Answers to some search questions
In a comment on one my earlier posts, Jerry Bryan had this to say:
I wonder if we could prevail on Anne to give us an update on the big picture of New Search vs. Old Search, where she thinks it stands, what she thinks it would take to make New Search into a viable replacement for Old Search, etc. I’m obviously not privy to ancestry’s plans nor to their usage statistics, but here’s my take. I try to use New Search as much as practicable, and I probably use it maybe 1/3 of the time. New Search has some really good features, and ancestry has been responsive about fixing some of the most egregious problems with New Search. But New Search is still a pre-beta product. It isn’t even good enough yet for a limited public beta. Unless there are major improvements in New Search, and unless those major improvements appear extremely soon, then I think that Old Search needs to stay around at least two more years, maybe three. I think that’s a realistic timeframe to get New Search into good enough shape that it’s a viable product. But I suspect that ancestry has a much shorter timeframe in mind. Indeed, what if there were a serious stability issue with the site a month from now that led to the premature withdrawal of Old Search from service? That would be a true disaster.
I’d like to try and address some of his points and questions so that everyone knows where we stand and what is coming.
There is no time line on turning off the Old Search UI. While I think we can solve the Search UI issues in a shorter timeframe than 2 or 3 years, they have to resolved before we turn anything off. There are still things that you can do in the Old Search UI that you can’t do in the New Search UI. The issues we are debating internally is how do we get to the point we want to be at with search? And does one size really fit all?
So we are working in two areas to improve your search experience:
1) Algorithms. Sometime in February we will launch a change in the algorithm so that dates are filtered better. You know that situation where you have an ancestor who was born in 1800 and you see 1930 census results? (It sounds so easy to solve, and it is actually not easy. I’m going to do some posts in the coming weeks on how a search at ancestry works. It made my searching a lot more efficient once I understood it, and it will also give you some insight on why things happen the way they do.)
2) User Interface. There are some good things in both UI’s. I think the ability to choose what you want to be exact in the New UI and what you don’t is incredibly useful. And, ahem, you all have been more than clear on what you love about the old UI. We could continue to tweak this and tweak that. And that would make the New Search UI better, but we are stepping back, looking at both UI’s top to bottom and then we will proceed. And proceed carefully. Change for change sake is not the answer. Search has to help you, the searcher, find the right content and all the content that might imply. There are 27,000 data sets on the site. I know the stats on usage, and unfortunately a lot of them are not looked at that often. And there is a lot of good stuff out there.
And a quick note on the state/country pages. I said this in a comment on the other post, but I am looking for a solution to replace those with something other than the card catalog and in a matter of days.


While I’m glad you will be working on a way to narrow dates, it has been nearly six months since I commented about the need to be able to search by locality. Instead of making that easier, your latest change appears to have made it harder!!
Hopefully, somewhere in there you will also be prioritizing a way to find people by locality.
There are three basics to genealogy: name, (at least an approximate) date, and (as specific as possible) place. That is (or should be) the first thing people are taught when they start trying to find their ancestors.
Without those 3 facts, it is impossible to do research, especially if one is dealing with a common name. (Most names are more common than we sometimes think…)
Your search engine, whatever it looks like, must be designed to address those 3 basics.
So far, Ancestry seems not to have figured this out???