New Search:Six months later
I had really hoped that some even minor improvements would be made to the 'New and Improved' search.
It is still a nightmare, I have looked at and read all the how to information, I have spent innumerable hours looking for information that I could have found in less than 5 minutes in August 2021.
I no longer search for records before creating a profile. I add a person and then hope that if I come back to them several days later, that some records will pop up.
Even if I search from an existing profile the results are still garbage.
You need to have the entire search form available from the beginning. The results need to reflect the information entered in the search form.
If I enter born in Lancashire England, lived in Lancashire England, married in Lancashire England, died in Lancashire England, I don't want or need results from North America. And when I enter a spouses name and or parents names I don't want results for other names.
I would like to be able to pick results categories from the list of results, but I can't I have to go back to the search form, re-enter info, wander around until I can find something that seems to make sense, however it often doesn't.
I have been involved in family history research for over 10 years and have never been so frustrated.
Answers
-
Even though I dislike the new Search interface about as much as practically every other user that comments here, I don't find myself struggling so much in getting the expected results.
The three main points that help in my searches are:
(1) Use of "Exact" match boxes - in conjunction with use of wildcards, where necessary.
(2) Only inputting one set of data at a time - e.g. if looking for a marriage, leave birth and death fields blank.
(3) Filter on "Type" and/or "Collection".
If I search for Lancashire records, in line with the above, I rarely find results from other parts of England, or the world.
Take the example below. I have inputted to two areas - Birth and Residence - and have not used any filters, but still most of my results relate to George Wilson individuals, born in Lancashire in the 1865-1867 period and found in Lancashire in the 1881 census. (Just two "irrelevant" results at the end, which would have been omitted if I had searched / filtered just on the 1881 E&W Census collection.
I wonder what types of inputs other users are making, when claiming their searches produce thousands of results from all over the globe.
2 -
It is now September 2022 and the results I get most of the time are still garbage. I have done the exact search and every other possible way of searching and I am still incredibly frustrated.
I have entered specific dates, locations and names, and most of the time get results for the wrong names sometimes wrong country, events 40-100 years different than the dates I entered.
I am working on a One Place Study and the current search results are largely useless.
0 -
Please share some of those examples with the FamilySearch community, including - hopefully - any of the engineers who might be reading your and similar posts.
Until recently, it was mainly the new presentation (design) of the pages that I was finding it difficult to work with (and still do). However, a recent tweak in the algorithm does now appear to be making the results of "exact" searches to not match ones inputs - i.e., a list of results including an increased number of unwanted ones, which not necessarily found at the end of the list, either.
As I suggest, it would be good to see your examples, so further evidence can be provided to show there is sometimes no way of filtering-out results for "wrong countries" and time periods, etc..
1 -
I have to say that I find complaints like this about the search function rather frustrating. It seems that people just want to complain but are not interested in seeing any improvements. Post like this are not helpful at all.
Would you take your car to a mechanic and just say, "my car doesn't work"? Would you go to the doctor, just say, "I'm sick," and refuse to say anything else? Would you summon a plumber to your house and leave as soon as he arrives without says anything?
If you expect the engineers to improve the search routines, you are going to have to:
- Report if you are using Search or Find.
- Post the exact search you set up. This works best by posting the URL of the search results page.
- Explain what you expected the search to produce.
- Explain why you thought the search should give those results.
For example, I work a lot in the records for Stord, Hordaland, Norway. I know that historical records are highly unlikely to have birth, death, marriage, and residence information so I would never include all of that when using SEARCH. However, Family Tree will often have all this information, so I would expect results using birth, death, marriage, and any place fields.
- Using FIND, since I am doing a Family Tree search.
- I set up this search: https://www.familysearch.org/search/tree/results?q.anyPlace=stord%2C%20hordaland%2C%20norway&q.anyPlace.exact=on&q.birthLikePlace=stord%2C%20hordaland%2C%20norway&q.birthLikePlace.exact=on&q.deathLikePlace=stord%2C%20hordaland%2C%20norway&q.deathLikePlace.exact=on&q.marriageLikePlace=stord%2C%20hordaland%2C%20norway&q.marriageLikePlace.exact=on
- I expect to get a list of people who were born, married, and died at Stord.
- I expect these results because I know from my previous research that there are a large number of people in Family Tree who did exactly this.
What I do find, is that the FIND routine worked perfectly. Scanning through the first four pages of results, all of them fit the criteria just as they should. Jumping to the last page of results, all of them match as expected as well.
On the other point that came up, I have to say that I am really impressed with the Hinting routine and the way that it almost does away with the need for searching historical records. It brings up a lot of high quality matches these days. So the search engine does work great. We just need to figure out how to set up searches the way the Hints routine does. That would be a great RootsTech presentation: "How to Search Like the Hints Routine." I wish one of the people behind the hints algorithm would take the hint and teach us how.
