Jump to content


Photo

County Check, Places, Geocoding et al


  • Please log in to reply
22 replies to this topic

#1 APerson

APerson

    Advanced Member

  • Members
  • PipPipPip
  • 267 posts

Posted 19 November 2012 - 08:44 PM

I must be the first kid on my block with RM6 and I'm VERY grateful that no new "changes" were made to the Edit Source windows (as that's the single most important thing as far as documenting genealogies is concerned). I really haven't had time to play with much else yet but it looks good. I have checked a couple of the more annoying things from previous versions - it now appears as if we can zoom in and out on media files by using the roller on the mouse (yipeeee!) but that very useful "hand" that you can grab and reposition media files isn't there yet (scroll bars must still be used). Also, it appears as if we still can't use the county checker if we include the word "county" in the place name. For example, in order to use the county checker, we must enter names as follows:


Erie, Erie, Pennsylvania

Los Angeles, Los Angeles, California


when the following is MUCH less confusing:

Erie, Erie County, Pennsylvania

Los Angeles, Los Angeles County, California


The problem when using the first approach (required by the County Checker) is that places are very nebulous when not all of the information is available (e.g., we don't know the city or county name [which might happened when a county name has been changed]). For example, what is meant by:

Erie, Pennsylvania

Los Angeles, California


Are these places refering to the city or county?


Anyway, it looks like RM6 has a lot of new and very good features - can't wait to try them out! I really like the research log as it looks VERY useful! Thanks! :D

#2 gerwally

gerwally

    Advanced Member

  • Members
  • PipPipPip
  • 104 posts

Posted 21 November 2012 - 07:05 AM

I currently have county check disabled in RM5 because you cannot enter the word county. It's too bad RM does not let you use the word county in the county check feature. Does anyone know why they will not let you do that?

#3 Renee Zamora

Renee Zamora

    Advanced Member

  • Support
  • PipPipPip
  • 8493 posts

Posted 21 November 2012 - 07:24 PM

Simply because the library we purchased with all the County Check data does not include the word County.
Renee
RootsMagic

#4 Paul Harris

Paul Harris

    Advanced Member

  • Members
  • PipPipPip
  • 172 posts

Posted 21 November 2012 - 07:39 PM

Bad purchase, encouraging bad habits.

#5 Jerry Bryan

Jerry Bryan

    Advanced Member

  • Members
  • PipPipPip
  • 3611 posts

Posted 21 November 2012 - 09:45 PM

Simply because the library we purchased with all the County Check data does not include the word County.


In that case,wouldn't it be possible internal to RM to strip the words County and Parish out of the place name before testing the place name?

Jerry

#6 Glenn

Glenn

    Advanced Member

  • Members
  • PipPipPip
  • 562 posts

Posted 22 November 2012 - 12:12 PM

Search for "_County," replace with "," if I understand your question. Then repeat the process using _Parish. Of course, remember not to use quote marks (") in the real search.

Underscore "_" = null space


Good real estate may be based on location, location, location;
but
good sources are based on citation, citation, citation!


Glenn


#7 TomH

TomH

    Advanced Member

  • Members
  • PipPipPip
  • 6268 posts

Posted 22 November 2012 - 04:27 PM

Search for "_County," replace with "," if I understand your question. Then repeat the process using _Parish. Of course, remember not to use quote marks (") in the real search.

Underscore "_" = null space

Basically, yes, but done on the fly in passing the Place value being tested to the county-implicit database. This can easily be done with a regular expression.

Tom user of RM7550 FTM2017 Ancestry.ca FamilySearch.org FindMyPast.com
SQLite_Tools_For_Roots_Magic_in_PR_Celti wiki, exploiting the database in special ways >>> RMtrix-tiny.png app, a bundle of RootsMagic utilities.


#8 Paul Harris

Paul Harris

    Advanced Member

  • Members
  • PipPipPip
  • 172 posts

Posted 22 November 2012 - 05:49 PM

This can easily be done with a regular expression.


I started using RM just before V5 came out last year and introduced the County Check Feature. For a solid year, observing this forum, the MOST ASKED FOR enhancement has been to allow the use of County Check without the word County flagging the need for a correction. A year later, a major upgrade to V6 does not include this request.

