BUG in source attachment - please be patient, this is a bit tough to explain

LegacyUser

Justin Masters said: Please be patient here... this can be a bit confusing.

I believe I found a bug with the source attachments, and I'm going to be going partly from memory.

I'm going to be referencing two SEPARATE people, both named William Roach. I have identified them for this post as William 1 and William 2.

Background:
I'm trying to build the family of William Roach ( GQWH-8GD ) (William 1)

Late last night (er, early this morning), I attach a probate source https://www.familysearch.org/ark:/619...

While attaching it to William 1, I read it for possible relatives, and I find a bunch of people receiving $6,000. So I go looking for them, and in the midst of that search I discover they many of them are siblings, children of Charles Edward Roach ( K48X-5FP )

Puzzled about why William 1 is giving money to people with no apparent relationship, I discover Charles has a brother named William F Roach ( MTPQ-DYH) (William 2), and while the birth year doesn't make sense (at that time), I got ahead and attach the probate source to those related individuals who are siblings to one another.

IMPORTANT:
So at this point, this probate is attached to William 1, and the 4 children of Charles Edward Roach (brother of William 2)

With the help of another person here on GetSatisfaction who I'll keep nameless (THANK YOU DOUGLAS!) :-), he has helped identify some additional info that helps me further confirm the nagging feeling that I've attached the Probate source to the WRONG William (William 1)

HERE'S THE BUG:
I go to William 1, and go to change the source attachment (source linker), and it says the source is already attached to Lewis Augustus Birely Roach ( MJS2-8QF ), one of the 4 siblings of Charles (and nephew to William 2).

(oookay... but it was attached to William 1 as well, why are you complaining about Lewis? So I go ahead and detach it, because it's stopping me from moving forward - I'll reattach it later)

I then change the right side of the source linker to William 2 (via his PID), and I get ANOTHER attachment warning, this time telling me Mary "Minnie" Elizabeth Roach (again, another one of the 4 siblings, and a niece of William 2) is improperly attached.

(I'm thinking "What the heck???" )

So I detach that one, and go on to connect the source to William 2.

Now, I *THOUGHT* I detached William 1 from the source, but it's still attached.

But I'm concerned about this confusion about attachments to other people who SHOULD be attached, and that preventing me from attaching it to the right person.

I think there's a bug in the source linker code somewhere. (between "seemingly" holding on to William 1 and attaching to William 2, but confusing the attachment with two others attached to the source.

If you've read this far, thank you. It's difficult to explain easier than that.

LegacyUser

Jeff Wiseman said: Justin,

I don't think that you are dealing with a bug here. It's just the very idiosyncratic behavior of the source linker that you have just tripped over.

Here are a few things to consider.

1. You cannot use unindexed sources with the source linker. In order to attach an indexed source to a record in the tree, you MUST have a specific person indexed in the record to attach it to a record in the tree.

2. Although the probate source you referred to was an indexed record, there is only ONE person indexed there (i.e., William F Roach). Although other names show up in the original document, since they have not been indexed, they have no citation files pointing to the original document for them. This will be obvious since they do not show up on the left side of the source linker.

3. The "probate source" file that you indicated in your original post is NOT really a source file, it is a citation that points uniquely to the indexed data for William F Roach as it was taken from the Maryland Probate Estate and Guardianship files.

4. In order to associate an indexed source to someone that has not been indexed in it, special handling is necessary. You cannot use the citation file to William as the citation file for any of William's nieces or nephews.

This is why if you just copy that citation into your source box and attach it to any person in the tree other than the William it belongs to, it is always going to link back specifically to William in the index regardless of who you attach that specific citation to. I suspect that this is what you are running into.

Look at it this way. The source linker will only allow you to link a person in the tree to a person that exists in the index file taken from some original document or image. If a person in the original document that you want to link to does not exist in the index for that document, you can not use the source linker to achieve that.

Instead, you have to create your own custom citation that points directly to the original document (or image), thus bypassing the index for that document which doesn't contain the name that you want anyway!

For example, in order to reference this image from Robert Edward Roach's source list, you could go to image 9 of the original document images and select the "Attach to Family Tree" button and follow the directions. This will create and attach a citation source to Robert Edward Roach's source list that points directly to image 9 without going through the index that Robert's name doesn't exist in. You will likely want to modify the title of the source to something like Robert Edward Roach in the Maryland Robate Estate and Guardianship File for William F Roach.

