Back to the Future of Handwriting Recognition (2016)

jackschaedler.github.io

43 points·at1as··9 comments

Oh hey, it’s Graffiti! https://en.wikipedia.org/wiki/Graffiti_(Palm_OS)

That Alan Kay video is unreal. It took h/w and materials engineering a fairly long time to catch up with the visionaries.

Yet I still cannot easily use handwriting recognition on my tablets, regardless of AI and machine learning algorithms everywhere.

It is still patiently word after word, correcting them every now and then.

And naturally only mostly works when using English, not any other human language that I would also like to use.

I can remember the early days of OCR and handwriting recognition. At that time, it was a difficult problem. Things have changed.

what is the state of the art now, that runs on-device? the article didn't mention the Q$ family, but I haven't found anything better than that plus mixing/co-judging with a dynamic-time-warping based approach

I tried to search for Q$ but google returned no results. What is it?

For my handwriting, the ParaGraph engine inside Evernote worked perfectly. I believe it's the same system that the US postal service uses.

https://depts.washington.edu/acelab/proj/dollar/qdollar.html

That looks pretty limited.

I found that AI can even decipher Suetterlin! Oh well, so much for my secret handwriting.

Reply on news.ycombinator.com