hnarayanan / gridpaper Goto Github PK

1. Once the app is functioning, test the app for common plot types and use visually interesting cases as initial example state.
2. Publicise the app on a few places (Hacker News, Product Hunt, Twitter…)

Take the source code of gnuplot, patch it and add custom wrapper JS to generate a powerful and lean WASM binary that can account for all the significant plot kinds and output formats that the app supports. This requires learning a lot about how emscripten can be used to compile gnuplot to WASM.

Go through gnuplot demos and identify commonly generated plot types.

For significant plot types, go through code used to generate them and figure out common aspects. Use this knowledge to create templates for significant plot types. These default templates should have beautiful default styling and should output in high quality to different formats.

These templates will help define the variables that need to be defined by the GUI.

The user interface could incorporate:

1. A list view, with examples one can draw from.
2. An edit view that also guides initial creation.

This interface will be used to set all the variables (within templates) and upload all the data that go into generating plots.

[notes photos]

Learn more HTML, CSS, JS (React) and Typescript (Blueprint) to layout the app’s user interface.

The interface allows users to define a lot of variables and upload data that go into the plots. The core logic needs to do multiple things:

1. Define models for graphs, multiple layers within graphs (functions or point sets), …
2. Convert the UI state into variables that go into the models
3. Merge the templates, variables and data files into a complete script that can be input to gnuplot
4. Take the output from gnuplot and render the plot on screen or offer a file for download
5. Persist all this in local storage so that the user’s state can be saved and restored

[notes photo]