You can use a similar approach (there are a couple of them some including the use of your source box) for any of the other millions of UNINDEXED historic records that the FamilySearch has.

Hope this helps some!

LegacyUser

Jessie Hearle said: Jeff is correct
See https://familysearch.org/ask/productS...

I did some screenshots showing the process. It is very helpful for attaching unindexed images to multiple persons, or even just one person

Adding it to your source box can be extremely useful too.
(Feel free to delete my source attachment & or edit & add your own reason statements.

LegacyUser

Jessie Hearle said:

LegacyUser

Justin Masters said: Jeff, you are a Wise Man! (sorry about the pun... but your explanation was ROCK SOLID!)

Thank you!

I get what you're saying about the index'ed name to THAT source ONLY being William F Roach., and we can't add on with our own "indexes", because... well, William is all that there is.

Okay, so you're saying we can attach an individual page to a person. If those people aren't indexed there, am I going to run into an issue where that source (in this case a page) is "already" attached to one person, as I'm trying to attach another person?

I don't want to have to build a custom source box entry for each person I attach, just because it happens to be an INDEXED document.

Again, thank you!

-----------------------
For those less computer savvy, it's like a card catalog card. That person is the only one indexed. I can't add names to that card and have it be a valid searchable option.

(I'm thinking... the pool of people who have SEEN a card catalog has probably diminished significantly.)

So maybe a phone book entry for a person when there's really a whole housefull of people there...

(Wait... what? you... haven't seen a phonebook? )

Okay, your smart phone has a phone program to look up people, right? You can't pile on more people on that entry for your aunt susie.

(somewhere a light bulb clicks on)

(aaand somewhere else is thinking... what a great way to enhance our phone program to include people who don't have their own cell phone number.. and increase surveillance and understand inter-personal relationships for suspect correlation)

LegacyUser

Jeff Wiseman said: Usually they just call me a "Wise-guy" :-)

Yea, many people look at and use the term "source" very broadly and it covers over a lot of details. When you look at a "source list", technically it is a list of citations, that each point to other citations, that can point to even MORE citations, each time getting closer to an original (usually paper) document.

When you open the document for the URL that you gave, it's important to know that you are seeing 3 "different" sources at the same time. They are combined for convenience, but that in itself causes misunderstandings. For example, the original paper document (not included in the URL), the digital image of the original paper document (on the right side of the page), the data from the index file that was derived from the digital source, and even some of the citation data that is visible are all DIFFERENT SOURCES. In this case, the window you see when you open that URL contains data pulled from up to 3 or 4 different sources, all chained together leading from specific to general.

If you go to a source list and for a given source select "View Source", you will get a page that usually doesn't even include an image of the original document (because it is not contained in that citation), it only has the indexed data for a subset of the original document showing.

An index file can be viewed as a "list of citations", each one pointing to one or more entries on some page of a digital image where it was indexed from. However, when you "attach" that source to a source list using the source linker, what you are really doing is adding a citation pointing to a citation (entry) in the index file (which points to some place in a digital image).

An index file is a source in its own right--even without the digital image source that it was indexed from. You can create a source to cite either an entry in the index file, or some page and line number in the digital image. In the latter case there's a bit more work as you need include information in your source that would allow someone to find that record, even if the original's URL fails.

Jessie gave some good examples of how you might handle your situation.

LegacyUser

Justin Masters said: Jessie, thank you for attaching that record. I needed to fix what you had done

(Alice was not one of the children listed in that source, and Charles Henry was left off.)

Too bad I can't ADD another person to the source, since it was not MY sourcebox it went into. And there's no way to add another person to an unindexed source.

(I *DID* report this issue a few months ago. It's ironic that you can delete people from a source, but you can't add people on when you view who all is attached to a source.)

So this leads into a problem I experienced, and which is thankfully also seen in Jessie's example above.

If you look closely, his 3rd screen shot shows 5 checked circles.

But if you look at the 4th screen shot, only 4 persons are selected. Charles Henry is *NOT* in that list. I'm not sure why. But I experienced the same thing when I tried to attach a record from Alice's PID.

I was going to attach a record from a source (baptism record showing the four extant children and the parents) in which the name was indexed, but no link to the specific page was found. And partway down the little preview box, there's a link to let you go search the images.

