The computer interfaces of my childhood
One of my motivations for building software like Gantry is that I love the computer interfaces of my childhood. MS-DOS 3.3, the pre-Windows era. What we now call TUIs running 'Character-Oriented Windows'. I grew up programming in QBasic and using Microsoft Works 2, early text-based Microsoft Office apps and other text-based apps for desktop publishing etc.
It was great. It had a stark kind of beauty. It was inspiring to see how far the software creators got within the character-grid paradigm. I think that was one of the interesting things, that using an interface like this, you are naturally aware of many of the trade-offs the interface designers faced, just because they are simple to understand and ever present in front of your eyes (how to economize with space on the grid, that you can only use text characters to create visual effects).
Graphical computer interfaces are inferior
We got Windows 3.0 installed a few years later but it never felt like home to me. Eventually I jumped ship to the Unix world, in no small part because I wanted a terminal-oriented operating sytem. These days I do almost all my work in urxvt terminals managed by i3. The things I use GUIs for are mainly web browsers and audio production tools.
And I think often about the fact that the text-based apps of my childhood were better than what I have access to today. My 10 Mhz, 1 MB RAM computer from 1989 had better TUIs than my present-day 3.7 GHz, 128 GB etc. Why is there not a text-based IDE as good as QBasic that I can run on a modern Linux machine? Why are there not multiple office suites implemented as TUI apps?
Because most people prefer graphical interfaces, I hear you say. Why, yes. But there should still be some people who want rich terminal apps for such things, enough to make up a market or an open source community. But maybe it's just me? Even if that's the case, I'll keep going. I'm dogfooding and daily-driving my apps right now, so I'm already getting solid dividends on the work.
Staking the company on the TUI
There is something magnificent about the software companies of yore having their own TUI toolkits, like Ratatui but 30 years earlier. They stood and fell with the quality of their proprietary system for implementing menus, windows, buttons, wysiwyg-editors, spreadsheets etc in the infertile land that was the text based operating system.
There is something about this that just strikes me as very romantic and beautiful. I, as well, want to stake my software business on my proprietary text-based UI toolkit. As it happens, terminal apps powered by modern toolkits like Ratatui are booming right now, some of them, the hot-shit coding harnesses, generating billions in revenues.
And there are some very strong technical points for TUIs. Notably, that they can be run remotely. You can ssh into a server and run a TUI there and have it render perfectly in your local terminal. Most people will not care about this at all, because they never ssh to other machines, but for some of the people who do, this is a killer feature.
So I decided to go for it
Some days I don't even care what the app I'm working on does, I'm just happy to be working on an old-school text based UI. But they are genuinly useful. Three of these apps I daily-drive: The editor, the agent sequencer and the file manager. I'm pretty sure I will be daily-driving the mail client soon enough. I also have a TUI app for backtesting trading algorithms, a multiplexing agent harness, an audio editing app and more.
This is my plan. To create a series of text-based apps, all using the same proprietary TUI toolkit. Including a full productivity suite. It will be a reimagined 80s software empire, reaching towards a full-fledged text-based operating system.
There you have it, my business idea. Now anyone can take it and run with it.
May the best text-based productivity software empire win!