Introducing the Geni Consistency Checker

Re: Could a fix for the missing surname at birth be added, by using the last name for it?

Would that be for women also? If so, that could be a problem for my areas and fill in Birth surname with husbands name, which it wasn’t. Inconsistency by the way is finding good errors to correct on this.

Erica,

We will have to think a bit about it.
I agree the fix could be a bit tricky.
The current fixes could introduce new inconsistencies also for the part of the tree I am working on and don't offer the only option that would be correct to fix the problem in many cases.
May be the wording of the fixes could explain a bit more about which one to use.

Hanna Moldsvor, +1

Getting an agreement on the order of the errors might be difficult - depending on what area of the tree you are working in, you will have a different opinion on which ones should appear at the top. I know that my personal preference would be different for different profiles.

You can easily shrink the ones you don’t want to work on or scroll down to the one you want to work on.

I would like the language depended ones at the bottom for now (they still need some tuning) and probably the ones that can be checked and fixed easily and/or have a large impact because they may switch profiles from private to public status (like too old to be still living) near the top.
I would like the ones that are good at spotting bad merges like event before birth and event after death also near the top because those don't give too many false positives so it would them before ones like siblings with the same name that can do the same but have more false positives.

Leanne M (Volunteer Curator - Australia) - on walkabout🇦🇺 - I appreciate you want to be helpful, but giving inaccurate information is not helpful! -- you said, in
https://www.geni.com/discussions/203863?msg=1342981
"the code for the inconsistency may have been altered since the consistency check was calculated on the profiles OR the date of birth on one of the profiles may have been altered since the consistency check was calculated.

When the consistency check is recalculated for the profile it will disappear, alternatively if you refresh the inconsistency (from the profiles inconsistency page) it will be removed."

The date of birth of the two people has not been changed at all in 2019 - yes, it might have been for one of them, but it has not. Various changes to the code for the consistency checker have undoubtedly been made - but -

The inconsistency has disappeared from the profile of the Spouse I explicitly told the Inconsistency checker to "ignore" -- I have refreshed the page while on my inconsistency checker page and while on the Spouse that still has the inconsistency -- the inconsistency on that Spouse remains -- both on its profile page and -
On my inconsistency page, it still shows "Large spouse age difference (2)" -- and the one for the Spouse I told it to ignore shows on the colored strip as "(1 ignored - show all)" while the one for the other spouse still shows as an Inconsistency.

In other words, the situation has not changed at all from how I described it above in https://www.geni.com/discussions/203863?msg=1342946

Either you were not able to explain things so I understood them, or your statement was totally incorrect. Either way, I would greatly appreciate it if you would stop doing such things to me.

Lois, try going to the inconsistency, reversing the ignore and refreshing the inconsistency. The inconsistency should then disappear. If it doesn’t please send me a link to the profile so I can look at it because 2 years is not a large spouse age difference and it should not be giving that inconsistency.

Sorry - I should have clarified - when I say refresh the inconsistency I don’t mean “refresh the internet page” I mean to “refresh the consistency check” by clicking on the refresh symbol on the right side of the screen just above the inconsistencies.

Job Waterreus - re: your statement that "event before birth and event after death" are ones that "don't give too many false positives" -- For Profiles I work on, that is the category that has the most false positives

-- When I write up info from Marriage Records or Announcements in the Timeline Event for a Marriage, I tag all people mentioned in the write-up in the Event. When I write up info from a Death Record or Obit or Death Death Notice in the Timeline Event for the Death, I tag all people mentioned in the About of that Timeline Event.
Even for a Birth -- if I am getting the info from JewishGen, for example, it may well list the child's father and his father, the child's mother and her father -- and I will include that in the About on the Timeline Event for Birth - and tag all mentioned to that Birth Event.
I find many advantages to this.

My guess - the reason these "inconsistencies" are not producing more false positives is because most folks do not do anything with the Timeline Events.

