Two problems with places in Narrative Web Site

Help
1

I am currently running Gramps 6.0.0 from Debian package installed on Kubuntu 25.04, although these problems have existed for some time. I just hadn’t noticed earlier because I have over 17,000 places and I don’t generally check the report pages very closely.

The only commonality I can find among the places that have problems is that the examples I’ve found involve places where there are two places of the same name but different types enclosed by the same place. The cases I’ve been looking at are in the St. Louis, Missouri area, in particular, the independent city of St. Louis and St. Louis county, and the city of Normandy and township of Normandy in St. Louis county. The first problem is that the Places index does not include both places, although some times the pages for both places are generated. The other problem is that sometimes links for the places will appear on other pages, but when I click on the link I get a File not found error.

I have a report at Crider / McDowell Family Tree - Surnames that exhibits the first problem. There are pages for both the city and county of St. Louis, but only the city is included in the index. Likewise, there are pages for both the city and township of Normandy in St. Louis county, but only the township is included in the index. This report does not include unused places.

I only noticed the second problem this week in some reports I generated after generating the report above and at first I thought I had messed up something in my database. But lots of testing with my current database and a backup of the database used to generate the report above exhibit the problem every time although not always the same pages. I also went back to a report generated in March using Gramps 5.2.4 on Kubuntu 24.04 and it had File not found errors for the cities of Normandy and St. Louis. I’ve also been experimenting with including unused places in the report as I realized I had not understood it, but I still get the same problems. Interestingly enough, I did not see either problem in a few cases I checked where there are counties and independent cities of the same name in Virginia.

Is there anything that can be done with regard to the first problem without adding the type to the place name? Are there any suggestions to debugging the second problem?

2

Can you make bug reports for these two problems.
For each report add a small .gramps which reproduce the problem.

You can’t have two places with the same name and a different type at the same level in the place hierarchy.

3

I will try to put something together but it may take some time.
“You can’t have two places with the same name and a different type at the same level in the place hierarchy.” This appears to me to be an artificial imposition that you have added to the Narrative report. It ignores reality and is not enforced by Gramps. Missouri has both a county and an independent city named St. Louis. Virginia has a number of independent cities with the same name as counties in the state. I don’t know about other states, but I know there are numerous counties in Missouri that have townships and cities with the same name. In some cases the city is within the township and I enter them that way when I am aware of it, but there are other cases where I haven’t done enough research to determine the correct hierarchy and it is easier to leave them separate until I do the research. Honestly, I don’t want to have to become an expert on every place that appears in my database before I can add it.

4

I have submitted bug report #13841 and attached a very simple tree that demonstrates both problems. It has only two people, five events, and seven places.

2 Likes
5

No

For this example, you should have the St louis city enclosed by the St louis county.

I think this is the problem. If you don’t know the hierarchy, use different names.

I look at the bug report.

6

I have to agree with Serge that to use the Place Tree and Hierarchy you
have to give considerable thought to your structure to make best use of
GRAMPS.
Bitter experience has taught me this, at times I have spent as long
getting Places correct as I have adding new names to my tree.
It is worth spending the time on this.
phil

7

Independent cities are one of the US oddities for GIS.

Although surrounded by St. Louis county, the independent city of St. Louis is not part of the county administratively. It was separated from St. Louis county in April of 1877 and is now enclosed by the US state of Missouri. So both St. Louis independent city and St. Louis county are at the same level of administrative enclosure, differing only by ‘type’.

Besides the 38 (thirty-eight!!) Independent Cities in the USA state of Virginia, there are 3 in the rest of the USA: Baltimore, Maryland; St. Louis, Missouri; Carson City, Nevada. Both Baltimore and St. Louis are surrounded by (but not enclosed by) counties of the same name.

There is an even more confusing example: New York City. It is composed of 5 boroughs, each of which is enclosed by a county that enclosed nothing else. Manhattan borough is enclosed by New York county, enclosed by New York City, enclosed by New York state, enclosed by the United States of America.

8

I decided long ago to forget “administrative areas” these generally are temporary phases of existence dependent on the whim of politicians, civil servants and others.

So I stick with Geographical areas (yes I know these can change but I tend to use the historical ones).

So for Me

Gorton		is enclosed by Manchester 
   (problem here is that there are at 
    least 30 Manchester's all over the globe.
Manchester	is enclosed by Lancashire
Lancashire	is enclosed by England
England		is enclosed by United Kingdom

phil

9

From what I could see, England’s politicians tend to be more whimsical than many other countries.

Although if I tried to include USA census enumeration districts or voter districting in the Place hierarchy, then the incessant gerrymandering would make that an exercise in frustration.

10

Exactly Brian
Don’t get me started on Census Enumeration Districts or Registration
Districts/sub Districts
phil

1 Like
11

I tried to fix the problem. It’s not natural. To have a complete index, I need to add gramps_id to the name.
@Nick-Hall I think we should avoid having more than one place name enclosed by a place. When we add a place, the name of the place must not exist.

with the current problem, I think “Saint Louis” should be a unique name with type “Independent City” enclosed by Missoury.
For the “Normandy”, I think it is the same with type of “city” or “suburb”

12

You could add place type to the name. So, pair “Place name + place type” will be unique and probably fix the issue.

13

I don’t know if that’s true. There are territories that were traded back and forth by countries/empire throughout history.

The placename+type+timespan would probably be unique though.

14

From my database:

St Louis [P001223] City
St Louis [P015182] County

And to be clear, from 1 Oct 1804 to 22 Oct 1876, St Louis the City was a part of St Louis the County.

