[NTLK] Love Notes ?2? Newton

[Matthias Melcher]

> On 24. Oct 2024, at 05:41, Larry Yaeger <larry.yaeger at gmail.com> wrote:
> 
> Maybe, but I doubt it's anything terribly sensitive or important, if there's any at all.  iOS is a very different OS, and none of the handwriting recognition code persists.  (And the handwriting recognition code has now been excised from macOS as well.)  The data structures and code we used for encoding and exploring our dictionaries/word-lists did get picked up and used for the Mac's "data detectors", and may have spread to iOS's spell-checking and word prediction, but I suspect even that low level stuff got rewritten at some point.  Concepts from the Newton's power management might persist, but I doubt the actual code does.
> 
> However, Apple doesn't even have all the source code.  At least if I remember correctly, Paragraph's cursive recognizer was always delivered as a binary and no one at Apple ever had access to their source code.  I think Apple should have the rest of the source code though.  I think.
> 
> It'd sure be great to get the source they have, to clean up a few minor problems.
> 
> And if source remains a no-go, at least it would be nice get a blessed release of a final generation ROM image and permission to put Einstein in all the App Stores.
> 
> Documenting those Quixotic quests could be pretty darn interesting, and who knows, having the process recorded might make someone decide to step up and be the hero that makes it happen.
> 
> - larryy

Thanks for the insights, Larry. I don't think that the missing Paragraph code would be an issue at all. There are great OpenSource recognizers out there, and if we were to ever write a NewtonOS for other platforms, we could just link one of those. 

My main reasons for having the source code (apart from just curiosity) would be fixing the year 2040 bug that is creeping up on us. A continuous fix for the Y10K bug would also be great. Lastly, if we could recompile the entire OS for other platforms, we could finally go from emulation to a native system, reducing CPU load by a factor of 50 and gaining the ability to actually add features.

As usual, here comes my little geek-out: we *have* the debugging symbols and quite a list of headers. The Assembler code is not optimized, it's really easy to read and understand. It is very much possible to write a disassembler that generates compilable C++ code for probably 80% of the Newton ROM. The remaining 20% would need to be done by hand. There is already tons of rewritten NewtonOS source code on GitHub.

So the question is, should we finally start a concerted effort to recreate the Newton source code? I know that there are only a hand full of people here who can write C++ and/or NewtonScript, but if we have a strategy and a goal, maybe it is doable with modern developer tools?

IANAL, but from my European point of view, the device is over 25 years old, and at least the Newton and MessagePad name patents have not been renewed in over 20 years. From my unqualified point of view, NewtonOS is abandonware, and I have no qualms to make the devices that I bought over the years survive beyond 2040 by fixing them, even if that means I have to understand the inner workings.


 - Matthias