Leanne M (Volunteer Curator - Australia) - on walkabout🇦🇺 - Why are you saying "2 years is not a large spouse age difference and it should not be giving that inconsistency" -- it is NOT a 2 year age difference. I did not say anything about how large an age difference it was.
Am I guessing correctly - is that from the (2) in parentheses after "Large spouse age difference" - which refers to how many such inconsistencies there are - as it does on each of the strips?

re: "“refresh the consistency check” by clicking on the refresh symbol on the right side of the screen just above the inconsistencies." - are you talking about the Inconsistency Page or one of the Profiles with the inconsistency?
For what it is worth, so far I do not see any such symbol anyplace on either, but would at least like to know what your talking about.

I really, really wish you had let Mike answer. I usually can understand his comments.

Private User,

I asked Mike not to give that inconsistency when there are multiple profiles connected to the event or to sort it so those are at the bottom.
He does not want to do that (at least not yet) stating you don't have to attach the profiles to the event, you can add a document to the event and add the profiles to the document (that won't trigger the inconsistency).
I think you mentioned that Geni does add the parents to an birth event, so it seems a bit illogical that if a user does something like that for a baptize event (or other event) it will trigger the inconsistency.
I think Mike argument about being present at the event is not very strong.
If you ignore the ones with multiple profiles the remaining ones will mostly be real problems.

I asked Mike to auto hide the inconsistency between profiles if you hide it from one of the profiles, but wrote it in a way that he did not understood.
I clarified it later, but that very shortly before the release, so I hope he will implement that feature.

Leanne M (Volunteer Curator - Australia) - on walkabout🇦🇺 - On the Profile Screen with the Inconsistency, to the right of where it says "1 inconsistency" there is a red box with the word "NEW" -- I clicked it and got to a page with just her inconsistency - that had a "refresh" symbol - I clicked it - the inconsistency stayed -- I "Unignored" the one I had ignored -- and then clicked "ignore" again -- and then yet again clicked "refresh" - and everything is exactly as it was.

Did you understand what I was talking about when I asked Mike " for ""Large spouse age difference" is there a way you can have your system count it as one inconsistency even tho both spouses are obviously involved - or at least have both inconsistencies marked "ignore" if one is?" ?

Job Waterreus - Thanks. I would modify your suggestion a bit, to suggest ignoring if the Event has folks tagged to it other than those Geni would auto-tag to it - (for birth, Geni will auto-tag 3 if all present, for Marriage 2) but doubt Mike would be any more open to that.

I find it useful to link to folks in the Timeline Event - even in cases when I have tagged them to a Document which is tagged to the Timeline Event. Clearly, it was not how Mike pictured it - but if he said anything about that anytime before this, I would love a link. And I think by never picturing it's being used this way, Mike has actually missed out on visualizing just how useful the Timeline can be!

Job -- re: "auto hide the inconsistency between profiles if you hide it from one of the profiles" -when I first read it, was thinking of it with regard to Timeline Events (and scratching my head)
- but if instead it was inspired by or relevant to my suggestion re: couples with a large age difference - Thanks again!!

I was reffering to that inconsistency (and others) between multiple profiles. If you hide it from one of the profiles it would help if it was hidden (in the same way [for the user or for everyone]) from the other concerned profiles as well.

I think you are right Mike not picturing that way to add multiple profiles to the timeline, so I hope he will change his mind about it.

Mike Stangel,

It seems if you ignore a inconsistency for everyone from the profile, you cannot undo that from the profile. Show all does not show you the ignored inconsistency.

Example: https://www.geni.com/inconsistencies/for_profile/6000000002175280007

Job, looks like when you press on show all it doesn't actually show all for you and me so once you ignore something you can't get back to it to reverse your unignore.
But it does seem to work for Lois as she is able to unignore them.
Confusing.

I want to thank Private User for a very nice comment here: https://www.geni.com/discussions/204014?msg=1343026

Leanne M (Volunteer Curator - Australia) - on walkabout🇦🇺 - I did "ignore" from the full Consistency Checker Page (ie the page that shows ALL of my inconsistencies) - NOT from the Consistency Checker from the individual profile (which just shows those for that profile) -
Did you try some from each and get the same result, or which did you do?

by "ALL of my inconsistencies" -- mean ALL the inconsistencies Geni chooses to specifically be telling me about [via link at the bottom of the Research pull-down menu, and the Home Page Notifications] - not that I caused all of them. Not sure what criteria they use - definitely not all of them are Profiles I manage, not all of them are Profiles in my Max Family - think one was a profile I neither followed nor managed nor was in my Max Family, but couldn't remember which it was when I wanted to check on that

Leanne - if you read Job's comment carefully - he is specifically saying - if you choose "ignore" and choose to have it marked correct for everyone -- and if you do it from the profile --then ... -- Whereas, as I mentioned above, that was not what I did.

