Annotation and Finding

When searching for data that has been OCR’d, you can often locate, once the page is in the viewer, with FIND. If the information has not been OCR’d or indexed, you may wish to make an annotation, so that you can find it more easily later.

ANNOTATION. When you find a listing, you can annotate it, and make it easier for you to find later. You could also save every directory page that you like to your gallery, perhaps in a folder that you name, “Frank F Sharbrough LA City Directories” so you can find them later. If course you could save them to your own computer, too.

FIND – maybe, maybe not. As a rule, when OCR text is involved, the snazzy Footnote viewer shows a FIND button in the toolbar at the top. It makes it easy to find where your Sharbrough is on the page. It’s VERY helpful with newspapers, or to find the Davises on the Vietnam Memorial. There are cases, and I saw a couple during this project, where the FIND button doesn’t appear. This means that Footnote staff considered the OCR quality for that page so bad that it is not available. This happens. I’ve suggested that they create a tool for volunteers to create such products, but it’s not an imminent development.

Psst! Hey Buddy! Over here! Wanna buy a Timeline?

One of the features of the Footnote site that I plan to write about someday are “Footnote Pages.” They are web pages, created by members and by staff. There is already a Footnote Page for Francis F Sharbrough, created by Footnote staff based on his entry in the SSDI. I could add these entries to it. I could, and I did. Further, if you saved the directory pages to your gallery, you could link to them as sources for your timeline. I could, and I did not. By entering some of these, I was able to determine that in 1905 Frank lived 7 blocks from work at the Broadway Dept Store. Here’s an example.

In this example, you can see the lollipops on the map, showing places that Frank lived and worked in LA. In the top right part, you can see the timeline for his life events. There are zooming and scrolling controls on both widgets, so you can drill down and see a long period of stability at a job, or back out and see the world events that happened in his life. I have to say that I kind of grinned at the idea that Frank might be in the Missing Air Crew Reports – a guy born in 1884 was almost 60 when WWII began for the US. Heck, FN recommends that I look for my grandmother in there! She was born 7 years after Frank, and passed away a year before he did.

Summing it up

By searching for information in city directories, you can learn a lot about a person’s life. You can watch them move around town, and you can tell how far they lived from work. If the directory listing isn’t indexed so that FN Search brings it up, you can still find it, relatively quickly and easily. Afterward, if you like, you can create a Footnote Page to help you analyze the findings further.

Last tip: these city directories didn’t include an address for the employer. You can look those up in the directory, or try google. I was able to find a picture of the Broadway Department Store from 1912 on Flickr, and was able to view a satellite image of 502 E 21st St on Google Maps. Now I feel like I’ve got enough information to have a deeper appreciation of the life that Frank made in LA. And I still don’t know much about him!

Here’s what worked

When I worked at Footnote, I used to routinely pester Peter Drinkwater about the shortcomings of the browse experience. Peter was kind enough to listen to me, and he used to tell me that they weren’t likely to do anything with browse because Search enhancements were going to replace it.

I remembered Peter’s comments while looking for my missing Frank’s, and it led me to try something else – searching. I thought, “I’ll just search LA directories for surnames starting with SH.” I figured that since FN keys the first surname on the page, that there might be some starting with “SH” that I could go to, and then browse from there – or even just page to the area where the Sharbroughs might be found.

I had a couple of problems with this approach. First, it wasn’t easy to limit the data to a specific directory – they’re all one title, right? So I decided to filter on year. That works, but the years are sorted in descending order of the number of images found for that year. It would make more sense to sort them in ascending order by year, no? I mean, you have to read the whole list, you can’t scan it.

Example: Years for LA Directories

In this example, you can see that the years are sorted by the number of pages, not by year. Finding a specific year, like 1909, is not as simple as I could wish. However, it’s not the kind of thing that would stop an Intrepid Researcher. There’s a workaround! If you type part of a year – say, “19″ – and press ENTER, you will get a list of years that start with 19xx, in numeric order.

Once I found and filtered on 1909, I was able to look for surnames starting with SH, and choose the page beginning with “SHANKLIN.”

First names on pages

In this example, I’m looking at images for 1909, whose “last names” start with “SH.” Notice that “Shop” meets the filter, even though the name starts with A B C. I didn’t find Frank in 1909, but in 1905 I did. There was Frank, living on S Flower St, 2 blocks from where his father Malachi B Sharbrough lived in the 1897 Directory. This method will generally take you right to the page that you’re looking for. In the 1921 directory, I assumed that “Shapiro Bros” was a classified page heading, but it wasn’t. The listing said, “Shapiro Bros (Isadore and NH) garment mfrs 546 S Los Angeles.” It was in the residential section – many commercial listings are. I should call it the “alphabetical listings” if I want to be more accurate.

I had problems

I had searched the LA city directories for Francis Sharbrough, but I was pretty sure that half of the entries were not OCR’d properly, and would require a browse to find the page for the Sharbroughs. It was a good idea, even if it wasn’t very practical.

I thought I would just “open the phonebook” to the right page. To do that on Footnote, you’d typically use Browse. That wasn’t as easy to do as I’d hoped. Each directory is identified in the browse structure by year, but within year, you have to go to the pages.

In this example, you can see that City Directory “browse structure” at FN is by state, then by city, then by year.

