Improvement to standardised place names volunteer opportunity
Comments
-
Tom Huber said: Welcome to the community-powered feedback forum for FamilySearch. FamilySearch personnel read every discussion thread and may or may not respond as their time permits. We all share an active interest in using the resources of this site and as users, we have various levels of knowledge and experience and do our best to help each other with concerns, issues, and/or questions.
This is under active discussion in another thread. See https://getsatisfaction.com/familysea...
Since the standards teams (who set up this project) do not regularly read these discussion threads, you may want to contact them at PlaceFeedback@familysearch.org with your suggestion (it is also the topic in that other discussion, and two more discussion threads are listed in that other discussion thread).0 -
JimGreene said: Entering this in here as well for Mike's benefit:
Tom and all:
Among other things, I am the Product Marketing Manager for "Improving Place names" so I would like to take a stab at an explanation.
First, some of you are reading too much into this exercise. All we want is some sort of standard in the place name field, even a high level standard to replace the data that is currently not a standard. If someone has placed "Fulton and Haight, San Francisco, California" in for the place name, it is obviously not a standard. All we want is something that can be recognized so that the automation, and community can take it from there. It does not need to be perfect, just right. Look at the other place names for the person, determine if what else is there is enough for you to enter "San Francisco, California, United States," or maybe just "California, United States" Either is a standard, though neither is 100% complete. Still, it is all we are looking for. Don't stress over this, if things still do not look right, then click that you can't do this one and move on. If you want to research this out that is your choice, we won't argue. But again, we just want good enough to get the ball rolling, not perfection. And while it would undoubtedly be better if you could select batches from places with which you are familiar, the cost of doing that, for a program that is very temporary and will only last until the finite identified group of names is done, it just is not cost effective. If this causes frustration, then don't do it. We appreciate everyone's help, and the more that help the sooner it will be done. Thanks!!
Jim1 -
Mike Dyble said: Ok thanks for the reply, now I have a more information I can see the problem, as an engineer i get idea of 'fit for purpose' rather than 'perfect' as some customers can only afford fit for purpose. I was trying to be too clever.... maybe a little bit more explanation on the web page banner would help.1
-
Adrian Bruce said: Jim - thanks for that. Phrases like "we just want good enough to get the ball rolling, not perfection" do help because wasn't obvious to me that "get the ball rolling" would suffice.
One interesting bit - you mention "something that can be recognized so that the automation, ... can take it from there".
Would you care to say what the automation in question is? And if it changes from run to run, I'll understand if I've got a basic view to start from.0 -
JimGreene said: Thanks Mike, actually we try to keep banners to a minimum, we were hoping to get the information into the community, and you helped us get it there!!! It is just too long of an explanation for a banner:) Thanks for your interest and help and ideas, don't stop!!!0
-
JimGreene said: Adrian, a couple of examples: Our hints are completely automated, and require a few key pieces of information, including a standardized place. Getting a hint out to someone who is related is a great way to get more and better information. Family would be able to improve the place name for sure. Another example involves members of the Church of Jesus Christ and their relatives. Temple qualification is also automated, and requires a standardized place.
Just a couple examples of the automations we have that are kicked off or enhanced by having a standardized place name.0 -
Adrian Bruce said: Oh, OK, Jim. Not quite the sort of automation I had in mind - I was wondering if the automation would automate (sure that's not good English!) a better choice of standard place-name either immediately or if the standard place-names were updated later on.
But I can see that your examples make sense by putting stuff in front of people who - hopefully - can contribute more.0 -
JimGreene said: sorry for the confusion, I can see why having automation to help the standardization process would be considered cool. It is pretty complex matching, and we did the best we could to literally take 100's of millions with bad place names down to 10's, but that is the best we could do, and now we need to either wait for better AI, or involve community and family.0
-
This is still broken, as a volunteer, we cant skip through 20 locations before we get to one that we selected. This sucks a bit as I want to help but waste time skipping and not helping!!!!
I select South Africa and get about 20% of locations in South Africa to resolve the rest is random, PLEASE SOLVE this or don't ask for volunteers to assist.
0 -
@SakkieJanse van Rensburg, chicken-and-egg: if there's no standard, then the computer has no clue where it is. That's the whole point of the volunteer opportunity: telling the computer at least an approximate location.
Sometimes, the algorithm can take a rough guess at a continent or even a country, but it'll be just that: a rough guess. Other times, it can't even do that -- and those are the entries that need help the most.
1