15

If you are going to say that the city of St. Louis is enclosed by the
county of St. Louis, you might as well say that any jurisdiction that
shares a border with another jurisdiction is enclosed by that
jurisdiction. Using an alternate name works as long as people are
accessing the places from links on other pages, but makes it more
difficult to find places in the index. Spelling out Saint Louis might
not be too much of a stretch for users, but what do you do with cases
like the city of Baltimore and the county of Baltimore? or the many
independent cities in Virginia? or many other examples that I could find
with a little effort?

Unfortunately for programmers, what seems natural to a programmer does
not always seem natural to others and doesn’t always fit reality. I saw
it many times in my more than 20 years of software development. How are
you building the index? In my database experience, it was always
possible to build an index that included two or more fields, whatever
was required for uniqueness. In this case, name and type would probably
make a good index and it would be useful to display both fields in the
index. Name and Gramps ID would be a better guarantee of uniqueness,
although there is no reason to display the Gramps ID.

Maybe you want to impose an unnatural definition of “enclosed by” on
Gramps? My objection to that is that most people who look at a
narrative web site are not going to be familiar with Gramps, or even
aware that it exists, and are going to be confused by terms that are
unique to Gramps.

In the case of Normandy, some additional research turned up that the
city of Normandy is inside the township of Normandy, so I was able to
fix that. My general practice has been to look up a place in Wikipedia
before I add it to Gramps and use the information I find there to choose
the enclosing place and coordinates. If I can’t find it there, I
usually turn to HomeTownLocator, but if I can’t find it on one of those
sites, I go ahead and add it with whatever information I’m given in the
source I’m using. There are very few sources I’ve been able to find
that provide the level of information that Gramps makes possible, so are
we supposed to leave places out when we can’t find enough detail?

Other situations that I have difficulty with and that reports don’t seen
to handle well are places that change types over time or that change
jurisdictions. I’m somewhat satisfied with what I’ve got for counties
in West Virginia by having them enclosed by Virginia before 1861 and by
West Virginia after 1861, although I haven’t checked what would happen
if I set the date on West Virginia as after 1861. Just one other
example that I’ve tried to deal with less successfully is Arizona which
was a territory before 14 Feb 1912 and a state after that date. I can’t
find any method to associate different types with different date ranges
for the same place, so I created two places with different date ranges
and different types. Of course, that is not handled well in reports either.

I’m thinking there is still considerable work to be done on the place
hierarchy and/or the documentation for it. There simply isn’t an easy
way to deal with the many changes over time that so many places have
experienced, and the ways I’ve tried don’t work well with reports.

16

Actually, I think it is even more complicated than that, although I
don’t know if Gramps is capable of dealing with the complete history and
I’m sure most reports can’t handle it. The county didn’t actually
become a county until 1 Oct 1812. Before that, it was apparently named
the District of St. Louis. The city gained municipal status on 9 Nov
1809 and city status on 9 Dec 1822. Missouri became a territory in 1812
and a state in 1821.

17

For the county, I have it named:

St Louis District from 1 Oct 1804 to 7 Dec 1812 and as St Louis Co after 7 Dec 1812

and Enclosed by
Missouri after 4 Jun 1812
Louisiana from 1 Oct 1804 to 4 Jun 1812

for the city, no name changes but enclosed by

Missouri after 22 Oct 1876
St Louis (the county) from 1 Oct 1804 to 22 Oct 1876
Virreinato de Nueva España from 1770 to 1800
Louisiana from 1760 to 1 Oct 1804

Louisiana goes through several name changes
including the Territory of Orleans from 1 Oct 1804 to 30 Apr 1812
Louisiana Purchase from 4 Jul 1803 to 1 Oct 1804
Louisiane from 1682 to 4 Jul 1803

and of course enclosed by the USA, CSA, and France.

Besides Wikipedia as a starting point for all places, I use..
Atlas of Historical County Boundaries
and the FaimlySearch Place Database

18

Here is the nub again the English Language to me Louisiana was never
“enclosed” by France it may have been a French Administered Territory
but was in a completely different area of the planet it is like saying
Australia or India were enclosed by the UK. Administratively Yes,
Geographically emphatically NO. Or like me saying Manchester was
enclosed by Italy when it was called Mancuniam back around 0 AD +/- 300
years .
We may need multiple Types of Enclosure, to cover the various scenarios
I prefer to stick with Geographical based on the “Gazetteer of
British Place Names” for UK, as the rest of the World is less than 5% of
my places I do not let that change my process.

19

:+1::+1::+1::+1::+1::+1::+1::+1::+1:

1 Like
20

My opinion: you have to consider that names are to be read and understood by humans. So giving a county and the townshipo therein the same name is going to confuse a lot of people, certainly if you would have omitted to place them in a hierachy in the Locations. Using the hierarchy is bound to clarify things (maybe not solve all the problems).
As far as changes in history occur, there is no straightforward way I think to solve this, apart from duplicating (or triplicating etc …) those locations.
Examples:
Dunkerque (Dunkirk - Duinkerke) in mediaval times was part of the county of Flanders, in 1659 it was assigned to the kingdom of England and in 1662 it became part of France. In my tree I have it once under France, although we have references to earlier times, for the sake of keeping things simple.
Borsbeek (in Belgium) was until 1/1/2025 a separate township, at that date it became a district of the city of Antwerp. Up to this moment I have no newborn family members there, but what if it happens??? I think people born before would object if I made them born in Antwerp…
I think no tool can make a decision for you what is the right way to handle such things.