LNAB", but to point out that the current naming system does not work correct even for American ancestors. The whole naming system should be discussed not just fractions of it, because what you change in one part of the system will ultimately effect the rest of the same system.

I'm sure, we are all serious about genealogy that is why we respond to these issues on WT. Brushing it aside does not make something go away.
+8 votes
Basically database managers like to select and sort and join records based on fields.  Whole fields, as stored and indexed.  So fields need to be designed that will work for these purposes.

Users won't feel that their great-grandmother's sort key was her most important attribute.   But they want sensible sorted lists.  If she doesn't appear where they expect, they won't scan the whole list and find her, they'll just assume she's missing.

But if the fields you need for database operations are user-supplied data, and can't be selected from a fixed menu, then you have a big problem.  As we know from experience, users will try to enter all sorts of junk in name fields without regard to function.

The LNAB field is essentially the first sort key.  Watchlists, category member lists etc will come out sorted by LNAB first.

(In fact category members used to be sorted by pagename, ie WikiTree ID.  This was the main reason to create a new page when changing an LNAB.)

So the LNAB rules are supposed to be an attempt to get the field populated with something that works right when sorting lists.

Unfortunately people have taken their eyes off the ball, including Chris, and extraneous and conflicting considerations have been dragged in.  Creating the current mess.
by RJ Horace G2G6 Pilot (550k points)
+6 votes

Je vais m'exprimer trivialement et en français. Wikitree est prise de tête. Moi, je m'en fous que l'enregistrement ne respecte pas les procédures qui sont si nombreuses qu'il faudrait passer un permis de remplir avant de se mettre au travail. Je veux des profils avec des informations justes et vérifiables (et sans payer, les liens vers Ancestry avec adhésion, j'en ai marre).
by David Quénéhervé G2G6 Mach 6 (63.5k points)

David, je ne suis pas sûr de comprendre. Tu veux dire que je ne devrais pas me prendre la tête avec ce genre de question, et utiliser le système en l'état en me préoccupant juste d'entrer des infos correctes avec sources, et me moquer du reste? C'est certainement la position la plus sage! Et sur Ancestry j'applaudis des deux mains yes

Bernard, it might be nice if you would translate for the rest of us.  Here is what Google tells me about David's answer:

"I will express myself trivially and in French. Wikitree is taken of head. I do not care that the registration does not follow the procedures which are so numerous that it would be necessary to pass a permit to fill before starting to work. I want profiles with fair and verifiable information (and without paying, links to Ancestry with membership, I'm fed up)."

And your response:

"David, I'm not sure I understand. You mean I shouldn't be bothering myself with this kind of question, and using the system as it is, just worrying about entering correct information with sources, and making fun of the rest? It is certainly the wisest position! And on Ancestry I applaud with both hands..."


It's obvious to me that for whatever reasons (technical, economical, sociological, psychological...) they don't want major changes. Wikitree is supposed to be International but remains very Anglo-Saxon. I guess that many Continental European users are in favor of some key changes. We use Wikitree because it's the only one of its kind and the idea is interesting but I don't really need it for my own researches.

Thank you, David.  Maybe I should have asked you in the first place, but not knowing you, I did not know anything about your proficiency in English.  I posted the Google Translate version because it was funny (to me), and I like to laugh!  laugh  I'm sorry if you were offended in any way.

Now that I can reply to what you have said:  First of all, as I mentioned to Bernard elsewhere in this complicated thread, he has not specifically addressed WikiTree management.  My own guess is that there are significant technical and economic challenges to making major changes to WikiTree's architecture.  On the other hand, though I am not really qualified to hold an opinion on the subject, I imagine that such changes will be inevitable if WikiTree is to survive.

Of course we don't particularly need WikiTree for our research, because it is not a research site.  By that I mean that WikiTree doesn't have its own archives, like for example Ancestry does.  Yet I use it because it is free, and I can share the information I have with more people, including my own relatives who don't want to pay for subscription websites.  Also I enjoy the community!

Finally, I appreciate that WikiTree is overwhelmingly Anglo-Saxon (or I'll take your word for it), but to be fair, so are its users.  So expanding options in a way that are useful everyone is a good goal, but not at the expense of the largest user group.

I'm glad WikiTree has international users.  That enriches the WikiTree experience, and, to put it selfishly, I like the opportunity to correspond with people from many parts of the world.  So I hope you will continue to find it worthwhile.

Julie, the main reason why I came to WikiTree, and the main reason I'm still here, is this, read at https://www.wikitree.com/wiki/Help:Collaborative_Family_Tree

"A collaborative family tree, or single family tree, is one that we all share. It's not your tree or my tree. It is our tree — a tree for the entire human family."

Highlighting of the three last words is by me. I took them at face value, please tell me now if I was wrong. The Anglo-Saxon world is just one part of the family, and, sorry to say, not the largest one. And as long as it is the largest user group, WikiTree is missing its declared target. The objective of WikiTree in the long run should be to have a community of users representative of the entire human family.

If I believe https://en.wikipedia.org/wiki/Anglosphere, the "Core Anglosphere" (USA, UK, Canada, Australia and New Zealand) represents 6% of the world population. I guess in WikiTree, it's rather 94% of users who belong to this world.