The next most often seen phrase over the last year on the Wish List has been, "Oh, I requested that over ____ years ago." New versions get new features that have not been requested, and old problems are ignored. From a development standpoint, I find this very discouraging.

Paul

#9 Jerry Bryan

Jerry Bryan

    Advanced Member

  • Members
  • PipPipPip
  • 3611 posts

Posted 22 November 2012 - 06:55 PM

Search for "_County," replace with "," if I understand your question. Then repeat the process using _Parish. Of course, remember not to use quote marks (") in the real search.

Underscore "_" = null space


I don't want to remove County or Parish from any of my place names. Since the problem that the Couny Check feature has with the words County and Parish seems to be with the County Check database, I was simply suggesting that RM itself remove County and Parish from an internal and temporary copy of my place names before doing a lookup in the County Check database.

Jerry

#10 APerson

APerson

    Advanced Member

  • Members
  • PipPipPip
  • 267 posts

Posted 22 November 2012 - 09:56 PM

Simply because the library we purchased with all the County Check data does not include the word County.


Shame on the library for using such shoddy documentation practices. This is a problem however, that should be very easy to resolve. Please permit us to use the word "county" in place names. Otherwise, this "feature" is useless and promotes bad documentation.

#11 Vyger

Vyger

    Advanced Member

  • Members
  • PipPipPip
  • 3436 posts

Posted 23 November 2012 - 04:06 AM

I don't want to remove County or Parish from any of my place names. Since the problem that the Couny Check feature has with the words County and Parish seems to be with the County Check database,


Being a non US user I accept I am probably not understanding this in as much detail as others but has the County Checker changed to allow "County" in the Place name?, see examples below.

At any rate I always maintained that any "Change to" value should be entered in the Standardized Place Name rather than the users preffered input format. This is what Gazetteer checks and uses on GeoCoding Places rather than replacing the users entry with a CONCISE entry which is beneficial for mapping.

I think it's a shame that the Rootsmagicians attempt to improve the quality of data has come under such fire and also believe if County Checker works on the Standardized Place Entry this problem could be overcome (maybe I am missing something)

Posted Image

Customers should never be frustrated by things they cannot do.

 

User of Family Historian 6.2.7, Rootsmagic 7.6.2, Family Tree Maker 2014 & Legacy 7.5

 

Excel to Gedcom conversion - simple getting started tutorials here

 

Root


#12 Jerry Bryan

Jerry Bryan

    Advanced Member

  • Members
  • PipPipPip
  • 3611 posts

Posted 23 November 2012 - 10:07 AM

Being a non US user I accept I am probably not understanding this in as much detail as others but has the County Checker changed to allow "County" in the Place name?, see examples below.

At any rate I always maintained that any "Change to" value should be entered in the Standardized Place Name rather than the users preffered input format. This is what Gazetteer checks and uses on GeoCoding Places rather than replacing the users entry with a CONCISE entry which is beneficial for mapping.

I think it's a shame that the Rootsmagicians attempt to improve the quality of data has come under such fire and also believe if County Checker works on the Standardized Place Entry this problem could be overcome (maybe I am missing something)


I am able to replicate your examples on RM6. I am also able to replicate your examples on RM5 (all upgrades applied). I do not have an RM4 system available for testing nor an early RM5 system. But it seems that RM does in fact accept the word "County" at the present time. I wonder if this is always the way it has worked, or if was changed at some point.

It is certainly possible that the problem with County Check has never been the presence of the word "County" but rather the absence of the words "United States" and that I and many others have been misinterpreting the cause of the problem. Here is precisely what I wish to be able to do:
  • Turn on County Check and leave it turned on at all times.
  • Enter a place such as "Knox County, Tennessee" and have it accepted without complaint by County Check provided that the event date is appropriate for the existence of Knox County and for the existence of Tennessee.
County Check will accept "Knox County, Tennessee, United States" without complaint, which I didn't not realize until I saw your e-mail. I feel sure that sometime in the past, I tried "Knox County, Tennessee, United States" and it failed. But that was a long time ago, and I could be wrong. :)

County Check will not accept "Knox County, Tennessee". It will not even accept "Knox, Tennessee". It insists on the inclusion of "United States", and it will accept either "Knox County, Tennessee, United States" or "Knox, Tennessee, United States".