In this example, you can see the browse structure for the 1907. Pages that don’t contain lists of names are entered as [Gen. Info]. Each case is followed by the page number. If you wanted to get to page 500, you’d have to click on the numbers at the bottom of the right-hand column. The layout supports moving by up to two sets of 12 pages at a time (in this case, from group 1 to group 3), with a jump to the end and the beginning. When there are 2118 pages, as is shown at the top of the column, there are 177 “groups of 12″ in the directory. You could get to any desired page by jumping to the front or the back, then moving two “groups of 12″ at a time until you get to the page. In the hypothetical case I’m explaining, where the user wants to go to page 500, they would start at the beginning, and move to the 42nd group of 12. This would require about 20 clicks on the bar. One could wish for an easier way to browse, but Footnote executives and designers know that 99% of users search for names, and almost no one browses. It’s not as painful as having a zit on your nose. In the big scheme of things, it’s hardly to the status of irritating. And when I get those three wishes to change the world, I won’t be using one of them to get the best browse experience on Footnote.

In a perfect world …

There are two changes that I would suggest, after the caveat that it’s probably not worth the time and trouble to make them. First, I’d separate the book into at least 3 “sections” – the Front Matter, the Residential listings, and the Classified listings (the “Yellow Pages”). Some directories also contain a “cross listing” where they show all of the names on a given street, in address order. This would reduce “the fan” somewhat, as there are about 500 pages of front matter and 500 pages of classifieds. Then a user would only have a max fan size of about 1000 pages in the residential listings. Second, inserting a browse level by first letter of last name, so I could just jump to the “S” listings. I wouldn’t ask for a second level, the “Sh”-names. Footnote could create the initials through software, without having to hire someone to key those.

The biggest reason to add the “sections” is to separate the classified and the residential listings. Footnote made a little error, which we didn’t foresee when we defined the browse structure for the first directories. The classified sections are not organized alphabetically by business names. They are organized alphabetically by category, and alpha after that. Footnote has instructed their editors to key the first surname on the page, so instead of seeing “Cigar Makers” the page might be headed “S O Wilson and Sons” in browse. If you are ever at the “back” of the browse for a city directory on Footnote, you’ll see jumbled names in this section. It might be too late for the directories that are already online, but it seems possible for Footnote to change the instructions for future books, to key the business categories instead of the business names. It would not only be intellectually less offensive, but it would enable users to use the search method that I’ll describe below to find Dressmakers as well as Sharbroughs.

Again, I want to emphasize that there are good ways to find people in city directories, and that Footnote can find better ways to use their resources than to act on my idle suggestions.

Next: What Worked

It’s not a secret. I’m a big fan of City Directories. Footnote has a lot of them. But it’s not always as easy to find what you’re looking for as one might wish. I’ve spent some time fooling with this, and I have some suggestions for you.

I have a very distant cousin named Frank F Sharbrough. His great-grandfather, Malachi Sharbrough, was my grandfather’s grandfather’s grandfather. That’s not important. What’s important was that I knew a bit about the structure of Frank Frazier Sharbrough’s family tree, and wanted to see how the LA city directories could improve my understanding of his life.

Frank lived from 1884 to 1968. His father, Malachi Bridger Sharbrough, was a prominent Methodist minister in California. I’ve found him listed in newspapers a number of time, generally listing appointments within the Methodist church.

Looking for Sharbroughs in the LA city directories gets hits on both Malachi Bridger and his son Frank.

The Sharbrough hits on the LA City Directories told this story:

frank1

Frank was first listed in 1906, when he was 22. Over the next 20 years, he was a clerk, then a driver, then a plumber. He moved a lot until the end of the Great War, and was pretty stable after that. I liked having this information, but I couldn’t help noticing that there were some gaps in the years. Also, in 1922, his son Francis E would seem to have been listed at his address, but Francis E was only 12, so I’m thinking it’s just a mistake, and that the record refers to Frank F.

So I decided to look for some more records. Before I explain those results, I need to define two terms.

“OCR” stands for “Optical Character Recognition.” That’s one of the methods that Footnote uses to process the images in city directories. Footnote produces pretty good OCR results, but “pretty good” can still be pretty bad if the original image is made from film that is blurry, scratched, too dark or light, and so on.

The words that OCR produces are treated as “keywords” by Footnote. There is an important difference between “keywords” and “keyed fields” such as names, publication dates, place names, and the like. The important difference is how they behave in Search at Footnote. When processing City Directories, Footnote staff and contractors keyed the first surname found on each page. Those names are searchable as LAST NAME. The rest of the names on the page are in a “text blob” that’s just a long list of words. These words can only be found by using a “keyword” search.

“Fan” is a term used to describe an aspect of Browse. When browsing at Footnote, each list is limited to 12 items. In the case of city directories, once you choose the 1907 directory, you see the first 12 pages.

I suspected that the missing years contained listings for Frank that weren’t found because the OCR produced alternative spellings. All I needed to do was to find the pages that had Sharbroughs in the other years and I could prove or disprove my hypothesis – and hopefully fill in the story.

It’s no secret that BROWSE on Footnote can be problematic. One of the problems is “fan” – how many options you have to go through to find what you need. We page them 12 at a time, and 100 isn’t bad, but 200,000 is ridiculous.

Well, the “new search” works as a terrific “new browse” in my experience. I can use the narrowing feature to “winnow” the search results in a variety of helpful ways. And every once in a while, you find an image using “new search” and want to get back to BROWSE.

There is a shortcut. In the examples below, if you click on the thumbnails, you can see a larger image.

Let’s say that you are looking at this page in the viewer:

and you want to get to that same document in a BROWSE context.

If you look at the screen between the image and the filmstrip, you will see a “breadcrumb” – a “browse path” to the document:

Click in the breadcrumb, and you can jump to BROWSE at whichever level you click. You’ll see something like:

Note that in this example, you can click on one of three different levels, separated by chevrons (>>). They are “FBI Case Files,” “Old German Files 1909-1921,” and “Frederick August Stock.”

The FBI Case Files are one of the more difficult browses. You might find the same kind of trick helpful when using city directories, naturalizations, or newspapers. Using this trick, you can start a search with NEW SEARCH, and widen it with BROWSE.

Good luck, and happy browsing.