Towards a personal knowledge interface: The sad state of personal data and infrastructure

I have been thinking of deriving value from what I read. I read a lot. A lot more than ordinary folks would care to admit.

A lot of what I read goes in the background thinking process. However, I haven’t been able to find anything worthwhile to integrate the process. There are several limitations of existing services which make it difficult to serve as an integration point. I won’t get into technical details (or the internet plumbing, to say). If you are interested, follow the link below the blurb.

Typically, once the service has your data, it’s siloed and trapped. You are completely at the mercy of service’s developers and management.

Within the same ecosystem (e.g. Google/Apple/Microsoft) you might get some integrations and interactions if the company spares them. Apart from these, integrations are virtually non existent.

The sad state of personal data and infrastructure | beepb00p

Ideally, annotations from the web or PDF’s need to make it through a revision system. I understand that software plumbing, API calls etc can become expensive as a majority of users remain siloed in various cloud services.

Lets start from annotations.

  1. Web Annotations- I could only find Hypothes.is as a worthwhile option. There are some other services that are in the education space but their software tools are clunky that interfere with the process on annotation itself. I am a stickler for a good user interface.
  2. PDF annotations- I had been using a Mac for a long time and the only decent software (that has a “standardised” system of annotations) is PDF Expert. There are several other services (like LiquidText and MarginNote) but these serve other niches. I need a system that would sync through Dropbox and is platform agnostic. So far, my choice fits the bill. I won’t recommend MarginNote because it is a Chinese app, forces  you to use a useless “search engine” and no sync options except “iCloud”. Get serious! iCloud isn’t even a decent service! I find it odd that Apple fanboys crow and whine about a service when better alternatives exist on the planet.

On mac, HighlightsApp finally got a refresh after three years, but like everything else, the developer slapped an expensive subscription. Besides, the Apple Pen support is buggy as the highlights jump off while scrolling. I didn’t have too much of a success in trying to get hold of the developer for “support” on Twitter and I didn’t bother much because I am planning to shift to Linux. Yet, the annotations, even if exported out in markdown (or any other format) to other services, require a great deal of cleaning up and work.

What to do with the annotations?

Readwise came up as a solution for “adding a second brain”; those ideas sound exciting when services push the dialogue towards “cognition” or “metacognition” and sound as “experts”. Readwise sends a daily summary of your annotations from several sources. It makes excellent sense to have a revision “flash card” like detail and personally, I like the idea. The best part is that they offer a generous 30 day trial that allows you to test drive it. However, the problems stem from a lack of PDF import. I was genuinely surprised to see a sync option with Hypothes.is (which can annotate on PDF.js in your browser) but doesn’t have any mobile interface!

I like the fact that ReadWise is bootstrapped and depends on subscriber contributions. I am told that PDF imports are “coming soon” but a lot of front-ends leave a lot to be desired. iOS App is clunky; seems like a browser wrapper. While the design is minimalist and focuses on what it does best, I think they can adopt a better way to explain concepts than YouTube videos. Does it justify a hefty $8 per month subscription? Not yet. There are a lot of rough edges and it is not ready for a prime time.

What next?

There’s an interesting service called as Histre. So far, the idea is brilliant- it serves as a collection of how you have scrolled from link to link. I was very lucky to connect with the developer on Telegram and we bounce a lot of ideas together. The premise of histre is that it helps you to visualise your search pathway. So far, it doesn’t add the “intelligent layer” to it, which hopefully is changing as I got a sneak preview on something exciting that is in the works. I am excited about it because it adds a meaningful layer to your search path (by use of tags).

Imagine if each of those tags had an RSS feed, for example. Lets say, you searched for hypofractionation, SBRT, Breast, Tumours (etc) and assigned tags to each one of them. Your RSS feeds would serve as a repository of information on your reader, you could automate a lot of things with IFTTT (like adding links to your Google Docs, for example) and compile the annotations to generate FlashCards/ pipe it through ReadWise. It opens up a world of possibilities. Each of those annotations could be shared publicly as well (Histre offers that option by tagging it as “pub”) and collect them through ReadWise for “revising”.

We aren’t there as yet for the “grand unified platform of knowledge” but I have been making a conscious decision to have a system that works in a browser as it is platform agnostic. I have also featured the services that preserve the integrity of your data and allow federation.

This is an developing thread, and I will revisit it once I have something in production.