Seems often to me now that non Anglo-Saxons are welcome here insofar as they can help with finding pre-1600 European ancestors of American immigrants, if possible famous ones, and decipher unreadable French, Russian, Dutch, Spanish ... records.

 I won't be rude, but I could have the same attitude, and care only for my own ancestors from Brittany, where there are excellent genealogy societies making a tremendous job about them, such as that one http://cghp-poher.net/

Bernard, I'll reply privately.
@Julie: If you take the chicken, I will take the egg...
Sorry, Michel, I don't know what you mean.
Julie, if WT remains this US focused, it will remain very hard to attract no US citizens. But as you say, as there are so little of us, WT does not feel the need to shift the focus more than it does now...


Out of the 7 Billion (currently) only a few are US based... so if the goal of WT is to create a world wide tree to enhance the feeling of 'one world', the rest of the world should be the focus more.

Thank you, Michel, I am familiar with the expression.  I wasn't sure what you meant by your use, but I suppose I should have understood.  No offense, but I think I've said everything I have to say in this thread and do not want to continue the discussion.  I hope Bernard will nicely summarize all the input on his free-space page.  smiley

2 questions:

1. Can there ever be one way of using the limited number of fields which can work for everyone?

2. Does there need to be?
Can there ever be one thing that pleases everyone everywhere all the time?

Smile. Enjoy WikiTree.

Well, I hope you're smiling, Chris.  smiley

Well Chris I think the question is whether the LNAB, as it currently stands, is trying to achieve too many conflicting aims, and creating smile reduction issues as an unintended side effect.
+5 votes
Bernard, I completely agree.  This is one of the reasons I've stopped contributing to Wikitree.  Being told over and over by Anglophone genealogists with no experience in the appropriate contexts that, say, my grandmother's grandmother should be "Rosen" rather than "Baronesse von Rosen" because of Wikitree's absurd LNAB indexing and adding to this the habit of "Data Doctors" butchering any attempt to create a reasonably correct noble surname-title combination in the current system was too much to bear.  If things ever improved, I'd be so glad to return!
by Kelsey Jackson Williams G2G6 Mach 1 (10.8k points)

Kelsey, please see https://www.wikitree.com/wiki/Help:Suspending_Your_Participation

You still have thousands of profiles on your Watchlist. It's important for the community that you not be managing profiles if you are not participating and collaborating with others.

+4 votes
LNAB works well for far more profiles than when it doesn't. If everyone used the LNAB and other last names as intended, it would be easy to find most people. It is very annoying to miss records because someone has arbitrarily decided what the last name really was or should be.
by Leandra Ford G2G6 Mach 2 (29.6k points)

Most people, I agree.

But many genealogy programs do much better. I use Legacy, and it does.

All that needs to be done is add one field, "usual name" that is searchable through any "word" it contains, a word being defined as being separated by first of field, end of field, parentheses, and space and only those. Curly braces would be prohibited.   The search would be specify using OR and AND as keywords, in expressions using curly braces as delimiters, as in  {John OR Iain} AND {McDonald OR MacDonald OR Lord} AND Isles for a certain ancestor of mine whose field would be "John MacDonald (MacDohmnialll) first (1st) Lord of the Isles".  Search would be not case sensitive except for the words AND or OR.  For non-Latin characters, transliterations would be given in parentheses in the field.

That would fix all of search.

This is not usually done because designers apparently like "things that are simple and give too many answers" rather than ones that give only the desired answer. I had numerous losing wars with Chem Abstracts database over this, as did lots of others (but at least they do have real unique identifiers for chemicals, that are searchable by chemical structure!)

Sounds a reasonable proposal to me.
Usual name provides plenty of opportunity for people to make stuff up. When someone lived before your time and you have 3 records for them if you're lucky, how do you know what their usual name is?

A Usual name field would be useless as who defines the Usual Name? I know of one currently living person who currently has 18 different legal names depending on Country never mind different languages creating more variants. 

Or another example we could use Sitting Bull as his usual name instead of the name used now even though it isn't his birth name (With no actual surname) . 

Or using an example in my watchlist I have someone who used five different Last names and I only know she used a specific name at times of records being made. There is great possibilities she moved between names and I can't find the records as she may be recorded under a different name than what I am thinking she used at that time. So she could have five Usual names. Which one could I choose could depend on which name I want to use. That may be different from those who are descended from one of the other families and could technically result in five profiles for the one person. 

So any User made name runs the risk of multiple profiles for the same person. 

 For those profiles that don't have an identifiable LNAB it becomes a question for the responsible projects to decide how they should be recorded as. If people don't like how the Project names the profiles then people should join the projects and prove their hypothesis and have the Project change the project decisions. Of course there is nothing to prevent Wikitreers to creating profiles based on what they think it should be but that means people may be inadvertently breaching the one person one profile mission of Wikitree.  

