Features

Current

These features currently have been implemented but may need improvements due to bugs, oversights, etc. as they have not been accurately tested yet:

Character, foreground colour, background colour, bold, italic and underline cell properties.
Pencil, text, eraser, line, rectangle, flood fill, picker and select tools.
Apply character, foreground colour and background colour cell properties independently of each other.
Export as plain text, JSON, shell script and image.
Import as plain text and JSON.
Zoom in, out and reset the raster level.
Pan around raster using scrollbars, middle click, arrows keys and automatically follow the cursor when moved.
Raw text and cell copy, cut and paste.
Visual grid that can be toggled on and off.
Switch between themes.
Undo and redo.
Keyboard shortcuts.
Resize and auto-crop raster.

These features should be straightforward to implement:

256 colours.
Strike-through cell property.
Export with option to remove unneeded empty cells (i.e. an extension of auto-cropping).
Apply other cell properties independently of each other.
Palette of characters to select from (useful for non-ASCII characters).

These features are expected to be harder to implement and would therefore require research into doing so:

Support web browsers other than the latest version of Google Chrome, such as Firefox and Edge.
Drawing mode that translates mouse movements into individual cell characters to achieve "smarter" drawing. Could use angles between points and/or OCR.
Import shell script, some features may be able to be read easily such as colours but it could be hard to interpret other syntax that is not as relevant.
Allow changing the font to other monospace fonts on the client, JavaScript does not seem to have permissions to know the fonts supported in the browser though.
Handle larger raster sizes as the HTML canvas is quite limited in size and there is also very noticeable performance degradation when the amount of cells increases. A possible way to address this is for the canvas to behave as a window that only renders what can be seen by the client. This means that extra work is required to handle non-native panning of the raster shown in the canvas.