# PDF importing

Eric experimented with Fable 5 + 5.6 Sol conversion to markdown

Julio experimented with https://www.datalab.to/ ... Will continue experimenting and will document the results in HM docs for future agents and humans

# Performance Fixes

hm://z6MkrbYsRzKb1VABdvhsDSAk6JK8fAszKsyHhcaZigYeWCou/:profile?v=bafy2bzacecoevvu2ebpntbrhm552uqp3ag5vrxvukpk7a4vvfmhazkk7smu5y&l will do performance testing for each of these 3 flows, to help prioritize

hm://z6Mkq9emq1yUBq4KSeiSH5yzgNBJSnidPVqFnTpzjCdLxB3R/:profile will build measurement tools for all 3 flows.

## Flow 1: Get the data onto the node!

Are we measuring the site load time on desktop?

## Flow 2: Load the content in the desktop app

the "local first" experience, everything should load in less than .5 seconds!

## Flow 3: Web: Load the content from the server from a fresh browser

Onyx and HM26

First: discuss the "why" of Onyx: Problem

Onyx = HM26

Introducing the Onyx schema system and deeply linking from the signed data is a huge step in the development of our protocol. This will be very painful or impossible to change in the future. But this will bring a level of self-documenting clarity and extensibility that we have been hoping for Schemas and Resources

Unified Schema system for RPC and Data?

Question from hm://z6MkgYR76i3w2Xc2fcvgFJWfYjsop6YwUonDhQx3AdLmQ4BQ/:profile

Because some types must be part of both RPC and data, a unified schema

Why Hypermedia Protocol

Why would anybody care about the Hypermedia protocol?