My solution to text editing on a mobile device is:
- Use a hardware keyboard.
- Install Termux.
- Install and use vim (or emacs, or nano, or whatever) to edit. That is, use a goddamned fucking text editor designed for the task.
The fact that devices similar to the 1990s-era Psion Series 5 are rare as hen's teeth doesn't help. This was pocketable (given sufficiently large pockets), had basic office productivity apps preinstalled (word processor, spreadsheet, database, email, contact and diary manager), as well as its own scripting language. It was optimised for power consumption and battery life. I know at least one journalist who used the device to file copy on the road --- this was no toy.
https://en.wikipedia.org/wiki/Psion_Series_5>
I'd love to see a modern update based on e-ink rather than LCD, as I wrote a few months ago:
https://news.ycombinator.com/item?id=36171805>
(I did receive a highly intriguing email suggesting such a device might be in development but have had no further follow-ups.)
There's a delicate balance between size, usefulness, and portability which is challenging to get right. In virtually all cases even the worst netbook-type compact laptop is superior for any significant composing task than the best smartphone or tablet. Touch-only interfaces are ultimately a dead-end.
The section of Jenson's article on how taps can be misinterpreted should be tattooed to the forehead of every UI/UX designer, PM, and marketer on the planet, as well as stenciled to all office walls, windows, doors, and whiteboards. I've railed about the ambiguity of selection vs. movement before (see: https://news.ycombinator.com/item?id=31771672>), but Scott highlights additional nuances. Does a screen touch: place the cursor? bring up the menu? start a drag? start a double tap? start a long press?
And what of when the user is disabled (Parkinsons, other motor/neuron disorder), tired, ill, drunk, in a moving vehicle (preferable not as the driver/operator), juggling children (don't drop), or in any other way at less than 100% focus, attention, ability, and dexterity?
I've had some variant of a smartphone or tablet since the mid-aughts. The two I've the fondest recollections of were the first (a Palm Centro, with a hardware keyboard) and my most recent (Onyx BOOX Max Lumi). Each has different strengths. The Centro was good for basic text entry, though its display was minimal. The BOOX is a large e-ink device which is excellent for reading and has Termux (a Linux userland for Android) installed. With an external keyboard it's passable local Linux system and can be used to SSH elsewhere, though the combined set is not something that can be crammed into a pocket. Slides easily into my messenger bag though.
As I'd listed out in table-form three years ago, there is no use-case for an emissive-display tablet for which it is superior to a different device, and the distinction is little changed for smartphones. Display + keyboard remains the ultimate productivity environment. For capability, privacy, and flexibility, I'd prefer separate audio, image/video, and comms devices, each of which individually can be quite small. Technological advances for each is slow enough that these need not be replaced frequently (once a decade or more should be sufficient). The main driver of late has been storage capacity, though with 1 TB being now widely available, that's no longer a driver for text, audio, or even to a large extent images, though video can of course consume prodigious storage.
The privacy advantages of airgapped generative / capture tools, with a separate comms-capable device also seem increasingly advisable.