Forum TI-Planet.org

At this point, I'm wondering if it's possible to add some functionality similar to Cesium's flash drive support and USB8x for the z80 series to the TI-89. (Titanium, because it has a USB port.)
Specifically, I'm wondering if it's possible to "mount" disks as folders somehow.
Of course, I'd be writing most of it, using BrandonW and Lionel's Linky as a starting point.

Any thoughts?

Linky is a TI-89 Titanium version of usb8x :
viewtopic.php?f=41&t=11769#p139751

Indeed, you had already found the starting point you need to use: Linky. All I did in Linky was fill in the keyboard detection and mapping code, as well as make it easier to switch between HID keyboard and HID mouse modes.

Yeah. Tari was nice enough to mirror BrandonW's repo on github (instead of SVN.)
If you recall, I asked you for the binaries a while back because I wanted to see what it did, but now that's unnecessary because I managed to build GCC4TI (I can't seem to find the KDE3 stuff, so I can't use KTIGCC, though. They're not available from apt.)
I already started looking at the specs for FAT32, and it's a mess. There are a boatload of different versions, from different operating systems. And there's the fact that most consumer flash memory is hideous and just the cheapest possible thing.
(Hamming codes, multi-bit cells, and the like)

Yes, making a good implementation of FAT12/16/32 is hard. I suppose you could leverage one of the existing ones... once you've found one which has sufficient portability, reliability, etc.

As for KTIGCC: just do what virtually everybody has done in the past decade or so, don't bother with this thing
Besides its adherence to an unpopular, heavyweight, poorly portable desktop environment and application framework which breaks API compatibility between minor releases, it suffers from both significant implementation issues - the code's quality is terrible, making it unmaintainable - and the fundamental design issue of trying to reimplement verbatim the... peculiarities... of the Delphi IDE, instead of fixing them.

After some searching online, I've found a couple possible starting points.
Unfortunately, even after examining the Linky source code for about an hour yesterday, I still don't entirely understand how it works.
One of the criteria for the starting/example implementation is that it has to be released under a GPL-compatible license. (There's something wrong with writing a proprietary program for a graphing calculator, even more so than a normal computer, because they're supposed to be learning tools.)
As for KTIGCC, whatever. I'll stick with my current setup of using tmux.

Unfortunately, even after examining the Linky source code for about an hour yesterday, I still don't entirely understand how it works.

Neither do I

The USB controller is indeed memory-mapped, in the 7100xx range.

Well, I'm going to take a break from this project to write an IDE that I can use to make it.
The progress thread is over on Cemetech:https://www.cemetech.net/forum/viewtopic.php?p=292032#292032