I clicked that, and it opened me up to a 30,000 foot view of a whole bunch of images, and so I double clicked on one, and then set about to finding the right page for Alice and siblings to be baptized (page 591).

I clicked the Attach to Family Tree Button, and followed the steps like Jessie shows, and in the end, after checking everyone relevant (parents and 4 kids - total of 6), I clicked Next.

The next screen shows the people I selected, but is missing Alice (I didn't notice it at this time). After completing the reason at the bottom, the next page tells me 5 people have been successfully attached to the record.

But wait.. I picked 6!

So this seems to be buggy behavior. It attached everyone except the person I came in on. (Now remember, I didn't attach the source (yet) as listed in Alice's page. I simply selected the little browse the film link to try to find an image.

I went back and attached the source, but had to manually attach the specific page source reference that the other 5 were successful on.

Is this the converse of the problem I had earlier where William was the indexed link to that record and his name was attached to everything? This instance seemed to link everybody EXCEPT the person who I was on when I started browsing the film to attach people to.

LegacyUser

Jeff Wiseman said: Justin, that behavior does sound a bit odd and if it is not just finger problems, it would be good for FS to look into it. Personally, until I was more familiar with the behavior, I would chalk it up to some strange side effect of the web site as there are many under this particular topic. There have been bugs introduced, fixed, reintroduced, and fixed again in this area (e.g., for a while there, the "Attach to Family Tree" button would not work consistently). Also see my summary below.

Note that I'm not real familiar with using these capabilities, but I believe that as you go about correcting your situation, the following behaviors (idiosyncrasies?) should be kept in mind:

When you create a reference to a digital image (i.e., not a citation to an index of that image) and attach it to a group of people, it can be structured in one of 2 ways and it's easy to get the two confused.

The first structure is where the same identical citation is attached to several different person's source lists (I believe that is what happens when you use the approach that Jessie described--you can get the same effect as placing a source into your source box and then pasting that same source to multiple people). In this case, the entry in each of those person's source lists are not only identical, they are in fact, the SAME citation. That means you can not go and put a unique Title on each one. If you change the Title, description, notes, etc. on one of them, it changes on ALL of them.

The second structure is where each of the citations is unique even though the referencing to the source image in each of them may be identical. Instead of a single citation being shared between multiple person's source lists, the original citation has been duplicated multiple times. This means that even though they may originally all contain the same information, the Title, description, notes, etc. can be uniquely modified on each one without touching any of the others. This could be created by using the "Attach to Family Tree" button for only one person at a time, or taking a Citation that you've put in you source box and duplicating it there multiple times and customizing each title as you go.

In summary, if while setting up your attachments, you accidentally got a mix of the first and second type structures I've described, it can present some views that would be quite confusing. You would need to go to each attachment and determine if it had the same UID or not to know for sure.

LegacyUser

Justin Masters said: Jeff, I understood everything you mentioned... except the "finger problems" part. :-)

I think I'll create a NEW bug entry, with a pointer to this conversation (I suspect it's too long for familysearch employees to bother following) or just reference Jessie's photos. (with some modification/emphasis added).

I can definitely agree that I ran into some weird things with Jessie's attachments, and I'm not even sure how to describe it (or if I could remember it correctly).

LegacyUser

Jeff Wiseman said: Note also that when it comes to indexed sources, theoretically the index for a single person should never be connected to more than one person as in the first structure I mentioned. However, due to bugs in the system and the loosening of criteria for hints, this does happen and with greater frequency than before. The situation that you got into where your indexed source was attached to both of the different William Roachs at the same time is an example of this.

In fact, you can even attach that same indexed record MULTIPLE times to the SAME PERSON! There was a long time bug where once you attached a hint to a person, it would keep coming back to the research helps area, and so that same identical source would get re-attached to the same person again and again and again. I don't know if it was ever resolved completely but I have been see less of that situation recently.

Copying an indexed source citation to your source box and then attaching it to various other person records will create particular problems as well. If you duplicate the indexed source in your source box and attach THAT one to other person records will ten to screw things up event more.

So to avoid issues here:

1. Do not copy indexed sources to you source box unless you REALLY know what you are doing

2. Remember that the source linker is only designed to be used with indexed sources, so if you come across a situation where an indexed source is attached to more than one PID in the tree, something is wrong and must be fixed.

