I started out trying to be a lumper, but the source and citation structure in RM makes that difficult. The Source Text field is at the Master Source level, so if you want to make a generic source for the 1940 census, you would have to transcribe the entire country’s worth of census entries. Obviously not practical for the census, but it works fine for an obituary.
I never thought of it that way, but @thejerrybryan’s advice about an image being a source is a good rule of thumb. If I have a census page with a family, though, I make the source at the family level and transcribe their info in the Source Text box. I also put the image in the Master Source section. I then make a separate citation entry for each member of the family, which is pretty easy.
Thank you! I looked at your site a little bit the other day and really appreciate you sharing the way that you do. I already adopted your suggestion of the Child/Parent Fact. I was wondering how to document the proof that a particular child is actually related to a specific set of parents, and this is a great way to do it. I’m thinking about your idea of storing media files in GitHub but want to discuss that with my husband (a developer) to see what he thinks since my knowledge of Git is very limited.
Thanks again for sharing so openly. It seems like one of the best parts of using RootsMagic may be the very helpful user community.
I like that idea. If you download the census page as a media file, do you put it at the source or citation level? I’m currently putting them all at the citation level, but then it’s three clicks to drill down just to see if there is one.
I’m not sure if this question is directed to me or to someone else. But either way, I think there is something that needs to be pointed out. Philosophically, an extreme source splitter such as me would never even use citations at all. They would just put every bit of their sourcing data into the source level and not even have any citations. But there is one very important piece of “sourcing data” that has to be at the citation level, even for the most extreme of source splitters. Namely, it is only a citation that can be linked to a person or to a fact. A pure source splitter might prefer just to ink the source to the person or fact. But that’s not the way the data model works. The source has to go through the citation to get to the person or fact.
And then there is another piece of “sourcing data” that can only be at the citation level - namely the proof data. A death certificate might be cited for the Death fact and for the Birth fact. But a death certificate is likely to be at a different proof level for the Death fact than it is for the Birth fact. So the proof data has to be where the link is located, which is at the citation level. I don’t use RM’s proof data because it doesn’t seem to me actually to do anything, but some RM users do rely heavily on using RM’s proof data.
But in answer to the question about whether I link a census image to the source level or to the citation level, even as a source splitter I link it to both. It really isn’t very logical because as a source splitter the image belongs at the source level. But I like the way the source images looks at the citation level when I publish via GedSite. Even with GedSite, I couldn’t see the uses of the images all on one screen if I didn’t link the images to the citation level.
The same question exists for transcriptions. I transcribe from outside of RM, mostly using Notepad on Windows. I think it’s very hard to transcribe directly into RM. I then copy and paste my transcriptions into both the source level and the citation level in RM. For an extreme source splitter, it makes logical sense to enter transcriptions at only the source level. But again, I paste transcriptions to both because of the way they look in GedSite. And even importantly, there is an option in RM to print the transcriptions in reports along with the endnotes from the citation level but not from the source level.
If/when I switch to source lumping, my transcriptions and my media files would go only to the citation level. This would not change my overall philosophy that 1 media file = 1 citation.
Thank you @thejerrybryan , that was very helpful, and I’m really interested to learn from as many people as possible about what works for them.
I definitely agree with you about transcribing and I ended up doing the same thing kind of accidentally. I opened a PDF file (which of course opened in my PDF reader and not in the document viewer) and had to figure out what I could use to transcribe it. Notepad has always been a favorite for me because it doesn’t add formatting, so copying and pasting is easier.
I was wondering if I could just skip the citation entirely, but as you point out, that doesn’t work because you can’t attach a source directly to a person or fact. I like the idea of attaching the media file at both levels, though, because it is a pain having to drill all the way down into the citation to see if a media file exists.
Thank you again for the help in figuring this stuff out. Given that I’d eventually like to also at least upload to Ancestry (and possibly FamilySearch but that’s less of a need), knowing where all this data is going could end up being a factor as well. Do you happen to know if there’s any documentation on which information goes where when you upload? After many years as a business analyst helping define interfaces, I almost want to just see the data mapping but that seems like too much to ask. (And it’s probably proprietary anyway).
Anyway, lots to think about and I’m very grateful for everyone who has commented.
I put the census document at the source level since it is the same for all citations, but I can understand putting it in the citation to save a click. Sometimes I’m tempted to just put it in the event.
Regarding transcriptions, doing them is an external editor like Notepad is a good idea. I also found I can get decent first drafts of transcriptions by uploading the image to an AI chatbot like Claude or ChatGPT and asking it to do the transcription. I still have to edit, but not as much as I would have typed. It doesn’t always work but it does enough of the time to be useful.