I'm perfectly fine with RM using standardized place names in its internal operations such as Gazetteer and GeoCoding and County Check Explorer and such. But I find Standardized Place Names to be totally unacceptable for reports. Well, I'm not perfectly fine with one thing. I would like to be able to copy and paste a place name out of Gazetteer that is not standardized. As it is, I have to copy and paste "Wayne, Pennsylvania, United States" out of Gazetteer and then change it to "Wayne County, Pennsylvania" - not having any previous instances of Wayne County, Pennsylvania in my database.

The reasons why standardized place names are so unacceptable for reports has been discussed many times before. I may try to put together yet another summary, but for now would just close out by thanking you for pointing out that County Check does accept County (haven't checked on Parish yet for Louisiana), and that the problem lies elsewhere.

Jerry

#13 Jerry Bryan

Jerry Bryan

    Advanced Member

  • Members
  • PipPipPip
  • 3611 posts

Posted 23 November 2012 - 10:36 AM

I am able to replicate your examples on RM6. I am also able to replicate your examples on RM5 (all upgrades applied). I do not have an RM4 system available for testing nor an early RM5 system. But it seems that RM does in fact accept the word "County" at the present time. I wonder if this is always the way it has worked, or if was changed at some point.


Oops.

My testing was not sufficiently complete nor sufficiently subtle. County Check does not accept the word County, even in RM6. Not really.

I wanted a simple follow-up test with as few complications as possible. So I chose Hamblen County, Tennessee which I know well. It was created in 1870. So there are no issues about whether the United States existed yet in 1870 or whether Tennessee existed yet in 1870. The only issue is the creation date for Hamblen County itself.

County Check will raise an alert about any event for Hamblen County, Tennessee before 1870, which is correct. For example, trying to enter an 1860 census entry for Hamblen County correctly raises an alert.

So my follow-up test case is an 1880 census entry for "Hamblen County, Tennessee, United States". County Check raises an alert because of the presence of the word "County", and I wish it not to do so. If the test case is an 1880 census entry for "Hamblen County, Tennessee", then there are two reasons County Check will raise an alert: the presence of the word "County" and the absence of the words "United States". But I wish County Check to accept "Hamblen County, Tennessee" without alert for any date after 1870 because "Hamblen County, Tennessee" is the way I wish for the place name to appear in reports.

Jerry

#14 TomH

TomH

    Advanced Member

  • Members
  • PipPipPip
  • 6268 posts

Posted 23 November 2012 - 01:08 PM

Jerry, I'm not sure if your last paragraph reads exactly the way you intended. "Hamblen County, Tennessee, United States" does NOT raise an alert. NOR does "Hamblen, Tennessee, United States".

Both "Hamblen County, Tennessee" and "Hamblen, Tennessee" DO raise the alert and BOTH suggest adding "United States".

You could simply accept all these and use global search and replace to remove ", United States" from Places as a workaround.

Data Clean (Beta)'s Place Clean gives you the option of adding or deleting a country from Places and should prove even easier to use than search and replace for this specific task. It is sensitive to US states, Canadian provinces and, likely, UK countries (don't know about shires), and Australian states, i.e., it will add Canada to Nova Scotia but not to New York and vice-versa.

But I think County Check should be smart enough to tolerate the absence of the Country except when there is ambiguity.

Tom user of RM7550 FTM2017 Ancestry.ca FamilySearch.org FindMyPast.com
SQLite_Tools_For_Roots_Magic_in_PR_Celti wiki, exploiting the database in special ways >>> RMtrix-tiny.png app, a bundle of RootsMagic utilities.


#15 TomH

TomH

    Advanced Member

  • Members
  • PipPipPip
  • 6268 posts

Posted 23 November 2012 - 03:32 PM

I just realised the fly in my foregoing test:
"Hamblen County" in the string "Hamblen County, Tennessee, United States" is not treated as a county but as a city and its existence is not checked. Likewise for "Hamblen" in "Hamblen, Tennessee, United States". But for both occurrences where a city is included, they are checked, e.g.,
"Erstwhile, Hamblen County, Tennessee, United States" is caught and suggested that it be
"Erstwhile, Hamblen, Tennessee, United States".
"Erstwhile, Hamblen, Tennessee" and "Erstwhile, Hamblen County, Tennessee" are both caught and likewise suggested to be changed to:
"Erstwhile, Hamblen, Tennessee, United States".
"Erstwhile, Tennessee" is also caught and suggested to be
"Erstwhile, Tennessee, United States".

