Who on the technical side of FS can fix an incorrectly indexed batch?
Over the past three years, whilst using FamilySearch to view Brazilian death/burial records, I have encountered many images where only one death was indexed when there were several more deaths recorded. Who should get notified when I find such images so that the appropriate techies can analyze to determine that the image or the entire batch needs to be reindexed?
As of today, I have collected snip copies of 24 images where only one of many of the deaths on the image was indexed.
Each time I have found such and image I wonder how many records I personally am missing because the death I'm looking for was one of the names that did not get indexed.
Answers
-
Your question will be forwarded to a Search team for review and assistance. You may be contacted by that team if they need more information.
If I am understanding your question correctly the records that you are referring to have already been published. The Search Team for Historical records can better assist you with your question.
0 -
I received an email today basically telling me that there is nothing that FS can do. Here is the text of the email:
"You reported indexing errors in 3 films of records from Brazil. We apologize for the errors. Our goal is to allow user corrections to our indexed collections so that you can fix the errors you find. When it is active for a record set, you see an active Edit button when you open the record details to look at the indexed information. In cases where the button is not yet active, we ask that you simply note the error for now. Hopefully, in the near future, more options for user corrections will open up. Thank you for using FamilySearch. We apologize that our volunteer indexers make mistakes from time to time. Do not respond to this email as it is not monitored for responses. If you have additional questions or concerns, please post in FamilySearch Community at https://community.familysearch.org."
The email mentioned the correction feature in many indexed batches, a feature I have used many times. However, this feature does nothing for death (and other) records that simply were not indexed in the first place, as I described in my original post. As mentioned, this is the second time that I've contacted FS about his issue on this site. Now, this is the second time the issue has been brushed off as apparently being something that FS doesn't want to understand, perhaps because it causes someone to have to think about the issue and actually do something besides return a canned response that doesn't solve the problem and improve FS record availablility to people with Brazilian connections.
One more time, the ISSUE IS - Indexed images where only one of many death records on the microfilm image was indexed. So far I've found more than 24 such images and have saved snip copies. I am simply wanting to know to whom I can send the snips with confidence that remedial action will actually take place. Based on my more than 30 years as a systems analyst and as former missionary with a VROC, to me this seems to be an easy thing to correct by simply having expert (not novice) indexers who know Portuguese re-index the batches.
The Church may be true, but its people aren't perfect - I understand that. What I don't understand is the reluctance of FS to understand this issue and make a REAL effort to resolve it....
0 -
@N Tychonievich is this one you can boost?
Thanks!
0 -
@Milton Allen Maughan Can you provide us with URLs so we can see the problem? When we report these sorts of things, we need to have URLs for those who are going to look at the issue and decide how to handle it. Thanks. (Screenshots are OK if they show the DGs number and image number. URLs are preferred.)
1 -
Yes I can. These have been found over the course of four years of Brazilian research. What do I do in the future when I find more ...?
Birth indexed as death:
Should have five records indexed:
Should have four records indexed:
Should have two records indexed:
Should have two records indexed:
Birth indexed as death:
Should have two records indexed:
Should have four records indexed:
Should have two records indexed:
Should have two records indexed:
Wrong two records indexed (first should have been indexed on for the prior image, second is indexed correctly, third didn't get indexed at all):
Should have eight records indexed:
Should have three records indexed:
Should have three records indexed:
Should have two records indexed:
Should have two records indexed:
Should have four records indexed:
Should have two records indexed:
Should have two records indexed:
Should have three records indexed:
Should have two records indexed:
Should have four records indexed:
Deve ter dois registros indexados:
Deve ter seis registros indexados:
Deve ter tres registros indexados:
0 -
Here is another one...
Birth indexed as death:
0 -
What happened to the post with 25 attachments that was sent, as requested, just prior to the last one with only one attachment?
0 -
@Milton Allen Maughan There has been an issue in recent days with comments/replies disappearing. It's happened to me a few times. Sometimes the comments reappear later; sometimes, not.
See this thread, among others: https://community.familysearch.org/en/discussion/comment/410762#Comment_410762
0 -
So, once again, I'm referred to a series of posts that tell of the problem, but not what to do about it from the people who can actually do something about it....
