editors

In times past, I would painstakingly learn about a single technology and keep trying to apply it until I find the bottleneck and give up, but now you csn spawn 10 agents in parallel to make demo implementations ro explore look snd feel.

It boils down to how we represent your text content. The most common example is a text buffer, also known as a continuous array of characters backed by a file. You need to implement all text operations on top of the text buffer snd any fancy functionality like inserting a TODO has to be implemented by scanning the text file for patterns and replacing them

The other extreme is block based where each row of the text is a different json object with extensive metadata. There are hybrids like Codemirror allowing parsing of a text buffer

After spending way too much time with note taking apps, you will come to see patterns of festure sets across apps. Clusters of apps would have very similar capabilities based on the libraries that were used. This makes some use cases basically free (and thus.. everywhere) but committing to a certain library or framework can completely block you from another set of features.

While we say you can duck tape everything (any festure to any framework) with sufficient code, I have come to realize there are fundamental limits to this. I would like to spend this blog post exploring the choices we can make when creating a web-technology based note taking app and how it would affect what you can create.