From the way you wrote your comment, there is no indication of any awareness of the different ways - you do not refer to trying it both ways, do not say anything about which way you did it.
And - If you were totally unaware of the two different ways - that may explain why your explanation to me about what you meant by "refresh" and how to find it was so majorly lacking.

Hi Lois,

Ignoring the inconsistency from the Consistency Checker Page or ignoring the inconsistency from the Profile's Consistency Checker Page does the same thing - it ignores the inconsistency for the profile.

There appears to be an issue that is not allowing some of us to see previously ignored ones.

What I have seen from spending hundreds of hours testing the consistency checker, going to the Consistency Checker page (the one that has "My Relatives" or "Managed by Me" on the top right - which gives you many many profiles) just displays inconsistencies previously calculated, it does not necessarily recalculate them all. So an inconsistency that is being displayed may have been created from 1) a previous version of the code 2) data that has subsequently been changed

What I have also seen is that opening the profile does not necessarily cause a recalculation in the consistency check. So an inconsistency that is being displayed may have been created from 1) a previous version of the code 2) data that has subsequently been changed

What I have also seen is that changing field values on a profile does not necessarily cause a recalculation in the consistency check. So an inconsistency that is being displayed may have been created from 1) a previous version of the code 2) data that has subsequently been changed

What I have found is that the only way for me to absolutely KNOW that the consistency check for a specific profile is using 1) the latest consistency check coding and 2) the data currently on the profile is to manually force a recalculation for the specific profile.

To manually force a recalculation I have to
* go to the profile
* click on the # of inconsistencies - which will open a screen showing just the inconsistencies for that specific profile
* click refresh (the little refresh icon on the right - not the webpage refresh)

If that does not remove the inconsistency then I know that the inconsistency is still a inconsistency whether that is a true problem, or a bug in the code or it is the fact which trigger an inconsistency (ie someone married someone 70 years their junior).

If you follow those steps then you KNOW that the inconsistency is using 1) the latest code 2) the current data on the profiles.

If you have already ignored the inconsistency, then we may need to wait for the issue (ie the one that is stopping some of us from accessing previously ignored ones) to be fixed so we can reverse the ignore, and manually force a recalculation.

Leanne M (Volunteer Curator - Australia) - on walkabout🇦🇺 - Much clearer. Thank you!

Most of the profiles I have checked had just one inconsistency - once I marked them as "Ignore" and chose "Mark them correct for all users" - then all reference to any inconsistency disappeared from the Profile in Question - so it is not possible for me to go from the Profile to any Inconsistency Page for that specific Profile.
Personally I think that is good. And hope it looks that way for everyone who looks at the Profile.

And for the example Job Waterreus provided, I can only see the Consistency Checker from the Profile's Page that he provides the Link to [and which nobody would be able to get to from the Profile if all had been marked "ignore" and "Mark them correct for all users" - is that correct?]

What I am wondering -- is it perhaps only on the Consistency Page from the Individual Profile where "Show All" does not work -- or is it also not working from a "Full" Consistency Page (from someone in its Max Family or Manager, depending on what was chosen) on which it appears - as I said, I have not been able to try both --
--- ie, we have two variables:
which Consistency Page did one use to say "Ignore"
which Consistency Page is one looking at after, and trying to click "show all"
-- have you checked all combinations, and is it the same for all combinations?

Now that my attention has been called to the Drop Down Menu to the upper right of my "Full" Consistency Checker page --
I see "My Relatives" and "Distant Relatives" as two separate categories/choices
-- I have 216 inconsistencies for "My Relatives" and 217 for "Distant Relatives"
so apparently "Distant Relatives" includes some not included in "My Relatives"
- What is meant by each of these categories? (ie Which profiles are included in each of these two categorles?)

