Fixing Crewe as an Event Place

LegacyUser

Adrian Bruce said: Ok, this is going to sound selfish because this is my hometown but there are bits that I don't understand about how places and place names are assigned and I can't ask for the right thing for any place with any conviction if I don't have that understanding.

Take this Historical Record index on URL https://www.familysearch.org/ark:/619...

There is an "Event Place (Original):" and an "Event Place:"

Q1 - can someone please tell me what the difference is?

I have a suspicion that "Event Place (Original):" comes out of the indexing project but "Event Place:" is the result of a separate process that attempts to standardise "Event Place (Original):" and puts the answer in "Event Place:"

Q2 - whatever the answer to Q1 is, once "Event Place:" has been created, is it fixed, or might it get - somehow - rebuilt? If so, how often?

Now we come to the question of how an "Event Place (Original):" of "St Paul, Crewe, Cheshire, England" got turned into an "Event Place:" of "Crewe by Farndon, Cheshire, England, United Kingdom". I should add that both Crewe and Crewe-by-Farndon exist in real life - they are on opposite sides of the county - the former is the big railway town, the latter is a tiny village.

I would have expected that "Event Place (Original):" of "St Paul, Crewe, Cheshire, England" would be turned into an "Event Place:" of "Crewe, Cheshire, England, United Kingdom". That seems logical and the closest match.

However, the record in question applies to 1936 and the place for "Crewe, Cheshire, England, United Kingdom" actually has a place-name "Borough of Crewe, Cheshire, England, United Kingdom" for 1877 onwards.

If I think about trying to map "Event Place (Original):" of "St Paul, Crewe, Cheshire, England" to 1936 names, the choices will surely be "Crewe by Farndon, Cheshire, England, United Kingdom" and "Borough of Crewe, Cheshire, England, United Kingdom". Which is the closest? I'm not sure but "Crewe by Farndon, Cheshire, England, United Kingdom" has ended up in "Event Place". So...

Q3 - why is "Event Place:" set to "Crewe by Farndon, Cheshire, England, United Kingdom"? Does it have anything to do with the date of the event?

LegacyUser

Lundgren said: Q1:
Event Place (Original): is created by taking the Event Place: and other meta data from the collection by using the place standards service.

he system that populates the Event Place (Original): is to my knowledge, automated, and can make mistakes.

Q2:
To my knowledge, the Event Place: is not rebuilt or changed as it should be what was on the record.

The Event Place (Original): can change as the standards system and data is rebuilt.

Q3:
I don't have an answer to that, and am not a place in England expert. If it is in error, our data team to look at.

(The use of the word "Original" is occasionally lamented in meetings at familysearch...)

LegacyUser

Jeff Wiseman said: Great question Adrian, as I have wondered too.

I wonder sometimes if when these FS "meetings" occur (that Lundgren mentions), if anyone ever uses a dictionary :-(

You have a block of data indexed from some image somewhere and it contains 2 different versions of a similar thing called "Event Place".

- The first version contains the data that was indexed directly from the source.

- The second version was indirectly derived and reconstructed (possibly by computer) from various places in the source (i.e., it has NOT been indexed from the source in spite of the fact it is included with all of the data that WAS indexed from the source as though it was indexed data).

Using a dictionary definition of the word "original", which of the above would be the "original" source data. To help answer that, here is part of the dictionary meaning of the word original:

1. Present or existing from the beginning; first or earliest
2. Not dependent on other people's ideas

Hint. There appears to be no justification at all for using the word "original" on the second bullet above.

It makes sense that the reconstructed information from the metadata is flagged to show that there was "interpretive" logic applied. But calling it "original" serves only to imply the complete opposite and is totally inappropriate.

Sorta like taking Find a Grave index data and adding an invented burial date (based on a guess) to it and displaying it as though it was original indexed data from the source.

Indexed data is being displayed there. Why do we have data which is definitely NOT original indexed data being labeled as "original"?

When engineers seem to constantly struggle with issues like this that should be very obvious, it kinda disturbs me.

LegacyUser

Adrian Bruce said: Thanks Lundgren - I think.