The fact that Erstwhile does not exist is immaterial to County Check. Gazetteer and GeoCode cannot find it, of course.

Tom user of RM7550 FTM2017 Ancestry.ca FamilySearch.org FindMyPast.com
SQLite_Tools_For_Roots_Magic_in_PR_Celti wiki, exploiting the database in special ways >>> RMtrix-tiny.png app, a bundle of RootsMagic utilities.


#16 Jerry Bryan

Jerry Bryan

    Advanced Member

  • Members
  • PipPipPip
  • 3611 posts

Posted 23 November 2012 - 09:00 PM

Jerry, I'm not sure if your last paragraph reads exactly the way you intended. "Hamblen County, Tennessee, United States" does NOT raise an alert. NOR does "Hamblen, Tennessee, United States".


Actually, the last paragraph did read the way I intended. But despite about an hour of what I thought was very careful testing, my interpretation of what was going on was not correct. Your observation that "Hamblen County, Tennessee, United States" does not raise an alert is correct, but see below.

I just realised the fly in my foregoing test:
"Hamblen County" in the string "Hamblen County, Tennessee, United States" is not treated as a county but as a city and its existence is not checked. Likewise for "Hamblen" in "Hamblen, Tennessee, United States". But for both occurrences where a city is included, they are checked, e.g.,
"Erstwhile, Hamblen County, Tennessee, United States" is caught and suggested that it be
"Erstwhile, Hamblen, Tennessee, United States".
"Erstwhile, Hamblen, Tennessee" and "Erstwhile, Hamblen County, Tennessee" are both caught and likewise suggested to be changed to:
"Erstwhile, Hamblen, Tennessee, United States".
"Erstwhile, Tennessee" is also caught and suggested to be
"Erstwhile, Tennessee, United States".

The fact that Erstwhile does not exist is immaterial to County Check. Gazetteer and GeoCode cannot find it, of course.


Armed with Tom's examples, I've been trying to abstract all the rules that County Check is using. I'm by no means sure I've found all the cases. Here's some further analysis.
  • "Hamblen County, Tennessee" alerts in all cases, suggesting adding ", United States" at the end. There is no suggestion to remove the word "County". There is no distinction between a blank date, a date of 1860 (which should be an invalid date), and a date of 1880 (which should be a valid date). Tom's interpretation that the string "Hamblen County" is not interpreted as a County name in this case seems to be correct. The string "Tennessee" seems sufficient to cause the alert suggest adding ", United States" at the end. I'm dying to see what happens with the string "Georgia" which is both a state in the United States and a country in Eurasia.
  • "Hamblen, Tennessee" alerts in all cases, and seems to be acting just like case #1 and for the same reasons. As Tom says, the string "Hamblen" appears not to be being interpreted as a County name seems to be correct.
  • "Hamblen, Tennessee, United States" never alerts, even when the date is 1860 which is before Hamblen County was created. This suggests that "Hamblen" is not being treated as a County name. However, see #4 below and #8 below.
  • "Erstwhile, Tennessee, United States" always alerts with any date or a blank date. So County Check is trying to treat Erstwhile as a County, fails to find such a County, and alerts. So in #3, it must be recognizing that Hamblen is a valid Tennessee county and it doesn't alert on the county name. But it fails to alert on the invalid date of 1860 for Hamblen County.
  • "Erstwhile, Tennessee" alerts no matter what date is entered and wants me to add ", United States".
  • "Tennessee" alerts or not depending on what date I enter. But the alert seems to depend on whether or not the United States existed at the time, not on whether Tennessee existed at the time. For example, Tennessee in 1750 alerts and says that the United States did not exist yet. But Tennessee in 1785 (after the United States existed) does not alert even though Tennessee was still a part of North Carolina until 1790 and did not become a state until 1797.
  • "Erstwhile, Hamblen, Tennessee" always alerts. If the date is after Hamblen County was created, it wants me to add ", United States". If the date is before Hamblen County was created, it warns me of the invalid date.
  • "Erstwhile, Hamblen, Tennessee, United States" seems to alert or not based on the correctness of the date with respect to the creation of Hamblen County, the creation of Tennessee, and the creation of the United States.
  • "Erstwhile, Hamblen County, Tennessee" always alerts. If the date is valid for Hamblen County, it wants me to add ", United States" at the end and remove the word "County". If the date is not valid for Hamblen County, it wants me to choose an adjcent county. So it understand that this really is the Hamblen County in Tennessee.
  • "Erstwhile, Xxxxxxx, Tennessee" always alerts no matter the date and tells me that Xxxxxx was never a Tennessee county. It doesn't say anything directly about adding ", United States", but it does offer me a menu of valid Tennessee counties of the form "valid county name, Tennessee, United States".
  • "Erstwhile, Georgia" wants me to add ", United States". But what if "Erstwhile" were a city in the country of Georgia in Eurasia? The country of Georgia is not a country in County Check (fair enough), but there appears to be no way to prevent references to the country of Georgia from alerting except to turn off County Check.
I was hoping to come up with a coherent set of rules of what works and what doesn't, but I don't think I've done so. The list above is pretty incomprehensible to me. I would like to say the following: County Check really wants to see a United States place of the form "city, county, state, United States". The county and state are checked against each other and the date, and the city is not checked. However, the presence of the city (even a totally incorrect city) seems to be necessary to make date checking work correctly in all cases. But even that summary doesn't quite seem to be a complete and accurate account of how County Check is working.