For most Genealogy programs it becomes user preference what is used. For Web based Genealogy sites (I am basing this on Family Search, My Heritage, Ancestry, Geni) they are all user preference entries and they all result in Duplicate profiles. I don't want Wikitree to become a Duplicates welcome type site. Having the LNAB reduces dupliicates (Not completely) in a far better fashion than any other site I have seen. Yes it isn't perfect but the other user preference ways are far far more worse. 

Two thoughts.

1.  If a search system goes through every single record, retrieving a field and doing a computation, it might be fine for a personal database, but with 100 million records instead of 10,000, it'll take 10,000 times longer.  Database managers index key fields with hash tables.  The trick is to exploit this.

2.  The Virtual Jamestown site has a database of the Muster.  You can't browse it, there's only a search form.  The only way to find anybody by name is to know exactly how it was written in 1625.  So you can't find anybody.  Well you can, but the procedure is first to look them up in the Hotten book using the browsable alphabetical index at the back.
+6 votes
We are back to this again I see.  Bernard I'm in agreement that the whole LNAB field doesn't work, for many people around the world.

After reading a lot of the comments/answers in here, I see both sides of the "computing issues".  I also just spoke with my son who is a "coder"/website creator who builds massive sites from scratch.  He has explained a lot of the ins and outs of working on a wiki.  It's very different from "smaller websites".  wikis store every change made, and they build the pages as they are about to be seen when someone searches for one in particular, from the data that is stored.  Any small minute change that we request, requires large massive amounts of work in the coding, on a "non-operating wiki" of the wiki in question for the needed changes (if that makes sense).  So say the change of just a simple button to click for "no LNAB" can create errors throughout the system on many different levels of the pages, sorting, sending, routing, etc, thus the need to do the changes on a non-active wiki first.  Potentially depending on the numbers of coders working on the needed change (just one change) it can take 30 or more days to get the one change done and correct, and still there could be massive unexpected errors that just weren't foreseen during the new change-build.

I still think the LNAB field should never have been mandatory.  What I find completely appauling in here is that this was brought up just last week involving Native People who don't have last names and didn't less 200 years ago.  Some in here now seem to thing an "exception" should be made for those of "Aristocracy" Notables etc.... My question is If an exception should be made for these groups, why then should this same exception not be made for Native Peoples and Other Peoples all over the world that don't have and Never did have a LNAB?

I just pulled a lot of Native records from just 1885- 1903 there are massive amounts of names, families, children births, deaths, marriages etc.... But the way the system stands.... not even half these ppl can be added to wikitree.. Because I will Not add ppl using a Tribe affiliation as their LNAB.  There are ppl in these lists who not only have no last name at birth or any other time in their lives, but there are also people who their fathers 1st name because of census takers- fathers 1st name became the last name  & they are NOT 1 word euro names.

so my next question is.... So are  "other cultures" just to say..  "I thought this was supposed to be a  whole world tree, why don't "we" fit anywhere?

Now I know ppl in here will say.. Use the Tribe as LNAB.. guess what, the records I pulled have ppl who were given the last names of things like Ojibwa/Chippewa/Oglala  ALL of these are their "assigned last names".. what will the last name do in searches now for these, assign them to the "Sioux" or the Chippewa, or the Ojibwa  Tribe Names?  BUT they were not living in those Assigned Tribe boxes...they were on other Reservations.... here is where another mess begins.

sorry so long... just another viewpoint on this whole LNAB
by Arora Anonymous G2G6 Mach 6 (67.3k points)
+2 votes
Would people be happy if the box on the form were labeled "Last Name at Birth (or substitute)"?
by RJ Horace G2G6 Pilot (550k points)
edited by RJ Horace
"Family name at birth"
OK, "Family name at birth (or substitute)".

Maybe we should at least work something into https://www.wikitree.com/wiki/Help:Name_Fields#Last_Name_at_Birth

These are good constructive ideas CM and RJ
My ancestors didn't have family names.
I think the "or substitute" is probably at least the easiest solution.  But how would it appear on the viewable profile page?  Because if it would still look like a "last name", without something stating it as a substitute we are still left with the issue of others coming to wt now or in the future, stumbling on a profile of someone with no last name, as showing as if... Lakota, Dakota, Chippewa, is the last name. Which will cause more issues since Ive got ones to add, who have some to add new profiles that have names that include. Chippewa, Oglala, etc that are their last names- but because of the natv project standard to use "Tribe Affiliation" in the LNAB field.. how will someone know that

a. All these other profiles are using tribe affiliation in LNAB field & its not actually a last name of any of these existing profiles

b. new profiles with last names that are also actually "Ndn- Tribe Nation names, but are also the actual last names of some of these people ( will b added, if this kind of solution is workable).. how will viewers of these profiles know the difference?

c. Will they still have to literally click every individual LNAB on each profile to see if it takes them to a tribe, or to a cluster of ppl with the same last name?

Related questions

+11 votes
1 answer
+10 votes
5 answers
+4 votes
0 answers
+8 votes
3 answers
+6 votes
1 answer
111 views asked Jun 17, 2018 in Policy and Style by Pip Sheppard G2G Astronaut (1.8m points)
+8 votes
1 answer
+9 votes
1 answer

WikiTree  ~  About  ~  Help Help  ~  Search Person Search  ~  Surname:

disclaimer - terms - copyright