1 -
Well, @Milton Allen Maughan several of us have reported it to those who can fix it. If we could have fixed it, we would have already.
0 -
So, why not tell me that in the first place?
0 -
Now that I've got you all angry with me, once the technical issue of disappearing posts is solved will anyone let me know so that I can post again? One platitude that I like goes something like this, "The problem with communication is the illusion that communication has actually taken place."
0 -
The thread to which I pointed you was an explicit reporting of the problem, so I DID tell you that in the first place.
0 -
@Milton Allen Maughan Mere speculation, but if you tried to include 25 screenshots in a message, it is likely that the system choked on adding so many images to your post.
Now, let me answer your initial question. Those of us who show "mod" after our names in Community are able to move your reports of errors on to the developers. But, of course, we operate under certain constraints. We have to be able to replicate the errors. The errors have to make the records unusable (there are exceptions to that, but not many).
The response you received when your issue was moved on initially is accurate if each of the 20-some errors you have found are errors made by indexers of the records. Closed indexing projects are not re-opened unless there are massive quantities of problems with the existing index. An index is, by its very nature, flawed. It is simply a tool to use to find the images. As long as it is able to do that, the expectation is that the user will open the original image and use that information to document the life of an ancestor. FamilySearch allows some editing of indexes. The hope is that editing will be expanded in the future. Of course, editing the index is not a cure-all either. What I see as a 7, you might see as a 1 and we could enjoy an editing war over it.
Now, we are happy to look at the errors you have found in Brazilian death and burial records to see if they are other than errors made by well meaning volunteer indexers. But, we need URLs. You can copy and paste a lot or URLs in a post and I believe the system will let them all through. If all you give us are screen shots, we then have to go find those records. Frankly, when that is all that is provided, it goes on a back burner for me since it is so time consuming to do all the needed replication. I'll always handle issues where the user provide me with a URL that I can copy and paste into my address bar before one with a bunch of screen shots.
1 -
OK. I give up.
0 -
Milton
I am just another 'lowly' User/Patron ...
Just in passing ...
As an aside ...
And, not regarding, your original 'Question'...Who on the technical side of FS can fix an incorrectly indexed batch?
I raised the post, on 29 November 2021, of:
FamilySearch Help
'Category' = Suggest an Idea
"Community.FamilySearch" Forum. FAULT / FLAW. "Comments" DISAPPEARING, once AGAIN ...
The problem/issue, of the DISAPPEARING "Comments", WITH "Images", NOW, appears to be addressed/fixed.
[ As of 9 December 2021 ... ]
And, your 'Comment', with ALL those MANY "Images", NOW appears ('Dated': 5 December 2021).
[ So, it DID NOT 'choke' the "System"; as, the "System" was already CRIPPLED ... it just needed a fix/nudge ... ]
Oh, well ...
Better late, than never ...
Just my thoughts ...
I know, that this certainly does not help/assist, with your original 'Question'; and, I am just letting you know (and, confirming), that one 'unrelated'; but, 'annoying', problem/issue, has been addressed/fixed - for now.
Brett
0 -
It's good to know that at least the disappearing post issue has been fixed. As for the original issue:
Ok, vocês ganharam. FS me derrotou. Tentei fazer um post semelhante no fórum da língua portuguesa, mas o FS não permitiu. Desisto totalmente frustrado em tentar obter ajuda para que alguns brasileiros possam localizar seus parentes por meio dos recursos de pesquisa do FamilySearch. Considere a questão deixada de lado. Eu não me importo mais.
0 -
Keep bugging them, Milton -- squeaky wheel & all. You did an excellent job of giving them examples of the problem. However, I've found that the wheels of FS grind _exceedingly_ slow. An issue like you raised about things-that-should-have-been-indexed-but-weren't is extremely important to fix.
Editing errors in something that got indexed is already implemented.
1 -
You gave examples/screen shots of the issue, what is needed is a link to the pages that have the issue. Copy the web address of the pages needing corrections, and paste them here. You may want to tag mod's in that same comment. Hoping these get fixed soon. :)
0 -
While the URL is optimal, there is enough information on these screenshots that someone could have found each of those images. There are film numbers and project titles and screen shots with the image numbers. I just looked at three of them within 5 minutes. For instance, here is the link to the 6th screen shot, Nelson Stanley Carpenter's death or birth certificate. https://www.familysearch.org/ark:/61903/3:1:3QS7-897P-986?i=61&cc=1582573 Here is the link to the index for the 4th screen shot from the bottom: https://www.familysearch.org/ark:/61903/1:1:6ZCQ-YZBZ where I think the OP says two records were left off. I don't have access to the film, nor do I understand Portuguese.
