|
|
This came my way due to a a mistake, but the comment might very
likely be of interest on this list as well
-------- Forwarded Message --------
Since I don't see many posts yet this weekend, please excuse one
of mine which isn't exactly on charter. Feel free to argue me out
of posting in personal (offlist) email.
In a previous job I got to see databases made up by all sorts of
other people and organisations. Every time I saw a field called
'firstname' or 'second name' or 'surname' or 'familyname' I
groaned. So I was nodding along as I read this:
<https://www.kalzumeus.com/2010/06/17/falsehoods-programmers-believe-about-names/>
I think this one is unusually well-written.
In case you want to know how best to handle personal names, the
current consensus seems to be to use a single field containing the
whole name, which can be searched by substring. Computer systems
for places with non-Roman character sets sometimes use two fields:
name in local characters (Chinese, Devanagari, etc.) and name in
Roman characters.
Also note that current privacy legislation in the US and EU means
you are not allowed to ask for anything like 'full legal name'
unless you cannot run your business without it. Ask them for their
name, and store what they tell you, with the words in the order
they gave them. If you need to sort people in name order (think
very hard about why, first), create a field called 'sort order'
and populate it yourself. Sorting is your problem, not that of the
people you're sorting.
Part of a continuing series including falsehoods about dates,
times, places, street addresses, gender, relations, phone numbers,
taxes, and amounts of money.
Good luck, and watch your back.
_______________________________________________
sqlite-users mailing list
[hidden email]
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-usershttps://gramps-project.org
|
|
Heheh - nice list. I've worked on a few systems where certain people's names broke the software. And I've known other people whose names broke other software. A colleague of my father's for example, had only one name. No family name. No surname. When I worked at Dell, all employees had to be entered into the HR software. This HR software integrated with many other systems, some of which had limitations on the names that didn't fit reality. For example, one of the common Indian surnames is Lakshminarayanan. That name, which is actually quite common, was too long for some of the software. Sometimes logins are created from people's names, giving an extremely long login if multiple names are combined for the login. When I worked at VMware, I welcomed the opportunity to work on the part of the application that dealt with names, particularly when Asian names, especially Japanese names were used. This time, the constraint was what the exchange servers would accept. I did learn about various ways of saving names in Japanese, not just for people but also for companies.
I think I may quibble with one of the items on the list: "People’s names are all mapped in Unicode code points." While that may be true, I think it's false for USEFUL names in software. If you can't store it in unicode, can you store it at all? Perhaps the artist formally known as the artist formally known as Prince changed his name for this reason. :) At some point, you have to make a decision for how the name is to be stored. Maybe it's not the name, but it's the representation of the name used in the software.
Re: my own genealogy, most of the names in my tree are Norwegian. Until the 1900s, people in my tree didn't use family names, with just a handful of exceptions. People had one or more given names, and they used patronymics and farm names. The farm name changed if the person moved. The spelling of patronymics usually was abbreviated in documents, and spelling of given names was also inconsistent. So for my tree, I decided when I first started, after consulting with a few other people, to use a system that was consistent. Well, consistent within my tree. It's not consistent with other records. In my tree, for verified names, I use the earliest one I can find, usually a baptism record. I spell out the patronymic and append it to any given names. Most genealogy software doesn't handle Norwegian naming well, so those go in the first name field, and the farm name goes in the surname field. But what is the farm name? I use whatever was written in the earliest document I can find. In case there is more than one (multiple church books recording a baptism), I pick one. These can differ because a small farm can be part of a larger farm. Sometimes the small farm name is used. Sometimes the larger farm name is used. Sometimes the current farm is used. Sometimes the farm where the father was born is used.
This, of course, confuses genealogy software that makes certain assumptions, like that surname passes from father to child. That's a common rule that I wish was easier to turn off.
Victor
This came my way due to a a mistake, but the comment might very
likely be of interest on this list as well
-------- Forwarded Message --------
Subject:
|
[sqlite] Things you shouldn't assume when you store
names |
Date: |
Sat, 9 Nov 2019 19:25:58 +0000 |
From: |
Simon Slavin [hidden email] |
Reply-To:
|
SQLite mailing list
[hidden email] |
To: |
SQLite mailing list
[hidden email] |
Since I don't see many posts yet this weekend, please excuse one
of mine which isn't exactly on charter. Feel free to argue me out
of posting in personal (offlist) email.
