Search for content in message boards

Help us assess Old Search

Help us assess Old Search

Posted: 20 Apr 2013 6:08AM GMT
Classification: Query
Edited: 20 Apr 2013 6:09AM GMT
Hello 2010 Search Beta Group,

My name is Dave Menninger and I've been with for about 5 years on the website product team. Right now my focus is on the's search functionality. We're currently trying to understand what makes Old Search so valuable for Old Searchers. We've initially consulted several people internally who have frequent contact with our customers and we've studied a detailed report about Old Search produced by an consultant and professional genealogist who is intimately familiar with Old Search and continues to value and use it.

To help us understand some of the issues we've created a script for Firefox web browsers which simulates some key aspects of Old Search (but within the New Search framework).

If you are a fan of Old Search or use it regularly, we'd like you to try out this simulator:
(Firefox is required; kindly do not post or re-distribute this link.)

Once you’ve had a chance to try the simulator, follow the link in the simulator to give feedback about the simulator and how you use Old Search.

I look forward to hearing your reactions and valuable insights.

Dave Menninger Product Team

Re: Help us assess Old Search

Posted: 21 Apr 2013 3:26AM GMT
Classification: Query
I have begun to use the Old Search simulator as requested. Here are a couple of first impressions. But even before that, let me repeat a little background I've posted before.

I was a dyed in the wool Old Search user who felt (and still does) that the first releases of New Search were extraordinarly dysfunctional. However, a number of significant improvements were made to New Search, and I now use New Search exclusively. However, I always use New Search such that my searches are done in exact mode and such that the results are displayed in Summarized by Category mode. I use wildcards extensively, and I use a number of the (relatively new) goodies in New Search such as searching for similar names and searching for a county and its neighbors.

I still find using New Search as it's more or less intended to be used with ranked searches displayed in Sorted by relevance mode to be totally intolerable. I don't see how anybody can accomplish anything searching in that manner. I say that in full knowledge that a number of researchers whom I greatly respect do use New Search with ranked searches displayed by relevance, and they seem to have great success. I can't understand how that can be, so I just do an Alice in Wonderland and choose to believe impossible things.

All of which to say that the Old Search simulator has very much the look and feel of the way I already use New Search every day of my life. So it seems fine at this point.


Re: Help us assess Old Search

Posted: 21 Apr 2013 3:52AM GMT
Classification: Query
Thanks for the helpful background information. From the sounds of it, the way you already use New Search circumscribes most of what the simulator is doing: basically mimicking the simple Old Search form and defaulting to the category view in New Search.

Re: Help us assess Old Search

Posted: 22 Apr 2013 4:05AM GMT
Classification: Query
Well, I finally read the directions well enough to realize that feedback was requested via the survey rather than here. So I took the survey. Sorry about posting here instead. :)

After taking the survey, I realized that what I was doing here was comparing the Old Search simulation with what I do now with New Search, and the survey was really trying to compare instead the Old Search simulation with the actual Old Search. I did try to answer the questions the best I could.

There was one question I liked especially well. It listed a number of features that Old Search had that neither New Search nor the Old Search simulation have, and it asked how important those features are. It's an extremely well researched list. Your consultant did a great job with it. It reminded me of how much I miss certain aspects of Old Search.

But if I may, I would like to mention one (and maybe two) items that I think your consultant missed. The one I'm sure about is the Old Search date range feature. What I would do with Old Search was to enter the date range as being from the person's birth date to the person's death date. This feature doesn't exist at all in New Search, the best I can tell. For example, if I search for a person for the date range of 1815 through 1902, Old Search will find a tax record from 1838 (tax records contain no birth dates or death dates or any other personal event dates - just the date of the record) and will not find a tax record for 1804 or 1912 - which is the desired behavior. I know of no way to accomplish the same search in New Search.

For the second Old Search feature I have in mind that might be missing in New Search and that might be missing from the consultant's list, I offer up the Old Search "Lived In" feature. Well, New Search obviously has a "Lived in" feature and I use it a lot. But the feature doesn't work quite the same between New Search and Old Search. For example, the "Lived in" feature in Old Search with "Lived in Tennessee" would find a census record for someone who was living in Tennessee irrespective of their birth place and also for someone who was born in Tennessee irrespective of where they were living. This was great, and is better than what New Search does (well, New Search will take "Lived in" down to the county, which Old Search would not do). The "Any event in" feature in New Search comes close to being the same as the Old Search "Lived in" feature. But in my experience the "Any event in" feature in New Search is not propagated properly down from generic search to a search of a specific database.


Re: Help us assess Old Search

Posted: 13 May 2013 5:34PM GMT
Classification: Query
Another item that I think your consultant missed in his list of items in Old Search that are not in New Search is that the search templates for Old Search often include searchable fields that are not listed in the search templates for New Search. These searchable fields that are missing in the New Search templates vary from database to database.

For example, the U.S. Public Records Index databases have a searchable field for the person's address. This field is present in the search template for Old Search but not for New Search. Such data seems to be searchable in New Search as keywords, at least some of the time. But searching for this data as keywords does not always produce the same results as Old Search or results that are as good as Old Search. Of almost more importance, searching by keywords is sort of a big black hole. You have no idea what kinds of things are searchable as keywords. With the Old Search templates, you could tell what was searchable.

This whole problem with the search templates for New Search came about because of the desire on ancestry's part to automate the production of the search templates and thereby to reduce the burden of search template maintenance. But automating the production of the search templates caused the loss of a lot of very useful functionality for the users.


Re: Help us assess Old Search

Posted: 25 Jul 2013 8:39AM GMT
Classification: Query
Edited: 25 Jul 2013 8:40AM GMT
I have recently filled in a typically flawed survey form regarding the old search function. "Improvements" to your search engine will be fruitless unless the problems with your transcription service are resolved. For example; the 'London, England, Electoral Registers, 1832-1965' should be a resource second only to the Censuses, but the transcriptions seem to have been done by a machine working across the pages while the data was recorded in 2 or 3 columns. Consequently, if you enter 'register of electors' into the Street Address box you get nearly 10,000 matches! Enter only 'register' and the matches approach a quarter of a million!! All of these will be invisible to a search engine looking for their real address however much you improve it.
Apparently only 2% of subscribers use the Old Search. How many of us are long time subscribers? It is apparent to me that there are a vast number of people who start their trees, using the New Search as they must, and give up within days because of the bewildering array of nonsensical matches with which they are presented.
I am unable to assess your latest version as I do not use Firefox, but if my experience of the new viewer is anything to go by (slow to load and often freezes) I will be disappointed and very likely to move my subscription elsewhere.
per page

Find a board about a specific topic