Top all-time · #2857
danvk.org blog
Personal blog · danvk.orgprogrammingdata
danvk.org blog is a personal blog at danvk.org about programming and data. Since 2008, 27 different Hacker News readers have linked it 47 times, most often “Training an Ocropus OCR model”. It is still writing: the last post was on 2026-09-10, and you can follow it by feed.
Too few sampled comments (22) to say how often HN recommends it.?
Why HN links it
HN readers link danvk.org for posts on OCR (ocropus text extraction and training), Boggle board analysis, dygraphs charting, and historical NYC and SF map projects.
Summarised by Claude from a sample of 22 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Most-linked posts?
- Training an Ocropus OCR model5 people · 5 mentions · first linked 2015-01-19
- Dygraphs: Better Time Series Graphs3 people · 3 mentions · first linked 2009-12-10
- /wp/2007-04-06/nebulabrot/2 people · 2 mentions · first linked 2010-07-23
- AI helps add 10k more photos to OldNYC2 people · 2 mentions · first linked 2026-04-06
- /dygraphs/2 people · 2 mentions · first linked 2009-12-22
What Hacker News says?
I've had really good luck recently running OCR over a corpus of images using gpt-4o. The most important thing I realized was that non-fancy data prep is still important, even with fancy LLMs. Cropping my images to just the text (excluding any borders) and increasing the contrast of the image helped enormously. (I wrote about this in 2015 and this post still holds up well with GPT: https://www.danvk.org/2015/01/07/fi…
Comment linking /2015/01/07/finding-blocks-of-text-in-an-image-using-python-opencv-and-numpy.html
on “Llama-OCR: Document to Markdown”Submitted Dygraphs: Better Time Series Graphs as a story · 181 points
HN discussionFinally someone's instantiated something like Vernor Vinge's "evocation" idea, quoted here: https://www.danvk.org/wp/2014-02-10/vernor-vinge-on-video-ca...
Comment linking /wp/2014-02-10/vernor-vinge-on-video-calls/index.html
on “Show HN: HiFiC – High-Fidelity Generative Image Compression – demo and paper”Submitted After 20 years, the globally optimal Boggle board as a story · 78 points
HN discussionYep. Back when I was at NYPL Labs[0], we developed a bit of an obsession with OCR. It turns out that so much of it depends on the context of the documents, original scan quality (unpaper and deskewing be damned), typography, how much handwriting you're willing to suffer (can someone please pick up the work of applying Convnets to large handwriting corpora that aren't just MNIST?). A lot of this comes down to how muc…
Comment linking /2015/01/11/training-an-ocropus-ocr-model.html
on “Ask HN: Open source OCR library?”Submitted Extracting text from an image using Ocropus as a story · 63 points
HN discussionhttp://www.danvk.org/2015/01/11/training-an-ocropus-ocr-mode...
Comment linking /2015/01/11/training-an-ocropus-ocr-model.html
on “Ask HN: Open source OCR library?”Submitted Mapsnap: Automated Georeferencing for Historic Sanborn Insurance Maps as a story · 22 points
HN discussionActual links instead of <censored> redirection: http://www.danvk.org/dygraphs/ http://www.highcharts.com/ http://cartographer.visualmotive.com/
Comment linking /dygraphs/
on “Dygraphs: Better Time Series Graphs”Submitted Breaking 3×3 Boggle as a story · 7 points
HN discussionThere are simpler ways to calculate a bound that don’t involve trees. You can read about the sum and max bounds in the WIP paper: https://github.com/danvk/hybrid-boggle/blob/main/paper There are some examples in these old posts: - https://www.danvk.org/wp/2009-08-08/breaking-3x3-boggle/inde... - https://www.danvk.org/wp/2009-08-11/a-few-more-boggle-exampl... These bounds are pretty effective at finding the global ma…
Comment linking /wp/2009-08-11/a-few-more-boggle-examples/index.html
on “Lone coder cracks 50-year puzzle to find Boggle's top-scoring board”Submitted What’s Going on with Twin Rates? (includes a visualization) as a story · 5 points
HN discussionThat would only be if the dice order didn't matter. Multiply by 16! for the number of ways you can place your rolled dice in the grid. Then you find that 26^16 is less than 6^16*16! https://www.danvk.org/wp/2007-08-02/how-many-boggle-boards-a...