In a previous job I got to see databases made up by all sorts of
other people and organisations. Every time I saw a field called
'firstname' or 'second name' or 'surname' or 'familyname' I
groaned. So I was nodding along as I read this:
<https://www.kalzumeus.com/2010/06/17/falsehoods-programmers-believe-about-names/>
I think this one is unusually well-written.
In case you want to know how best to handle personal names, the
current consensus seems to be to use a single field containing the
whole name, which can be searched by substring. Computer systems
for places with non-Roman character sets sometimes use two fields:
name in local characters (Chinese, Devanagari, etc.) and name in
Roman characters.
Also note that current privacy legislation in the US and EU means
you are not allowed to ask for anything like 'full legal name'
unless you cannot run your business without it. Ask them for their
name, and store what they tell you, with the words in the order
they gave them. If you need to sort people in name order (think
very hard about why, first), create a field called 'sort order'
and populate it yourself. Sorting is your problem, not that of the
people you're sorting.
Part of a continuing series including falsehoods about dates,
times, places, street addresses, gender, relations, phone numbers,
taxes, and amounts of money.
Good luck, and watch your back.
_______________________________________________
sqlite-users mailing list
[hidden email]
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-usershttps://gramps-project.org
|
|
On Saturday, 9 November 2019 14:11:54 PST Victor Engel, an eminent
manifestation of divinity, wrote:
> Heheh - nice list. I've worked on a few systems where certain people's
> names broke the software. And I've known other people whose names broke
> other software.
MY name breaks software.
legally, my name is Bruce Salamandir-Feyrecilde, which is too long for most
systems. people frequently enter "Salamandir" as my "middle name", but it's
not. i have no middle name. my "last name" is hyphenated, but most systems
don't take hyphens.
it is also not my "familiar" name (anyone who calls me "bruce" or "mr.
feyrecilde" is likely to be met with blank stares, as i haven't been "bruce"
for closing in on 40 years now...
and nobody knows how to pronounce "feyrecilde". most commonly i get "fair-
clyde" or something like that. hint: it's pronounced "fair-child". 😉
my name is salamandir. no capital letters, odd spelling, no "second name" or
"surname" or "family name" or anything like that. it's on my credit cards. the
only place that it's NOT is on my drivers' license and my passport.
but it's also not the name i was given when i was born, which is also on my
birth certificate. i have a birth certificate AND a certification of name
change, in case anyone asks, but nobody has in almost 50 years.
and the only people who have trouble with it are people who demand a "first
name" and a "last name". 😉
--
namaste
salamandir
[hidden email]
A: Because it breaks the logical sequence of discussion
Q: Why is top posting bad?
() ascii ribbon campaign - increased security
/\ - against html e-mail
- against proprietary attachments
--
Heute die Welt, Morgens das Sonnensystem!
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-usershttps://gramps-project.org
|
|
Victor,
I'm norwegian and I know the problem with Norwegian names all to well... What I started to do in Legacy familytree (the software is really bad regarding names and UTF-8), was to make Name Events (Custom Events) for all names and name changes I found, in addition to add the full name as an AKA. That way I could record the "full" name, the date of the document, the place and I could add a note to it... Luckily, Gramps support all this in its name editor, inclusive source sitation... ------------------
I still argue with other Norwegians about the patronymicon and farm names .... lol... even Norwegians don't understand that the farm name was more of a address than a name and shouldn't be places in the "surename" field...
I started asking Legacy familytree dev to change their name field 15 years ago, they still haven't managed to change it
And the problem is not only for Scandinavian names, but lot of other countries has the same or even more advanced structure...
Only software I have found, and I think I have tried everyone except TMG, that support most all name standards is Gramps...
-------------------------------------------------------------
------------------------------------------------------------- PS.
The list was great... I think I have been troubleshooting all of them in my time as a Windows
Technician and a Project Management Software
Technician...
Jaran
Heheh - nice list. I've worked on a few systems where certain people's names broke the software. And I've known other people whose names broke other software. A colleague of my father's for example, had only one name. No family name. No surname. When I worked at Dell, all employees had to be entered into the HR software. This HR software integrated with many other systems, some of which had limitations on the names that didn't fit reality. For example, one of the common Indian surnames is Lakshminarayanan. That name, which is actually quite common, was too long for some of the software. Sometimes logins are created from people's names, giving an extremely long login if multiple names are combined for the login. When I worked at VMware, I welcomed the opportunity to work on the part of the application that dealt with names, particularly when Asian names, especially Japanese names were used. This time, the constraint was what the exchange servers would accept. I did learn about various ways of saving names in Japanese, not just for people but also for companies.