Jerry

#17 Jerry Bryan

Jerry Bryan

    Advanced Member

  • Members
  • PipPipPip
  • 3611 posts

Posted 25 November 2012 - 09:10 AM

Being a non US user I accept I am probably not understanding this in as much detail as others ........

I think it's a shame that the Rootsmagicians attempt to improve the quality of data has come under such fire and also believe if County Checker works on the Standardized Place Entry this problem could be overcome (maybe I am missing something)


I agree that it's unfortunate that County Check has come under so much fire, even though I'm one of the sources of that fire. :) I think your excellent comments certainly deserve some additional discussion. Much of these questions have been asked and answered before (and by many others, not just by me), but I thought I would make an attempt at consolidating some of the ideas. So, ....

If users are so upset with County Check, why don't they just turn it off and quit complaining? I can only speak for myself, but I really like the concept of County Check and I really do want my place names checked. The problem is that to make County Check quit alerting about my valid place names, I would have to adopt Standardized Place Names as the way place names are stored in my database. And having accepted Standardized Place Names as the way place names are stored in my database, I would have to accept Standardized Place Names as the way place names are printed in reports. I cannot abide the way reports appear if all the place names are Standardized Place Names. So I do turn off County Check, and therefore I lose out on all of its benefits.

Standardized Place Names do not include the words "County" or "Parish". Why are users who are so upset about County Check so insistent on including the words "County" and "Parish" in their place names. Again, I can only speak for myself. It seems to me that there are actually several different reasons.
  • The most commonly cited reason and the reason based most on logic is that without the word "County" or "Parish", many place names are ambiguous. Many counties have the same names as other political subdivisions in the same state - the city of Los Angeles in California and Los Angeles County, California, for example. The city of Los Angeles is in Los Angeles County, but there are other cities in Los Angeles County. And there are many cases where the same named political subdivision is not even in the same county that is its namesake. The forum archives are full of such examples.
  • There is another much less commonly cited reason that is based less on logic than reason #1 and which I suspect comes closer to being the "real" reason people want to keep the words "County" or "Parish" in their place names. Namely, including the words County or Parish is the way place names are used by normal people every day in common usage. For example, I live in Knox County, Tennessee. Nobody speaks of the "Knox Schools" or the "Knox Courthouse". Everybody speaks of the "Knox County schools" or the "Knox County Courthouse". I live very near the Anderson County line. If somebody asks me which county I lived in, I would almost never say that I live in "Knox". I would say that I live in "Knox County". So it is very jarring to see in a family history report created by RM that a couple was married in "Knox, Tennessee" rather than in "Knox County, Tennessee". Nobody says "Knox, Tennessee" in the real world, and family history reports that are manually created (with typewriters in the old days, with word processors these days) are not written that way.
  • Some implementations of Standardized Place Names get into what I call "counting commas". RM doesn't really go quite that far, but I certainly cannot abide a family history report saying that somebody was born in ",, Tennessee, United States" or that they were married in ", Knox, Tennessee, United States". The business of the extra commas comes into place when an implementation of Standardized Place Names becomes so rigid as to insist on providing a place for every possible political subdivision, including those that are unknown or are otherwise omitted. Such rigid implentations of Standardized Place Names would, for example, write the city of Oliver Springs, Tennessee as "Oliver Springs,, Tennessee, United States" if the