@beirne thanks for the suggestion. I have been typing them manually but I might give the AI a try. I tried out Abby FineReader recently for a very long document - a family history that my great-aunt wrote - and it did a pretty decent job but it’s $100 a year. I don’t mind paying for software, but that’s more than my budget allows for a hobby.
Anyway, thanks again for joining the conversation and helping me out with suggestions!
@nkess , Windows 11 but I tend to disable that sort of thing just out of habit. I worked in IT for decades and learned to be skeptical about things like AI. That said, some of the newer tools are providing some decent results for family historians so now that I’m retired and don’t have to worry about compatibility with the software I’m testing, I might give it a try. Thanks for the suggestion!
RM does not publish a doc but it’s a simple task to create a test event with a citation using the Ancestry Record template and populate each template field with the RM field name, then treeshare the event to ancestry, and view the source details. I suggest adding an event note as well and some formatting as part of your test. It’s easy to see which fields upload, what gets lost, and how fields are mapped. Once you are comfortable with how the Ancestry Record template maps, then it’s easier to wrap your head around how other templates map. You’ll see that formatting gets stripped out by ancestry, so anything fancy tends to get lost.
btw, in case you are not doing so already, using slide in workflow makes editing a little less clicky. In addition to enabling the setting, you need to click on the event image (pen for citations) to engage the workflow.
I think that being a source lumper on RM can be very effective, but you have to think about it a little differently. The source (called the Master Source in RM) would be something like the entire U.S. Federal Census for all years. Or the entire census for 1940. Or the entire census for 1940 in Tennessee. Or the entire census for 1940 in Anderson County, Tennessee. The source would contain no transcriptions and no media.
Then things like transcriptions would go into the Research Note field which is specific to the Source Details (essentially, to the citation). Images would be linked to the Source Details (i.e., to the citation).
In concept, it’s a good system, and it operates very much in accord with the way sources and citations are stored in GEDCOM files.
I do a lot of genealogy research but have limited time. Being a lumper allows me to tell the readers of my material where to find my sources, but allows me make the sourcing quick and easy which allows me additional time for researching rather than documenting. One can document in detail while having poor research overall.
Say I find my person or family in the 1880 US census, and I use for the Master Source:
1880 U.S. Federal Census (Washington, D.C. National Archives microfilm publication) and then add for the source details, the place of the census, i.e. Chenango, Coventry County, New York
If there is material from the census page that I want to describe or discuss, I describe it in the note, i.e. Persis Crawford Pike was living with her daughter Elvira, husband Grover Maben and their children in Ellington Township, Iowa in 1880.
Generally I would have no need to print out the media when I do a narrative report. If one has many generations, printing media is not particularly useful. Also as Jerry mentioned, this type of documentation exports well in gedcom files.
Since I share most of my research in a couple online shared databases, I try to make sure my sources export in a readable format. Simple works better than complex presentations. It is also easier to teach beginners to start their genealogical journey by a simple sourcing approach.
For US census, I’m a lumper at the year level. So my source list includes 1900 U.S. census, 1910 U.S. census, etc. These do not contain any notes, transcriptions, or media. My citations are named by head of household in this format: Surname, Firstname - Town, County, State. Each citation contains the media, which is named for the head and the year; and the Research note contains the transcription for that family. As others have pointed out, RM10’s nice search box for citations let’s me filter for surname or state or county or town.
This is an excellent method, and it’s pretty much the way any lumper would enter censuses in RM8/9/10. As I have mentioned, it’s really hard not to do it this way in RM8/RM9/RM10 because the user interface in RM8/9/10 makes it very hard to be a splitter.
Each user could have minor variations on this excellent method by lumping by state instead of year, or by lumping by year and state instead of by year, etc. But they are still lumping. Any user using this excellent method would be putting notes, transcriptions, and media into the citation and not the source. Most users as well as Evidence Explained and RM’s own source templates cite down to the family level. As I have described before, I have chosen to cite down to the census page level. I think either way is fine.
I was an extreme splitter prior to RM8 for the reasons that you noted – it was the only way to figure out how a source was being used. I changed to lumping with the design enhancements in RM8/9/10. Rather than changing my old sources, I just add “(Shared)” to the end of all new source names. I may regret this shortcut in the future if I ever focus on publishing reports but, for now, it has made the transition easier.
I also enjoyed your observations about using the FS API. I just recently stuck my toe in that water and my initial reaction was to make manual source updates for the reasons you describe. Your comment of “forgetting about source lumpers” is well said. At the same time, the shared nature of the FS tree makes me a bit more hesitant to automate updates.
As always, thank you for all the excellent guidance that you share.
I guess the exception for me would be two pages (such as front back of WW2 draft card).
Maybe Front back of head stone also vs different plots would have different citations