2 -
After using the search for so long, I find that I have lowered my expectations so much that I can't even remember all the reasons I hate it so much. Basically, your best bet is to search in specific record collections whenever possible.
0 -
To that point, I would reply that it has always worked better, ever since the release of any FamilySearch search engine, to work in a specific record collection. That is the best filter you can put on a search. If you are looking for records in the 1940 census, why would you want results from the 1920 census? The only time I do a general, database wide search is either at the very beginning of research to get a general lay of the land or as a last resort if no other search is bringing anything up.
Your comment also makes me wonder if the reason people are having less success with general searches is that the complete set of all databases has just gotten too huge. There are just too many possible results.
Also, each individual database has its own quirks of data presentation. If you search in just one collection, you can learn in about ten minutes how the data is arranged and what the most effective searches are going to be. That is impossible to do if you are searching all the databases at once.
But whether you search a specific record collection or search everything at once, the search engine is the same and works the same.
Here is a little experiment.
I am going to look for Anders, born between 1820 and 1830 in Stord, Hordaland, Norway. If I do a general search, here are the results: https://www.familysearch.org/search/record/results?q.birthLikeDate.from=1820&q.birthLikeDate.to=1830&q.birthLikePlace=Stord%2C%20Hordaland%2C%20Norway&q.givenName=Anders
There are 122,966 results. The first page of results all have the right name, the right place, and all the christening dates fit the date range quite nicely because christening dates are treated like birth dates. The ones that have baptism events are out of the date range, but that is because baptisms are custom events, not birth date equivalents, so those records are treated as if they have no birth date at all.
If I go to a specific collection, Norway Births, I get these results: https://www.familysearch.org/search/record/results?f.collectionId=1467014&q.anyDate.from=1820&q.anyDate.to=1830&q.anyPlace=stord%2C%20hordaland%2C%20norway&q.givenName=anders
There are 2,315 results. All the first page have the right name, the right place, and dates are all within range. Which it to be expected because all of these have christening dates which are treated as birth date equivalents.
But if you look at this result page closely, you will see that this really looks just as the same as the other result page except it has a collection filter on the top. If I remove that filter, the updated results are: https://www.familysearch.org/search/record/results?count=20&q.birthLikeDate.from=1820&q.birthLikeDate.to=1830&q.birthLikePlace=stord%2C+hordaland%2C+norway&q.givenName=anders which are the same 122,966 search results it got from my original search.
So searching from a specific collection is just using the main search engine, which works quite well, and adding specific filters.
One advantage, I will say, of starting from the main search, is that you can filter on more collection than one. For example, I can filter by all the collections that might have the right birth date and place and look at all of them at once: https://www.familysearch.org/search/record/results?count=20&q.birthLikeDate.from=1820&q.birthLikeDate.to=1830&q.birthLikePlace=Stord%2C%20Hordaland%2C%20Norway&q.givenName=Anders&f.collectionId=1467014&f.collectionId=1468081&f.collectionId=1468080&f.collectionId=3756102&f.collectionId=4135961&f.collectionId=1529106&f.collectionId=4067726&f.collectionId=3744863&f.collectionId=4237104
It works great, allowing for the feature of the search engine that puts in additional records that sort of fit but are missing part of the criteria. As can be seen here, the first three records look like the dates are out of range, but they are not, because they are being treated as if they have no birth date since baptismal dates are not treated like christening dates which are treated like birth dates. Also that particular collection has a problem in that the birth date was indexed with just the day and month so some places in FamilySearch treat that first Anders birth date of 25 March as March 0025. Here I'm assuming the search engine is just ignoring it.
2 -
Gordon has provided some excellent analogies above concerning what your expectations should be if you refuse to assist in any diagnosis of your problem(s).
Unfortunately, no one can assist you if you persist in making complaints without any examples to illustrate your difficulties.
1 -
No disrespect to Gordon or Paul, both are capable and creditable, but I actually support the statements made by @Monica Ross_1 without needing her to submit a detailed snag list. There are thousands of exisiting comments made by users which describe the problems adequately. Monica's point is valid: After 6 months It seems that nothing has been done to improve matters.
It is now more than 12 months since the new and "improved" search presentation was forced onto us without being requested, and the issues reported by users are still unresolved.
1 -
The problem here is the many people suggesting work a rounds that might help to solve a problem created by the engineers/programmers.
In the past I have been involved with businesses where new software was created to "solve" a problem that did not exist.
As the people involved in solving 'the problem' did not understand and did not wish to know how the software was used by the customers the end result was not satisfactory. And the new software was often ditched because it cost the company more money to use the improved process without it providing improved results.
I have no interest in spending many hours of my time providing details of how the 'new' search process does not work. It is quite obvious that it doesn't. And it is equally obvious that FS is not interested in 'fixing' the problem.
I belong to a family history group with over 50,000 members and for the most part we are all having varying degrees of problems with getting results that fit the information searched for.
I am merely expressing my continuing dissatisfaction with the current 'improved' search.
0
