I have one, and I find the writing part nearly perfect. But there are a couple of downsides mentioned in the article that mean I barely use it:
- Searching and organisation is really difficult. I wish everything would be OCRed transparently and you could instantly search for it (while keeping the original graphics)
- It's unfortunate that the OCR runs on their cloud. It would be really perfect if it was a web app that you could install on your own servers, if you have hightened security requirements for example
- In the UI, there is an on screen keyboard and you have to press keys. Why can't you just write in the text fields?
I think most of these problems come down to the fact that the OCR is some "secret sauce" provided by a third party. I wonder if there is any viable free handwriting OCR one could use instead to build a better experience (open source, source available, or even just some research papers)?
I have the same issue with the OCR. I've been using it for taking notes for my masters courses, and ironically as part of the course just finished making a machine learning model for OCR on armv7 devices that works as well if not better then the proprietary OCR service they use, and could easily run on the device without needing the network at all.
My biggest issue with the rm2 is not any specific feature, but the total and complete lack of support for 3rd party software. The people behind the device make great hardware, and nice device drivers, but kind of suck at user facing software. They should stop trying to do the front end stuff, and just give us the ability to let the community do it ourselves.
If you made this into an open source project (even just some hacked together scripts) I would definitely take a crack at helping tweak it! Especially if you're able to do on device -- that'd be incredible. Otherwise, I'm happy to just pull out the files and do it via SSH.
Seriously though, the value of some scripts that set a baseline would be enough for many people to finish the rest of the work!
Wanting your documents and writings to not be fully accessible to CCP intelligence services....is not exactly "heightened security requirements" to me.
This device is basically completely un-trustable.
Leaking information related to financial stuff, research, invention, product design, or legal/government work. For a publicly traded company any manner of documents / notes on a device like this could spell trouble.
Plus, if they're able to identify something you don't want others to know, they have leverage on you.
- Searching and organisation is really difficult. I wish everything would be OCRed transparently and you could instantly search for it (while keeping the original graphics)
- It's unfortunate that the OCR runs on their cloud. It would be really perfect if it was a web app that you could install on your own servers, if you have hightened security requirements for example
- In the UI, there is an on screen keyboard and you have to press keys. Why can't you just write in the text fields?
I think most of these problems come down to the fact that the OCR is some "secret sauce" provided by a third party. I wonder if there is any viable free handwriting OCR one could use instead to build a better experience (open source, source available, or even just some research papers)?