HPI

Human Programming Interface project by karlicoss https://beepb00p.xyz/hpi.html

Why?

TLDR: I’m using HPI (Human Programming Interface) package as a means of unifying, accessing and interacting with all of my personal data. HPI is a Python package (named my)

The package hides the gory details of locating data, parsing, error handling and caching. You simply ‘import’ your data and get to work with familiar Python types and data structures.

Table of contents:
Why?
How does a Python package help?
Why don’t you just put everything in a massive database?
What’s inside?
How do you use it?

The main reason that led me to develop this is the dissatisfaction of the current situation

How does a Python package help?

the hardest bit is actually getting your data in the first place. It’s inherently error-prone and frustrating. But once you have the data in a convenient representation, working with it is pleasant – you get to explore and build instead of fighting with yet another stupid REST API.

Why don’t you just put everything in a massive database?

while databases are efficient and easy to read from, often they aren’t flexible enough to fit your data. You’re probably going to end up writing code anyway.

Here’s the (incomplete) list of the modules:

check out my infrastructure map. It might be helpful for understanding what’s my vision on HPI.

In essence, I’m mirroring most of my online data like chat logs, comments, etc., as plaintext. I can overview it in any text editor, and incrementally search over all of it in a single keypress.

orger is a tool that helps you generate an org-mode representation of your data

promnesia is a browser extension I’m working on to escape silos by unifying annotations and browsing history from different data sources

*Timeline is a #lifelogging project I’m working on.

I want to see all my digital history, search in it, filter, easily jump at a specific point in time and see the context when it happened*

Ideally, it would look similar to Andrew Louis’s Memex, or might even reuse his interface if he open sources it

Ad-hoc and interactive

My friend asked me if I could recommend them posts I found interesting on Slate Star Codex. With few lines of Python I can quickly recommend them posts I engaged most with, i.e. the ones I annotated most on Hypothesis.