But, once this was passed on to a Search team in November they could have seen the problems. The email that Milton received simply didn't address the problem that he presented. Allowing user corrections was not what he was asking for; he also reported missing records that could never be found. The subsequent posts did not provide any satisfaction and he surrendered.
"You reported indexing errors in 3 films of records from Brazil. We apologize for the errors. Our goal is to allow user corrections to our indexed collections so that you can fix the errors you find. When it is active for a record set, you see an active Edit button when you open the record details to look at the indexed information. In cases where the button is not yet active, we ask that you simply note the error for now. Hopefully, in the near future, more options for user corrections will open up. Thank you for using FamilySearch. We apologize that our volunteer indexers make mistakes from time to time."
It would be very easy for someone who has access to all the FamilySearch films, and who is familiar with the language, to take on this project and see if there are multiple records in these films that are indexed incorrectly or missing. Maybe that will happen. And if anyone does take on the project, post the URL's. Working together works.
0 -
The problem is still a real issue. I keep using the images that I can find, but I've quit documenting the ones where some people were indexed but others got missed entirely. Perhaps when enough Brazilians get upset about this issue, someone will actually get involved.
Lately, I've found a number of death records from the state of São Paulo, Brasil, that appear to have been indexed by computer. The result is very poorly indexed names and dates for published batches. My searches have found that some of the indexed and published names are nonsense, but looking at the images themselves show that there are real names and dates - most of which are completely legible on the image.
I've thought about sending the issue to Elder Soares, but that probably is inappropriate. Who knows, since he is from São Paulo, some of his relatives may be among the not-indexed-in-the-first-place and poorly indexed records....
1 -
I am not connected with FamilySearch
Lately, I've found a number of death records from the state of São Paulo, Brasil, that appear to have been indexed by computer. The result is very poorly indexed names and dates for published batches. My searches have found that some of the indexed and published names are nonsense, but looking at the images themselves show that there are real names and dates - most of which are completely legible on the image.
Someone on Community recently posted a screen shot of the transcription of a person where there was a note that the record had been produced by a computer. There was also a Feedback button available. If you have issues with the computer produced data, the comment (from memory by a FS moderator) was to use the Feedback button to report the errors.
Edit: FS video https://www.familysearch.org/rootstech/rtc2021/session/insights-in-archives-and-computer-assisted-indexing or https://www.youtube.com/watch?v=k78HooANi60
0 -
Interesting video. But it still doesn't answer the question about why the poorly indexed data passed the "Review" step and is still unusable for a normal FS user. Nor does it answer how the poorly indexed data is to be corrected.
I'm not searching by a person's name, but by the Brazilian state of São Paulo. If I were searching by name, many of the computer-indexed names and dates from 1979 until current, would not be found.
An Example, the indexed data:
attached to the human-readable image:
is totally unrecognizable if someone were to scan by the name "Noete Marques de Moraes" for the birth year of 1920. The name as indexed by the computer is totally illegible and the indexed birth year is some 70 years off.
While working as a computer/systems/data analyst for some 30 years, I had a personal saying that "Bad data is worse than no data!" If no data is available, then the user knows to look elsewhere. Whereas bad data in a "trusted" database misleads the user into thinking that the data doesn't exist, or the user comes to mistrust the "trusted" database and quits looking in frustration.
Just as in the video, some Latin American (Portuguese and Spanish) record keepers had beautiful, clear caligraphy. However, my experience indicates many government clerks or parish registrars who were under pressure to get the forms filled out quickly which resulted in barely legible (if at all) records.
I think it is great that systems are being developed to read natural language. However, there still needs EXPERT human review before the indexed batch gets published for general use. So, back to the original question. Why are machine indexed batches not reviewed by language experts and, if not, how do published bad batches get corrected.
0 -
Oh, yea - the only Feedback button that I've found sends me directly to this forum.
0
