Hi HN,

For a long time I have been perplexed, why is that when we interact with text content like articles or blog posts on social media, we are expected to react to them in their entirety – either like the whole post, or dislike the whole post; agree with the whole post or disagree with the whole post. A post may contain multiple ideas we agree or disagree with independently, yet we are forced to generalize.

What our interaction could be like, if reactions could be attached only to particular selected parts of the text instead? Using classic reactions like buttons or emojis all over the text seemed like a dead end idea as even imagining it can make one's eyes hurt. But there's a traditional and universally understood way of text selection – marker-style highlighting. What if we simply convert such selection into a reaction? Not only highlighted parts can represent individual reactions, they also can be naturally aggregated into heatmaps, so besides expressing their own attitude, users are able to satisfy social curiosity – see what other people like or dislike.

This idea gave birth to Re:Likes:

https://github.com/reinventinglikes/relikes

Live demo (scroll a bit down to the playground area):

https://relikes.com/#demo

Additionally, created as the base of this project but found deserving to live as a standalone text selection library – Clean Selection:

https://github.com/cleanselection/cleanselection

To make the Re:Likes project usable on touch devices, I wanted to find a solution for text selection to be more user-friendly and visually attractive.

When we interact with the text type of web page content on a mobile device, all we typically can do with touch swipes is scrolling up or down. Some interactive elements of the page may want the horizontal swipe, but not the post/article area in most cases. When we read a text, horizontal touch swipes are just there, doing nothing at all, lonely and forgotten, begging to be used... So why not make them the standard way to select the text? What if, like an airbrush tool in a graphics editor, horizontal swipes could draw pretty soft clouds and select the text inside them? Thats what Clean Selection does.

Demo:

https://cleanselection.com

Particular use case examples:

Imitation of a forum interaction, selecting text to quote in response – https://cleanselection.com/forum/

Selection of multiple fragments from a long text – https://cleanselection.com/multi/

Appreciate your feedback about these, especially with the touch device experience~

Patreon Blocks AI Crawlers from Copying Content: 'Creators Deserve Compenstion'

Texas wins court order to suspend domain name for violating age-verification law

How Cathode Ray Tubes Work [video]

Kaiser nurses say AI, workplace surveillance are making their jobs, care worse

Solving Dexterity: A Full-Stack Approach

Maintaining the code of the man who wrote "How To Write Unmaintainable Code"

Facebook Anonymous

Netflix Paid $587M for Ben Affleck's AI Company

Torvalds: 'AI Is a Tool, Just Like Other Tools. It's Clearly a Useful One'

MG Siegler: 'OpenAI Makes ChatGPT ChatGPT Again'

HKT Launches Data Centre AI Superhighway in Hong Kong

A toy diffusion model for text gen using Karpathy's shakespear data

The CDC Has a Cyclospora Lab. DOGE Downsized It Last Year

WordPress – Polls for ActivityPub

The US Grocery Slowdown Is Real

AirDows – Instant P2P file and clipboard sharing between devices

HollowByte DDoS flaw bloats OpenSSL server memory with 11-byte payload

Claude Fable 5 vs. Kimi K3 vs. GPT 5.6 Sol [video]

Show HN: PrintBlocks – API and MCP server for your thermal printer

Investigating Flock: America's Mass Surveillance Company [video]

What does the Riemann zeta function have to do with the distribution of primes?

Show HN: A local browser extension that protects critical thinking from LLM's

JobLog

Two critical SQLi vulnerabilities in WordPress

Necklace

TP-Link Kasa cameras leaked home GPS via unauthenticated UDP for 6 years

US seeks share of Korean chipmakers' 'excess profits'

Show HN: Re:Likes – open-source library to be particular about reactions to text

Show HN: Learn hadware digital design online (VHDL, systemverilog)

Hegseth wants a "High-T" military; doctors call it a clinical minefield

Ask HN: Do you use LLM-Wikis?