I think I may quibble with one of the items on the list: "People’s names are all mapped in Unicode code points." While that may be true, I think it's false for USEFUL names in software. If you can't store it in unicode, can you store it at all? Perhaps the artist formally known as the artist formally known as Prince changed his name for this reason. :) At some point, you have to make a decision for how the name is to be stored. Maybe it's not the name, but it's the representation of the name used in the software.
Re: my own genealogy, most of the names in my tree are Norwegian. Until the 1900s, people in my tree didn't use family names, with just a handful of exceptions. People had one or more given names, and they used patronymics and farm names. The farm name changed if the person moved. The spelling of patronymics usually was abbreviated in documents, and spelling of given names was also inconsistent. So for my tree, I decided when I first started, after consulting with a few other people, to use a system that was consistent. Well, consistent within my tree. It's not consistent with other records. In my tree, for verified names, I use the earliest one I can find, usually a baptism record. I spell out the patronymic and append it to any given names. Most genealogy software doesn't handle Norwegian naming well, so those go in the first name field, and the farm name goes in the surname field. But what is the farm name? I use whatever was written in the earliest document I can find. In case there is more than one (multiple church books recording a baptism), I pick one. These can differ because a small farm can be part of a larger farm. Sometimes the small farm name is used. Sometimes the larger farm name is used. Sometimes the current farm is used. Sometimes the farm where the father was born is used.
This, of course, confuses genealogy software that makes certain assumptions, like that surname passes from father to child. That's a common rule that I wish was easier to turn off.
Victor
This came my way due to a a mistake, but the comment might very
likely be of interest on this list as well
-------- Forwarded Message --------
Subject:
|
[sqlite] Things you shouldn't assume when you store
names |
Date: |
Sat, 9 Nov 2019 19:25:58 +0000 |
From: |
Simon Slavin [hidden email] |
Reply-To:
|
SQLite mailing list
[hidden email] |
To: |
SQLite mailing list
[hidden email] |
Since I don't see many posts yet this weekend, please excuse one
of mine which isn't exactly on charter. Feel free to argue me out
of posting in personal (offlist) email.
In a previous job I got to see databases made up by all sorts of
other people and organisations. Every time I saw a field called
'firstname' or 'second name' or 'surname' or 'familyname' I
groaned. So I was nodding along as I read this:
<https://www.kalzumeus.com/2010/06/17/falsehoods-programmers-believe-about-names/>
I think this one is unusually well-written.
In case you want to know how best to handle personal names, the
current consensus seems to be to use a single field containing the
whole name, which can be searched by substring. Computer systems
for places with non-Roman character sets sometimes use two fields:
name in local characters (Chinese, Devanagari, etc.) and name in
Roman characters.
Also note that current privacy legislation in the US and EU means
you are not allowed to ask for anything like 'full legal name'
unless you cannot run your business without it. Ask them for their
name, and store what they tell you, with the words in the order
they gave them. If you need to sort people in name order (think
very hard about why, first), create a field called 'sort order'
and populate it yourself. Sorting is your problem, not that of the
people you're sorting.
Part of a continuing series including falsehoods about dates,
times, places, street addresses, gender, relations, phone numbers,
taxes, and amounts of money.
Good luck, and watch your back.
_______________________________________________
sqlite-users mailing list
[hidden email]
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-usershttps://gramps-project.org
|
|
On 09/11/2019 23:20, salamandir wrote:
[...]
> and the only people who have trouble with it are people who demand a
> "first name" and a "last name". 😉
Software designers who think they know personal names (and placenames)
might consider a course in onomastics
https://blog.oup.com/2016/03/name-studies-discipline/You cannot (reasonably) validate names, so it's pointless to try. You
can ask for a "called name" and optionally a "family/clan name" and
that's about all. By all means add optional fields for "other names".
You also cannot (reasonably) parse inside a name field. You can
certainly remove control characters (d000–d031 and a few others: see the
XML specification for examples) but otherwise all characters are
permitted, *including space*, and the only limitation should be that
they are all Unicode characters in whatever encoding you specify (except
the Private Use Areas, because there is no font for them).
Peter
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-usershttps://gramps-project.org
|
|
I extensively use the name editor. Names build up not only during
an individual life but also during a "family" History. They are
not carved in marble.