county were not known, and as "Oliver Springs, Anderson, Tennessee, United States" or "Oliver Springs, Roane, Tennessee, United States" or "Oliver Springs, Morgan, Tennessee, United States" if the county was known (Oliver Springs straddles three counties). And then the counties alone if the city were not known would be ", Anderson, Tennessee, United States" or ", Roane, Tennessee, United States" or ", Morgan, Tennessee, United States". As I said already, I simply cannot abide such names in reports. I think that "Oliver Springs, Tennessee" is just fine if the county is not known, and that "Oliver Springs, Anderson County, Tennessee" is just fine if the county is known. And I think that "Anderson County, Tennessee" is just fine if further political subdivision is not known or is not applicable. Indeed, sometimes there is not a further applicable politial subdivision and writing ", Anderson County, Tennesee" is like the sound of fingernails on an old fashioned chalk board.
Standardized Place Names do include the words "United States" for place names in the United States. Why are users who are so upset about County Check so insistent on omitting the words "United States" in their United States place names. Again, I can only speak for myself. It seems to me that the basic reason is the same as reason #2 above for including the words "County" and "Parish". Namely, Americans in everyday speech typically do not say that they were born in "Tennessee, United States" or that they were married in "Knox County, Tennessee, United States". They would typically just say that they were born "in Tennessee", and as long as the state was clear from context they would typically just say that they were married "in Knox County" - (and certainly not "in Knox" :) ). And in addition, including the words "United States" for place names in reports quickly makes the reports read in a fashion that is extremely redundant and ponderous sounding.

I will grant that including the words "United States" does not introduce any ambiguity into a place name as does omitting "County"or "Parish", quite the contrary. It might sometimes be necessary to include the words "United States" to distinguish the state of Georgia in the United States from the country of Georgia in Eurasia. But normally a state name is sufficient to disambiguate a place name, even for folks who are not from the United States. It was certainly my experience living and traveling in Europe that everybody understood American state names. People would ask me where I was from, I would say "Tennessee", and they would know exactly what that meant. In fact, they usually had already figured out that I was an American from my accent and from my clothing and from my body language, so when they were asking me where I was from they really were asking me "which state?". (Well, folks in Europe typically knew "exactly" that Tennessee was an American state. But they often didn't know "exactly" where it was beyind the fact that it is "somewhere in the middle".)

So what's the solution? Well, even if you don't use Standardized Place Names it is already the case that RM is storing Standardized Place Names for you along side your preferred place names. I think Vyger has it right on with his suggestion that County Check might check against the Standardized Place Name rather than against your preferred place name, and that an alert might be raised only if County Check detects a problem with the Standardized form of your place name instead of complaining about the preferred form of your place names. Or maybe there could be three options instead of two - check, don't check, and check only against the standardized form without alerting on the user's preferred form.

Jerry

#18 Vyger

Vyger

    Advanced Member

  • Members
  • PipPipPip
  • 3436 posts

Posted 25 November 2012 - 11:49 AM

A long and very well considered post and I believe a very valuable discussion.

Presently the Standardized Place Name does not seem to have an option in reports or sentencing so again, at the risk of repeating myself, I believe the long winded but concise Place Name should populate that field leaving the users preferred input alone.

Also Like Gazetteer, County Check should apply it's check to the Standardized Place Name and populate it in the case of an acceptance.