Lois,

Geni did not provide any information on the difference between my relatives and distant relatives (that I am aware off).
What I've seen it seems the distant relatives go back further in time (before 1600)

"Event after Death (165)" and "48 ignored - show all' - I clicked "Load More" - and as it Loaded More, the number listed as "ignored" went up -
It finally has apparently stopped, ie reached the end of the loading, and now shows
"215 ignored", even tho 165 is still being shown as the total number that was found in that category - and lots and lots are showing still needing to be changed or ignored

Job Waterreus - any idea what is going on there? If not, should I tag Mike, or ??

Also - in the Blog, it says "We will soon back fill inconsistencies for older profiles in batches, until all profiles on Geni have been checked." -- Given that, would expect the numbers to keep jumping up as the finish more batches, until "all profiles on Geni have been checked". Has all the back filling happened yet? If not, any idea what percent of the way we are? And, will we be told when it is 'all caught up'?

Lois,

I can't answer all your questions, but I will try a few.
This is from what I saw while testing, not from what Mike told use, so it could be wrong!

It seems that when the checker is run for an inconsistency it determines the total that you see in the header. It appears this will stay the same till the inconsistency check for that inconsistency is run again. When you ignore the check it seems that is marked in the list for that inconsistency. When you reload the page it will load the first five inconsistencies and show you how many of them where ignored. When you scroll down when it is loading that will trigger the loading of the next few and if any of those where ignored you will see the ignore count rising.
When you alter data in a profile in many cases (but not all) the inconsistency will automatically be rechecked. You can also from the profile force it to recheck.
But the checker is still running so any new profile or profile change for a profile that did not appear in any of the inconsistencies may trigger an new inconsistency.
So there is the checking for profiles that did not change with new code for the checker and there is checking for profiles that did change (that would be with the new code)
I have no idea how long it would take to recheck those profiles which did not change but are rechecked because of code changes. You might be able to see some indication of the progress by noticing a change in the total that is shown in the header for an inconsistency.

Torbjørn Igelkjøn it turns out we had already fixed the date comparisons to handle that "born 1953, parent died 1952" issue. We just needed to re-check the inconsistency. Right now we're re-checking all of the Children With Different Last Names to account for our new Lithuanian patronymic / family name conventions. Once that's done we'll re-run all of Child Born After Parent Death inconsistencies.

Private User yes we will have it ignore reciprocal Large Spouse Age Difference inconsistencies when you hide or ignore one.

Torbjørn Nilsen you wrote, "I have seen several instances from the 18th and 19th centuries where the age difference at marriage is 40-50 years or fathers who had children when they reached the age of 80. The sources show that the information is correct." Our response to that is that sure, it's *possible* but the point of the consistency checker is to point out possible problems. How often is a 75-year-old father actually the father, and how often might this indicate that a generation was skipped? The point is to draw your attention to unusual data, not to tell you that it must be wrong (at least, not in all cases).

Job Waterreus yes we will add those Dutch titles. Can you point me to an example of where you'd prefer to use the last name as the auto-fix for Missing Birth Surname?

Job Waterreus we've fixed the Show All problem that you pointed out on Hendrina van Ingen.

Private User and Job Waterreus we need to document the groups somewhere... My Relatives is your max family walk (up to 3rd great grandparents, out to 4th cousins) while Distant Relatives is max family PLUS up to 20 generations of ancestors.

Private User the ignored count in the header increases as it loads more, because it is encountering hidden inconsistencies as it loads. We didn't have to do it this way, but the most convincing reason we had to only count as it loads, is because when you click "show all" you're going to expect to see that many more appear. So if the ignored count has gotten up to 20 and you click "show all" it's going to show 20 more inconsistencies. Whereas if we'd given you the total there, say there's 75 hidden or ignored inconsistencies, you might click "show all" and not see ANY because they haven't been loaded yet.

Mike Stangel -- You explained why the count on ignored goes up while loading --
but -- how can the number ignored (of one type of inconsistency) go up to more than the total of that type of inconsistency (in the case I was looking at, lots more, and with lots still left not ignored)?

Private User it appears to be loading the same inconsistencies multiple times. We'll figure out what's causing that.