You have confirmed that there is a rebuild process that goes on "as the standards system and data is rebuilt", which was one of the things that I was really interested in. (Clearly it would make corrections only half useful if there was no retrospective rebuilding, so this is good.)

Now, as for the rest, I really, really apologise for this - but I wonder if wires have got crossed? Your explanation and the use of the English language would make perfect sense if we swap "Event Place" and "Event Place (Original)".

I don't doubt from your words that you have explained the situation as you understand it - reference to lamentation in meetings demonstrates that. But I wonder if someone has changed something in the data loading of historical records and not told you guys...?

The record in question comes from film 2262737. The catalog shows the first 4 items on this film as:

"Items 1 - 4: Parish registers of St. Paul's Church, Crewe, 1869-1971
"Author: Church of England. St. Paul's Church (Crewe, Cheshire); Cheshire Record Office"

There is no mention there of Crewe-by-Farndon. The catalogue text, from the meta data, I presume, matches closely to the value in "Event Place (Original)". As it is in the meta data, it presumably should not change.

The value in "Event Place", which is "Crewe by Farndon, Cheshire, England, United Kingdom" bears no relationship to "what was on the record" other than having Crewe and Cheshire in. Note in particular the presence of "United Kingdom" which appears in the standardised place-names but seldom in the original data, suggesting that this is what comes out of an automatic process using the standards service and the "Event Place (Original)".

In other words, the values of "Event Place (Original)" and "Event Place" are swapped round from your explanation but, instead, match what your lamentations had (sensibly) wished for.

Can you possibly alert someone from the team who load the data and ask them if they are now loading data the "right way round" - the way you guys wished for, Jeff wishes for and the way I'd guessed it was? Or if maybe just this film was loaded that way round?

Again, I apologise - I'm not doubting your sincere belief but, knowing the data and places, everything is screaming at me that we're missing something - such as a change in loading.

LegacyUser

Lundgren said: I accidentally added a bit of confusion here!

If I could delete the post above, I would.

You can reverse all of the Event Place with the Event Place (Original) above. Sorry about that!

So to clarify:

Event Place is generated from Event Place (Original) and meta-data on the collection.

The Event Place can change as the place data and standards system changes.

The Event Place (Original) is preserved. (Even if there are user edits/corrections.)

If the interpreted places are coming back wrong, we can work to improve those. I will ask the data team to look at this instance.

You can see browse the standards place service in the link from the last post and get an idea for what is in the standards as well as why the results came back the way they did.

Sorry again about the confusion in my first post.

LegacyUser

Lundgren said: Please see my clarifying response to my own post above.

LegacyUser

Lundgren said: Looking for
St Paul, Crewe, Cheshire, England in the research site here , you can see that the first entry after 1936 is:
Crewe by Farndon, Cheshire, England, United Kingdom
Parish
1801-Today

If another place is missing you can suggest that place in the bottom left corner of the place research site.

Sorry again for the added confusion.

LegacyUser

Adrian Bruce said: Phew! Thanks for the clarification...

LegacyUser

Adrian Bruce said: Thanks - and thanks for the insights that (a) mapping "St. Paul..." is the thing to do and (b) that the date is relevant.

As for the "missing place", well, I have actually posted a suggestion in the places community but nothing's happened with it yet. It may be that the suggestion won't quite hit the mark - I don't know. That was partly why I raised this query here - to try to understand the downstream effects of place-name changes.

LegacyUser

Jeff Wiseman said: Lundgren, Thanks for the clarification!

However, when a person opens a citation to an index, it would be really great if data that was not indexed directly from the source would be partitioned away from calculated and/or derived data based on some programmers algorithm. A separate flag or asterisk identifying it as non-indexed data would be useful so that we can discern between actual indexed data and guessed at, or speculated data.

LegacyUser

Lundgren said: I can see how that would be valuable. There are additional changes coming to these results as well I believe.

Those make help to further clarify what comes from where. I am not on the team that is doing that work though, so can't really speak to what they are doing. I would watch the beta site and give feedback from that.