What is considered now as a family name started as a convenient
designator for a group of relatives,usually a "demonym" based on a
topological item (a road, a valley, a wood or forest, a bridge, a
ruined house), a physical trait (short, tall, red complexion, hair
colour) or a particular belonging (a donkey, a sword, …). This
name was not really fixed and could vary during lifetime. It might
even not be transmitted to descendants.
Things began to settle when clerks recorded scrupulously the
records, citing parents and in the best of cases grandparents
(though very often only male ancestry, women being considered as
wombs only -- times have changed). However litteracy of these
clerks is also a factor for variation. Another factor is
translation: records were written in Latin while people were known
in the common language and there is no one-to-one correspondence
between Latin names and common names!.
With time, some family branches became so "prolific" that
additions to the base family name were necessary to make a
distinction.
In my present geographical area and period of time of interest,
people were considered attached to a land and, without being
under serfdom status, very seldom moved, resulting in a high
ratio of endogamy. Also, children received only one name during
their baptism (at least until first third of XIXth century) and
this name was automatically the godfather's fro a boy or the
godmother's for a girl. I found only one exception duly
annotated "ex voto matris" in the record. This tradition lead to
many homonyms simultaneously living.
In some prolific branches it was necessary to append new demonyms
to remove ambiguities with "dictus", becoming a hyphen or even
dropped. But there is no rule and the connector could reappear on
clerk's initiative.
This is why I keep all variants I meet as alternative names with
the Name Editor, even subtle differences such as presence or
absence of hyphen. Thanks to the Citation system, evolution can be
monitored. I spotted thus how a woman changed its first name
(though immutable in principle because imposed on her at baptism
and consecrated) after her fifth child and this "call name" became
her only designation.
One item with which I am not at ease is the stauts of
nickname(s). They start as a personal additional name for a
specific person (such as "the elder" or "the nephew"). But they
happen to be transmitted and to become part of the "family" name.
I don't know exactly "when" to make the switch from the Nickname
entry box to a component of a "multi-name".
In presence of several name variant, my convention is to tag as
Preferred Name the one used in the oldest record for that person
(then in principle the birth record), even if this is not the same
as the birth name of the parents.
Gramps does a pretty good job in that I have not to artificially
choose a "canonical" name for an individual. It also contains a
very useful feature. The Name Editor contains a "Sort as" box. By
default, it is preloaded with the main part of the "family" name.
But its contents need not be related at all with the name itself.
Consider it as an arbitrary key for grouping people. This is were
you can enter a "canonical" name. It will not appear in the
various Gramps lists on screen but you can use any part of this
key in filters to cause display of the "tribe", no matter the
effective name variants.
This "canonical" name is used in NarrativeWeb report to group
individuals in the Surnames list. My regret is that the main name
subsequently does not appear after clicking on the surname. It is
then more difficult to disambiguate people with same first names.
Regards,
Patrick
Le 09/11/2019 à 23:11, Victor Engel a
écrit :
Heheh - nice list. I've worked on a few systems
where certain people's names broke the software. And I've
known other people whose names broke other software. A
colleague of my father's for example, had only one name. No
family name. No surname. When I worked at Dell, all
employees had to be entered into the HR software. This HR
software integrated with many other systems, some of which
had limitations on the names that didn't fit reality. For
example, one of the common Indian surnames
is Lakshminarayanan. That name, which is actually quite
common, was too long for some of the software. Sometimes
logins are created from people's names, giving an extremely
long login if multiple names are combined for the login.
When I worked at VMware, I welcomed the opportunity to work
on the part of the application that dealt with names,
particularly when Asian names, especially Japanese names
were used. This time, the constraint was what the exchange
servers would accept. I did learn about various ways of
saving names in Japanese, not just for people but also for
companies.
I think I may quibble with one of the items on the list:
"People’s names are all mapped in Unicode code points."
While that may be true, I think it's false for USEFUL names
in software. If you can't store it in unicode, can you store
it at all? Perhaps the artist formally known as the artist
formally known as Prince changed his name for this reason.
:) At some point, you have to make a decision for how the
name is to be stored. Maybe it's not the name, but it's the
representation of the name used in the software.
Re: my own genealogy, most of the names in my tree are
Norwegian. Until the 1900s, people in my tree didn't use
family names, with just a handful of exceptions. People had
one or more given names, and they used patronymics and farm
names. The farm name changed if the person moved. The
spelling of patronymics usually was abbreviated in
documents, and spelling of given names was also
inconsistent. So for my tree, I decided when I first
started, after consulting with a few other people, to use a
system that was consistent. Well, consistent within my tree.
