# 202204181144 Licat on [[Obsidian]] Philosophy ## Description Licat describing the philosphy and core principles behind [[Obsidian]], discord chat excerpts. ## Notes - Core principles - Local-first data - Data in plaintext - Links are first-class citizens - Secondary principles - Extensible - Performance - Split screen workspace - Powerful search - fast fuzzy auto-completion - graph view - keyboard-centric - easy publish and sharing ## Conversations ### [[2022-01-16]] [source link](https://discord.com/channels/824179057620811806/824179205993922601/932426759335444540) > I can also offer my perspective if that helps. > > When we created Obsidian it was really a combination of core principles for a note taking app that we felt like should go together, but we haven't found a satisfying and polished app that had the combo of all of them, at least back in 2020. > > 1. Data being local first. > 2. Data being stored in a plaintext format. > 3. Links being a first class citizen, preferably using a simple-to-type format like `[[wikilinks]]`. > > To elaborate: > - Data ownership is extremely important for personal notes to avoid lock-in. A lot of companies are dismissive of true data ownership, and they will say something like "you can export your notes any time", which is very different than actually owing your data. They will also have reasons why end-to-end encryption can't be implemented in their cloud platform, often something like "well we won't be able to offer instant searches if our servers can't read your data". > - Part of the advantage of being proper/real markdown is that your notes will work with every other compliant markdown app out there should there be a need to switch. Non-supported syntax is gracefully unsupported (it'll show up as readable text). > - I think `#3` is what brings Obsidian from a text editor to a Tool for Thought ™️ It's also something Roam popularized right around the time Obsidian was created (~Jan 2020), which led to the boom in TfT apps. > > The other major part of being a plaintext format like markdown, on top of being local first, is that you can now live-interop with an unlimited amount of other apps, tools, and scripts that works on the local filesystem. Changes made with another app reflects instantly in Obsidian. Can't do something in Obsidian natively? You can use VSCode, Sublime text, vim, grep, awk, Windows Search, Spotlight, and thousands of other apps in tandem with Obsidian. Backup with literally any PC backup solution. > > I want to add here that there's a class of "local first, markdown based" apps like Bear that actually skips out on the benefit above, because they store all of your notes in a sqlite database or some kind of proprietary format. > > There were a list of secondary principles as well, on top of these three, which were initially discussed, planned, and scheduled, but not shipped until later in the development cycle, because we knew we wouldn't be able to do everything at once. > > - Extensibility. We've grown used to how customizable modern IDEs are for coding, and figured that this would be greatly beneficial for notes. At the very beginning from a technical perspective, we put in a lot of effort to architect the internal API such that it can be directly exposed to plugins to augment functionality within the app. A lot of optional internal functionality were written as "plugins" to alpha test the API, and also make it easily disable-able. > > We also designed the theming system to allow anyone comfortable with CSS to style the app however they want, with some core-defined variables used throughout the app. ### [[2022-01-18]] [source link](https://discord.com/channels/824179057620811806/824179205993922601/932882210980245525) > So just to put it down here, the other things we wrote down at the very beginning: > - Performance (I ended up spending countless nights performance snapshotting and debugging lag sources) > - Some kind of split screen mechanism, preferably very customizable (at first it was a forced left/right source/preview split, we spent about a month after public beta to come up with the workspace/pane system that exists today) > - Powerful search (support for complex combination of operators) > - Fast fuzzy auto-complete (quite a few apps lacked this for creating internal links back in the day, you had to type the note's name in full without errors!) > - Graph view with good looks (and fast) > - Facilitates keyboard centric usage (things like fully customizable hotkeys) > - Easily presentable (can be shared into a website/blog easily, this became Obsidian Publish)