Now to sentencing:

I cannot get the Standardized Place Name to appear anywhere in reports so I am assuming it's purpose is to be a recognisable entry for Online Mapping and in this suggestion County Checker.

Both the Place (users default input) and the Abbreviated Place can be used in sentencing so in this case:

Place (user input) - Bedlington, Northumberland, England
Standardized - Bedlington, Northumberland, England, United Kingdom
Abbreviated - Bedlington, Northumberland

< [Place:First]> gives Bedlington
< [Place:Short]> gives Bedlington, Northumberland (the abbreviated place field)
< [Place]> gives Bedlington, Northumberland, England

I cannot find an option to produce Bedlington, Northumberland, England, United Kingdom, the Standardized Place Name, So I don't think users should be too concerned about what is in that field only that is recognised by Online Mapping services and County Checker.

Now you see someone is going to find a way to get Standardized Place Name to appear in reports :D

Bottom line is that if County Checker was looking at the Standardized Place Name in the example above it would not keep nagging me to change Bedlington, Northumberland, England to Bedlington, Northumberland, England, United Kingdom

Customers should never be frustrated by things they cannot do.

 

User of Family Historian 6.2.7, Rootsmagic 7.6.2, Family Tree Maker 2014 & Legacy 7.5

 

Excel to Gedcom conversion - simple getting started tutorials here

 

Root


#19 TomH

TomH

    Advanced Member

  • Members
  • PipPipPip
  • 6268 posts

Posted 27 November 2012 - 04:54 PM

I am really transfixed by this discussion and the quality of the analysis and writing. The attempt to consolidate prior threads is commendable and I would like to add some others that bear on the issue. I'll set aside the consideration of County as that, I think, can be readily handled by a filter between the values entered and County Check, allowing a user freedom to include or not such words or abbreviations as "County, Co., Township, Twp...." In my own somewhat muddled thinking, I keep wondering if we are using fewer fields than are needed so as not to mix up the purposes and end result. Certainly, as it stands, the structure does not work well for everyone. I think what we have as data are:
  • Historical Place Name - this is what County Check should use or fill.
  • Current Place Name - this is what Geo-Code, Gazetteer and On-line Map should use or fill.
  • Vernacular Place Name - this is what the place is/was commonly known as.
  • Is there more?
Note that I have not used the RM terms Place, Standardized and Abbrev, although there might appear to be a 1:1 correspondence. The problem from a database design is that the Current or Standardized Place Name can have multiple Historical Place Names, and maybe multiple Vernaculars. Data normalization would dictate a different structure than the single table currently used. A site (Place Detail) may have gone through several changes in the Place it is situated, i.e., it belongs to multiple Historical Place Names but only one Current Place. The current structure requires a separate record for each Place + Standardized Name combo plus a record for its Place Detail; if there were 4 changes in the Historical place name, then there are 5 sets of two records to be created and maintained. If, instead, there was one record for the Current Place, one record for the Place Detail belonging to that Place and multiple records for the Historical Place Name belonging to that Place, the data for Current Place and Place Detail need be entered and maintained just once.

For Vernacular, a general solution would be to support one for each Historical Place and one for the Current.

Then we have outputs:
  • Currently for reports and websites, we only get Place and Abbrev. A problem with that is if it is the Historical value that is used in Place, it may not be relevant to current maps and GPS systems. OTOH, there are problems with the Standardized Place, the modern name not suited to the historical fact. In some sentences or even table reports, it would be desirable to have both.
  • Other?? (this is stream of consciousness that started a couple of days ago and was interrupted, now off-track!)
I think this argues for some change in database design, definitions, procedures and sentence template variables; this list is by no means exhaustive:
  • Replace the current Place, Standardized Place, Abbrev with Historical Place, Current Place and Vernacular, replace Place Detail with Site where Current Place supports many Historical Places with Historical Vernacular and a Current Vernacular. A Site relates to one or more Historical Places and/or to a Current Place.
  • County Check operates on Current Place and event date, manually from the fact edit pane in addition to the automatic option as at present. The manual check basically launches County Check Explorer from the fact edit pane with the Current Place value automatically entered.
  • County Check tolerates the inclusion of "County, Borough" and other common political entity names.
  • Somehow, make available the selection of the Current Place an option with a drop-down list of the Historical Places used from which one is chosen. If the list includes dates, it's possible this process could make an automatic offer which the user could take or change.
  • Enhance the sentence template variables to support modifiers :Historical, :Current; :second, :third, :one, :two, :three, :Vernacular.
I'm sure this is a flawed proposal but maybe adds something to the considerations. I'm too blotto to continue...