It's not consistent with other records. In my tree, for
verified names, I use the earliest one I can find, usually a
baptism record. I spell out the patronymic and append it to
any given names. Most genealogy software doesn't handle
Norwegian naming well, so those go in the first name field,
and the farm name goes in the surname field. But what is the
farm name? I use whatever was written in the earliest
document I can find. In case there is more than one
(multiple church books recording a baptism), I pick one.
These can differ because a small farm can be part of a
larger farm. Sometimes the small farm name is used.
Sometimes the larger farm name is used. Sometimes the
current farm is used. Sometimes the farm where the father
was born is used.
This, of course, confuses genealogy software that makes
certain assumptions, like that surname passes from father to
child. That's a common rule that I wish was easier to turn
off.
Victor
This came my way due to a a mistake, but the comment
might very likely be of interest on this list as well
-------- Forwarded Message --------
Subject:
|
[sqlite] Things you shouldn't assume when you
store names |
Date:
|
Sat, 9 Nov 2019 19:25:58 +0000 |
From:
|
Simon Slavin [hidden email] |
Reply-To:
|
SQLite mailing list [hidden email] |
To:
|
SQLite mailing list [hidden email] |
Since I don't see many posts yet this weekend, please
excuse one of mine which isn't exactly on charter. Feel
free to argue me out of posting in personal (offlist)
email.
In a previous job I got to see databases made up by all
sorts of other people and organisations. Every time I saw
a field called 'firstname' or 'second name' or 'surname'
or 'familyname' I groaned. So I was nodding along as I
read this:
<https://www.kalzumeus.com/2010/06/17/falsehoods-programmers-believe-about-names/>
I think this one is unusually well-written.
In case you want to know how best to handle personal
names, the current consensus seems to be to use a single
field containing the whole name, which can be searched by
substring. Computer systems for places with non-Roman
character sets sometimes use two fields: name in local
characters (Chinese, Devanagari, etc.) and name in Roman
characters.
Also note that current privacy legislation in the US and
EU means you are not allowed to ask for anything like
'full legal name' unless you cannot run your business
without it. Ask them for their name, and store what they
tell you, with the words in the order they gave them. If
you need to sort people in name order (think very hard
about why, first), create a field called 'sort order' and
populate it yourself. Sorting is your problem, not that of
the people you're sorting.
Part of a continuing series including falsehoods about
dates, times, places, street addresses, gender, relations,
phone numbers, taxes, and amounts of money.
Good luck, and watch your back.
_______________________________________________
sqlite-users mailing list
[hidden email]
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-usershttps://gramps-project.org
|
|
On Sat, 09 Nov 2019 15:20:52 -0800
salamandir < [hidden email]> wrote:
Hello salamandir,
>MY name breaks software.
If software can't handle your name, then the software is already
broken.....
>and the only people who have trouble with it are people who demand a
>"first name" and a "last name". 😉
....apparently, by design.
--
Regards _
/ ) "The blindingly obvious is
/ _)rad never immediately apparent"
It's the age of destruction, in a world of corruption
Neuromancer - Billy Idol
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-usershttps://gramps-project.org
|
|
If your name breaks software then you were born to be a beta tester. -Brian On Sun, Nov 10, 2019 at 4:56, Brad Rogers On Sat, 09 Nov 2019 15:20:52 -0800 salamandir < [hidden email]> wrote: Hello salamandir, >MY name breaks software. If software can't handle your name, then the software is already broken.....
>and the only people who have trouble with it are people who demand a >"first name" and a "last name". 😉
....apparently, by design. -- Regards _ / ) "The blindingly obvious is / _)rad never immediately apparent" It's the age of destruction, in a world of corruption Neuromancer - Billy Idol --
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-usershttps://gramps-project.org
|
|
On Sunday, 10 November 2019 04:04:05 PST Emyoulation--- via Gramps-users, an
eminent manifestation of divinity, wrote:
> If your name breaks software then you were born to be a beta tester.
i was a tester for several years, before the computer industry crashed, in
2001... and then i had a brain injury...
--
namaste
salamandir
[hidden email]
A: Because it breaks the logical sequence of discussion
Q: Why is top posting bad?
() ascii ribbon campaign - increased security
/\ - against html e-mail
- against proprietary attachments
--
Know thyself? If I knew myself, I would run away.
-- Goethe
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-usershttps://gramps-project.org
|
|