Comment linking /wp/2007-08-02/how-many-boggle-boards-are-there/
on “Lone coder cracks 50-year puzzle to find Boggle's top-scoring board”Submitted Training an Ocropus OCR model as a story · 4 points
HN discussionHis blog post announcing it: https://www.danvk.org/2025/04/23/boggle-solved.html FT article: https://archive.ph/siaAO His blog: https://www.danvk.org/blog.html > Driven “by the thrill of discovery”, Vanderkam has searched for this board, essentially alone, since 2004. He would scrape together computing time on Google’s hardware for heavy Boggle computation, all along documenting his efforts on his blog. > “As far as…
Comment linking /blog.html
on “Lone coder cracks 50-year puzzle to find Boggle's top-scoring board”Submitted Introducing Imnowave - Crossword puzzle collaboration with Google Wave as a story · 4 points
HN discussionWow, that's an amazing level of dedication to mapping this all out. Looking at the about page, it looks like an good exercise in machine learning. The developer talks about the process of generating the test data here: http://www.danvk.org/wp/2013-04-20/generating-training-data/
Comment linking /wp/2013-04-20/generating-training-data/
on “A Street Map of New York City in the 1800s”Submitted AI helps add 10k more photos to OldNYC as a story · 3 points
HN discussionI really enjoyed OhLife. Reading your own writing is the best way to improve it. OhLife helped you do just that. The one additional feature I'd like to see in an OhLife replacement is the incorporation of some cues about my day to help me get writing. Show me my calendar entries, photos, check-ins, etc. to jog my memory. This would be especially helpful when back-filling in entries a few days later. It would also di…
Comment linking /wp/2012-07-01/ohlife/
on “OhLife is shutting down”Submitted Arbitrary Precision Hex Dec in 65 lines of JS as a story · 3 points
HN discussionHere's a great example of something that is most natural in mapreduce: http://www.danvk.org/wp/2007-04-06/nebulabrot/
Comment linking /wp/2007-04-06/nebulabrot/
on “MapReduce from the basics to the actually useful (in under 30 minutes)”Submitted Boggle Roundup: My Fifteen Minutes of Fame as a story · 2 points
HN discussionThis one is pretty fun: http://danvk.org/dygraphs/ - the best part IMO is that the charts are interactive.
Comment linking /dygraphs/
on “Ask HN: What charts javascript library to use?”Submitted Nothing is ever so simple as it seems: Developing the OldSF Slideshow as a story · 2 points
HN discussionYou can read about some of the recent development work on OldSF here: http://www.danvk.org/wp/2013-01-21/developing-the-oldsf-slid...
Comment linking /wp/2013-01-21/developing-the-oldsf-slideshow/
on “OldSF – Pictures of San Francisco from 1850-2000”Submitted Which States should support the National Popular Vote Bill? as a story · 2 points
HN discussionHave you looked into ocropy[0]? Here's a nice intro[1] that later talks about how it achieves higher accuracy using an LSTM model[2]. [0] https://github.com/tmbdev/ocropy [1] http://www.danvk.org/2015/01/09/extracting-text-from-an-imag... [2] http://www.danvk.org/2015/01/11/training-an-ocropus-ocr-mode...
Comment linking /2015/01/11/training-an-ocropus-ocr-model.html
on “Show HN: Convert scanned documents into searchable PDFs”Submitted Puzzle+: Crosswords in Google+ Hangouts as a story · 2 points
HN discussionYou'll enjoy my follow-up post then, which talks about training: http://www.danvk.org/2015/01/11/training-an-ocropus-ocr-mode...
Comment linking /2015/01/11/training-an-ocropus-ocr-model.html
on “Extracting text from an image using Ocropus”Submitted Sunrise/Sunset Onebox (a Google engineer's story) as a story · 2 points
HN discussion