Tom user of RM7550 FTM2017 Ancestry.ca FamilySearch.org FindMyPast.com
SQLite_Tools_For_Roots_Magic_in_PR_Celti wiki, exploiting the database in special ways >>> RMtrix-tiny.png app, a bundle of RootsMagic utilities.


#20 Jerry Bryan

Jerry Bryan

    Advanced Member

  • Members
  • PipPipPip
  • 3611 posts

Posted 02 December 2012 - 10:40 AM

  • Historical Place Name - this is what County Check should use or fill.
  • Current Place Name - this is what Geo-Code, Gazetteer and On-line Map should use or fill.
  • Vernacular Place Name - this is what the place is/was commonly known as.
  • Is there more?


I confess I had not fully considered the impact of all this on Geocoding and Mapping. I may not be thinking about it entirely correctly, but I wonder if Gazetteer really belongs in your item #2.

But in more general terms than just whether or not Gazetteer belongs in your item #2, consider the following. As is sometimes the case, this is slightly easier to describe if I use the names of real people. My maternal grandfather was Lester Peters. The Peters family moved to Tennessee from Greenbrier County, Virginia in 1797. Tobias Peters was married in Greenbrier County in 1797 shortly before the family moved to Tennessee, and Henry Peters Sr. was married in Greenbrier County in 1788. It has not been possible thus far to determine for certain who Henry's parents were nor Tobias's, but DNA is consistent with them having been brothers or at least closely related. The family was staunchly Methodist, and both men were married in the Rehoboth Church in Greenbrier County. The church is still standing and is the oldest church west of the Allegheny Mountains. It was built in 1784 and was dedicated in 1785 by the Methodist circuit rider Bishop Francis Asbury. Bishop Asbury had founded the Methodist church in Greenbrier County.

So what does any of this have to do with RM? Well, we start with the fact that Greenbrier County was split in 1799 with the northern part continuing to be called Greenbrier County and the southern part being called Monroe County. The Rehoboth Church is in Monroe County, so there is immediately a geocoding challenge that the church was built in one county and is now in another county, and the church has never moved. Also, there are geographic features called Peters Mountain and Peterstown which are both in Monroe County rather than in Greenbrier County. I don't yet have any early deeds Greenbrier County (don't even know yet if they are still extant), but there has to be an extremely strong suspicion that my ancestors lived in what is now Monroe County rather than in what is now Greenbrier County. To make matters even more complicated, the state of Virginia was split during the Civil War to create West Virginia and both Monroe County and Greenbrier County are now in West Virginia. Suffice it to say that RM's geography tools do not deal with this situation very gracefully. And in fairness, I'm not sure I expect RM's tools to be able to deal with Greenbrier County, Virginia. It's an extremely complicated and convoluted problem. But RM will not enter a standardized place name for Greenbrier County, Virginia even if for testing purposes I enter the name as fully standardized already and even though I'm giving a valid date when Greenbrier County was still part of Virginia.

But also in all truth, I guess that in a way I don't "get it" with respect to why geocoding and mapping are so important in RM. I do "get it" why County Check and Gazetteer are so important in RM. But geocoding and mapping do not result in any data that appear in narrative reports. I geocode lots of things, including individual tombstones and old family cemeteries and old family homesteads. But I have to put the GPS coordinates in notes in order to get them to print. There really isn't anything that RM can do to assist this situation. And when I do wish to do mapping with GPS coordinates, I much prefer Google maps to Bing (and I totally understand why RM needs to use Bing). So I just do all my own mapping by hand. (I use Acme Mapper 2.0 rather than using Google maps directly. Acme Mapper 2.0 is really Google maps under the covers with a new skin. The Acme Mapper 2.0 skin is much more useful for genealogical mapping than is the standard Google maps skin).

Jerry