Evernote OCR: A quick look

As a doctoral student in the late 1980s, I began reading biographies of scientists. The first biography I read was Charles Darwin’s Autobiography ($0.00 in Kindle store). Early in the book (p 54 L 800) Darwin talks about his organization system and concludes with:

"... by taking the one or more portfolios 
I have all the information collected during 
my life ready for use."

After reading this I sat back and thought “Whoa! What is today’s equivalent to Darwin’s organization?” and shortly thought “A database.” The next day I purchased the only portable computer I could afford, a Radio Shack TRS-80 Model 100 with 32K of RAM and a four-line, 40-column display.

I would type in the good passages of every article, every book, and every magazine that I read. Then, when the Model 100 was full (about 4 hours of reading and typing) I would ride my bicycle from MSU’s library to my office in the bowels of Hubbard Hall and upload the data to my desktop computer via RS-232, then erase the Model 100’s ram, and bike back to the library and repeat the reading and typing.

Because laptop computers were totally out of my reach, after building a text database in AskSam of all these passages, I printed out all the key passages I had read on 4″x6″ cards in 8 pt courier font. I picked 8 point font in order to squeeze as many characters on a card as I could. I printed approximately 1,200 of these cards on my HP DeskJet (1988) printer. The DeskJet entered my life in 1988 when first introduced, I think I paid $800 for it. My wife had pre-authorized the purchase of an inkjet “… as soon as it is under $1,000.

Here is an example DeskJet card:

I kept the cards in long boxes and then went through the boxes repeatedly, card by card, and making connections across cards. These connections were then captured on other 4″x6″ cards in hand written notes. I have approximately 800 of these “linking” cards making bout 2,000 4″x6″ cards in total. Here is what a linking card looks like:

Going paperless with Evernote, I scanned in all my 4″x6″ cards and then recycled them. The purpose of this blog post is to show the results from a quick investigation of Evernote’s optical character recognition on my machine-printed and hand-written cards.

Evernote OCR: Machine-Printed Cards:

First, the recognition on machine-printed cards. I “tested” evernote by opening the card and then typing the content of the card into the search box for that card. Here is what that looked like:

You can see the text I’ve typed in to search for, in the upper right hand corner of the note, and the yellow rectangles in the note indicate recognition hits.

The result is that Evernote has a pretty hard time on 8 point Courier font text. After doing this quick test and thinking about it, I’m going to have to re-read these cards in order to sift through their content. I can’t count on Evernote to find words for me when they are printed small. This is not a criticism. Evernote is always growing and adding capabilities. I just need to keep in mind the current capabilities in accessing my information.

I’m not just being pollyanna about Evernote. When I first scanned my files, I had probably 8 gigabytes of files to upload, and could upload only 1 gigabyte per month. At the time I wrote Evernote and said “Can I please give you money so I can add all my gigabytes?” to which Evernote replied “Thanks for offering, but not yet.” And within 12 months, they created the “gigabyte amnesty” program where you can pay $5 to upload a gigabyte of extra stuff.  So I’m patient with Evernote.

Evernote OCR: Hand-Written Cards

Next, I tested how well Evernote could read my hand-writing. No *flames* please for my awful writing. I gave up on cursive at the first possible moment, and bought an electric type writer to substitute. Consequently I “print-scribble” rather than write.

Opening the note for the card at the very top of this post, I obtained this hit list after typing in every word on the card:

Again, there are quite a few hits, but, a text search is not yet ready for prime time. As of 3/2014 I must read rathe than search information captured in hand writing.

Using Evernote is a Janus-thing. Janus was the Roman god of doorways, beginnings, and transitions.

Using Evernote is great. A dramatic improvement to my organization and productivity. What I’m already accomplishing is face #1 of Evernote. But, as an Evernote user, I’m on the cusp of doing so much more. The so much more is face #2 of Evernote.

Click here to see Liz Parrish’s two headed elephant. The Janus-experience of using Evernote.

bill meade


2 thoughts on “Evernote OCR: A quick look

  1. Pingback: Links Roundup #19 - Personal Knowledge Management for Academia & Librarians

  2. Pingback: Janus-ed in Evernote | RestartGTD

Leave a Reply

Your email address will not be published. Required fields are marked *