LegacyUser

Jeff Wiseman said: "finger problems" are issues created when your fingers aren't doing what your mind thinks they are while working in the system :-)

LegacyUser

Justin Masters said: Well, I think to clarify/correct your point #2...

If an indexed source (with a single-person referenced in the index) is attached to more than one PID in a tree,...

And about the finger problems... yeah, I have those too.
(mind goes faster than the fingers)

LegacyUser

Jeff Wiseman said: Yea, in the case of an indexed source, if it shows up as being attached to two or more PIDs in the tree, that can only be if:

- they are duplicates and need to be merged, or
- they are not the same person that the source applies to and so the source should be detached from the inappropriate ones.

LegacyUser

ATP said: Excellent explanation, Jeff. Clear, succinct, point by point. Thanks! It much increases my understanding of the issue and how to best utilize the information. Thanks, again!

LegacyUser

Justin Masters said: Okay, Jeff, I've apparently not understood your part about multiple "source" links on a page. So let me share this one...

I'm attaching an index source for a draft record to a man here:
https://www.familysearch.org/ark:/619...

I understand I can only attach Herbert to this record.

But his wife is also listed on that, and I want to make a source for her.

But if I click the document image, or the "View the original document" link below it, I'm taken to this page:
https://www.familysearch.org/ark:/619...

Which is image 827 in that film.

But no "Attach to Family Tree" link is active.

So I'm missing something in the loss of that "Attach to Family Tree"... unless it's being "held back" by the latter part of the URL above... lemme try something...

https://www.familysearch.org/ark:/619...

.... nope. Stripping the stuff after the "?" in the URL does NOT enable the image to have additional references (as a source) to it.

Am I to understand that my ONLY recourse is MANUALLY go to my source tab on the wife's page, click "Add a Source", and then in the course of that process add the link to this image page?

LegacyUser

Jessie Hearle said: The easiest way is for an indexed record which includes unindexed names is to copy the source to your source box
Once copied to your source box, it can be attached to as many persons as you wish to include.
The image URL will remain the same

Please note that other persons can edit the source.

LegacyUser

Jeff Wiseman said: Jessie, this is where you run into problems (i.e., copying indexed sources into the source box and attempting to paste them elsewhere). Let me get an example here...

LegacyUser

Juli said: Notice that the title has been modified: it says "Copy of...". This means that it has been de-linked from the original citation, and changes to it will not affect that original.

This is an example of "one person's bug is another person's feature": I consider the linked/identical citations generated by the "Attach to Family Tree" button to be a Good Thing. The linkage means that if I notice a typo in the transcription, I can fix it once and have it corrected everywhere.

LegacyUser

Jessie Hearle said: You are correct my way can cause issues.
I said it was the easiest way, not the best

You can also go straight to the image and add it to your source box, & create the data for the source manually.

LegacyUser

Jessie Hearle said:

LegacyUser

Jeff Wiseman said: With the change that Jessie made by attaching Herbert's INDEX source directly to his wife, you now get this situation. Go to Herbert's source list and review the attachments:



You can now see that the same INDEX source (i.e. the Herbert specific one) has now been linked to the Tree for BOTH Herbert and Myrtle:



Basically what you have done is that you have documented the "Fact" that Myrtle in the Tree is THE SAME PERSON as Herbert in the index file! This is obviously not correct. You cannot use that Index file as a source for Myrtle, because Myrtle is nowhere in that file! The source for her is a different source. It is NOT the index file, but rather it is the image file that the index was derived from.

Why would you attach a source to someone, when the source (an index file) contains NOTHING about that person ?

LegacyUser

Justin Masters said: Uh, Jessie, that's not a good sign...

Because look at the 1-1 relationship behind the modal window, where Herbert (in the record) appears to be improperly attached to Herbert (in Family Tree). (the Detach symbol)

LegacyUser

Jeff Wiseman said: Exactly. The IMAGE is the source for Myrtle, the INDEX is the source for Herbert

LegacyUser

Jessie Hearle said: Sorry, I lost the text with my last image.
Jeff is correct, I said my way was the easiest way, not necessarily the *best*
way.
Good reason statements can go a long way toward explaining why a source is attached to 2 people.

