How I might keep track of a bunch of old letters
I have a bunch of family letters. The letters in the attic from this site. I wish to set up some sort of systematic way of digitizing them. This post is an articulation of a possible scheme for doing so. It’s not even a proof of concept, it’s just the concept– a plan of action thinking through what I’d like to try to do.
Scan all those letters? I am a masochist. But if I want to scan them all and somehow make sense of them digitally, that makes me a systematic masochist.
The letters are mostly to my grandmother. Some are to my grandfather from his sister and parents. But they’re basically the attic-box correspondence with their family. My great-grandmother wrote her daughter weekly. In addition to the news of doings from home and her teahing jobs, the dispensing of advice and how to cook this or that, my great-grandmother’s letters include fascinating details such as prices for objects, newspaper clippings, and a prodigious dose of nagging “Why don’t you write me?”, and more.
When dealing with physical materials, one archival rule of thumb is to keep them in the order they were found in the first place. Those who stored ‘em this way had some sense or order to them, so maintain them in that order. Or you file them according to date, or by sender and date. The point is, you need to have some organizing scheme to keep track of them all.
Once you scan a letter, that store-in-some-order requirement is no longer necessary.
So. The project. Scan letters. Which become images. But, once they’re scanned pixels, how to organize them or name them? I thought about MemoryMiner and its ability to attatch a place and one or more persons to an image, and supply a date for that image. Why not do the same for a set of scans of letters? But wait a second. But can you attach two places? Because the letter starts in once place, say, Billings Montana, and it travels to another, say, Schenectady, New York.
In an IM conversation with John Fox, Mr. Digital Asset Management, I discussed the particulars. Alas, MemoryMiner will only attach one location object to an image. So the location will be the location the letter originates from. But then he told me something I didn’t know about in my fifteen years of using Photoshop.
You can put metadata into a photoshop file. That’s right, add text to a photoshop file with descriptions, and the text becomes a part of the file itself. Sweet! Then when you bring it into MemoryMiner, the metadata is automatically imported along with the image. Doubly sweet!
I think there’s other software besides Photoshop to add metadata to images; when I first looked into this, I created a set of bookmarks for metadata in my del.icio.us collection. So that’s a starting point for research if you’re not a Photoshopper.
Since I use Photoshop, I’ll dig into more specifics> In Photoshop’s File menu is a command, File > Info… That opens up a dialog box where you can put metadata in the following categories:
Caption—has a text area where you can write whatever you want to describe the image.
Keywords— creates a list of different keywords.
Categories—add a category or list of categories for the image.
Credits—byline,title, credit, source
Origin—lots of information about location and date. Good for where the image originated and when. Will this be the letter’s origination or destination point?
Copyright and URL—Mark as copyrighted or not, and give an image URL.
Photoshop has a “save” and “load” section of the dialog box, so once I come up with the right metadata format, I can save versions for G-Gma Fannie to Gma Florence, and load that up with each new scan, and fill in the date and subjects discussed. I can create preliminary metadata files for each correspondence relationship of who writes whom and from where.
The beauty of this undiscovered metadata country is that the information is retained with the image itself. John Fox tells me that MemoryMiner will look at the metadata and import and display the information when it imports an image. And MemoryMiner will (either in the just-released 1.2 version—not tried yet; later today is boyfriend’s mother’s second memorial service; I’ve been busy—or in a later version) search in the metadata.
So, with that, here’s what I think I will do:
- Scan letters to photoshop .psd document. Each item will have to be scanned: envelope, each sheet of letter. I should invent a short naming scheme to keep letter elements together. Something like FTFtoFFB-210304-1.psd FTF to FFB is initials of who wrote to whom, then the date 21 for 1921 210304 is for March 4 1921. Putting the year first means that all images when listed alphabetically ‘ll list in order of creation. Oops, no, I’ll have to reverse the date and who-to-whom: 210304-FTFtoFFB-1psd
- Open in photoshop and do basic post-scan image adjustment, cropping, rotation.
- Add metadata. I could, in the caption, type out the basic gist of the letter. Topics discussed, etc. Make a tag for enclosure or newsclipping that’s included with the letter. Maybe even create a base-level metadata object. You can import and export them from the metadata dialog box. I think that the date of the object will be the postmark, if that can be read.
- In one of those metadata areas, put the letter’s destination, in addition to its source.
- Save the image, import into MemoryMiner. See if John Fox was telling the truth (!) when he said that MM imports all the metadata.
My first set of tests will help me find out how and what information in the metadata is imported into MemoryMiner. MemoryMiner assigns “location” to an object. Does it get that from the “Origin” part of metadata? If so, then I’ll need to decide how I’ll indicate the letter’s destination. Keywords? categories? This will take some playing around before I create a system.
Once I get things figured out, I’ll be able to arrange letters this way and that. Letters sent from a particular person. Letters sent from a particular place. I can focus on a narrow date range to see letters sent during, say, my grandmother’s college years. Or marriage before children.
MemoryMiner, since version 1.1, will save sets of images as a separate library. That’s good, because I wish to keep a correspondence library separate from the main photos library.
There it is. An idea. Now (cough, cough) all I need to do is implement it.
UPDATE: After some careful consideration, I decided to go for an analog method of processing the letters. It can become digital later, but the perfect is the enemy of the good, and the perfect is the enemy of done, and I’d prefer to do it.
This is very interesting. Some suggestions:
(1) Make sure that the file name is encoded in such a way that the file name itself can id the letter.
(2) Use ISO date form for easier sorting.
(3) Add a few key tokens to the file name
Also, using sub-folders for this within a single over-arching folder would be useful IMHO.
While the PS idea is very good, I would store as HiRes JPEG or JPEG2000 as the PSD files are not a universal standard. Metadata does live in JPEG files.
Sorry for more confusion.
You may want to consider something like PaperPort (http://www.nuance.com/paperport/standard/). It’s a bit pricey (about $100) but it will convert scans to PDF format and handle multi-page documents. It also provides organization, search and share funtionality. I have found it a valuable tool for organizing and managing not only family documents, but also important personal papers - keeping scanned copies of documents you normally store in safe-deposit boxes.
I have found it a valuable tool for organizing and managing not only family documents, but also important personal papers - keeping scanned copies of documents you normally store in safe-deposit boxes.
I’ll try your methods.
But if I have some transparency - what can I do to organise them?