While she isn’t named in the *index* she IS included in the record.
When you attach the record, you the option to include additional facts or notes.
Some of those notes may include reasons known to you personally, ie spelling errors

You might also point out relatives or the fact that the address matches the address on letters from *your John Smith* to help differentiate same named persons

LegacyUser

Justin Masters said: I went ahead and manually created a new source in Myrtle's entry (and referenced Herbert), and attached Herbert as well.

I think this is looking like the only way to do this....

Which speaks to FamilySearch needing to find a way to make this MUCH easier for everyday people to get right...

...because a couple of experts are fumbling around, and Jeff... well, other than his finger problem, he's got NO problem! :-)

LegacyUser

Jeff Wiseman said: Basically, yes. You see, you need to create a citation for the image and NOT for the index. But when you are on the image, the “Attach to Family Tree” button is greyed out. Why? Because the index for that image is ALREADY attached to the tree—at least that is why I suspect you are seeing the inconsistency. the site software can only see the index for that image that is already attached and since it is an Index, it’s not going to let you do it anywhere else.

Again weird behavior, but you are mixing index sources with image sources on the same page. You didn’t have the problem earlier because in that Estate record, the index only pointed to the first page and everyone else was on a different one.

LegacyUser

Justin Masters said: Hmmm... I wonder if it causes problems when you get to the image from the photo or link in Herbert's attachment to that source.

Hmmm... When I copied that link from Herbert's source page, I get the following (uhhh..I modified this to change the token - I wonder why that's added???)
https://sg30p0.familysearch.org/servi...

Clicking on that image and following the link I get this:
https://www.familysearch.org/ark:/619...

It had a LOT of extra stuff (after the question mark in the URL) that I wonder might just undermine what we're talking about in linking a source more directly.

That link I used strips out all the stuff including and after the question mark in the URL. (But users shouldn't have to do that).

LegacyUser

Jeff Wiseman said: Actually, this isn't a matter of good, better, or best. It is Right or wrong. Remember, the index is as much a source as an image is. The very fact that they could be owned by different organizations is proof of that. Frequently we have indexes but no images because the images are restricted. FamilySearch has done the indexing on images that someone else owns and even if we take the images out of the picture (i.e., not allowed to see them) we can still see the indexes and use them for sources.

Now imagine that situation here. Lets say we don't even have access to the actual picture of the draft registration cards (a common occurrence). We only have the index that was taken from it. Myrtle's name is nowhere to be found in that index. In fact, you don't even know that she exists on those cards that were indexed. Can you claim that the index data for Herbert is a "source" for Myrtle? If someone went to that index looking for Myrtle's information, it doesn't exist there!

Now, later on sometime, you actually get your hands onto images of those cards and you say "Hey look! Myrtle is referenced in this image here!" you could then create a citation to that IMAGE that you attach to Myrtle's source list.

You cite the source where the data comes from. Not a source such as an index where it "could have been" but isn't.

Remember, when you open a "source" in a source list, you get a window that is data collected from maybe 3-4 different sources. For convenience that is really nice, but if you don't understand the underlying structure, people will just gloss over the entire thing treating it as a single monolithic "source". inconsistent terminology as well as unusual structures of citations and sources on FS's site all contribute to a lot of confusion here.

LegacyUser

Jeff Wiseman said: I might point out that it probably isn't necessary to include both the image source as well as the index source for that image as separate sources attached to Herbert. By definition, the index source already cites the image source just by definition of an index. If the citation to an image source is missing, you can still get to it via the citation to the index. Adding the redundant reference to the image is just extra work IMHO

LegacyUser

Jeff Wiseman said: Juli,

Thanks for showing up! I always hope that if I'm screwing up my analysis of things when it comes to handing unindexed sources, that you'll show up and correct me as you have far more experience in this area than I do :-)

You are certainly correct about the "Copy of..." However, because it is a citation to the exact same place in an INDEX, it still isn't great because in theory, it should never be attached to any more than one record in the tree anyway. And of course, it just makes the title longer...

By the way, have you looked at the example that was shown here of the Draft cards where the "Attach to Family Tree" button is disabled? That sounds an awful lot like some things you were talking about not too long ago. Was your issue with the dysfunctional button ever resolved?

LegacyUser

Jeff Wiseman said: I didn't look real close, but sometimes the UID just gets you to the source file, and the other stuff gets you to the correct PLACE in the source file (i.e., image number)