Reddit's image hosting implemented blocking direct linking to image files and started substituting an HTML-ish page like imgur does a few months back. After much complaint they actually fixed it so that the image in the hijacked URL's HTML is a real image tag (<img src="ht..) with a real URL to the image file. For a while it had been a javascript defined custom html element (web components) instead of an <img>.

So at least they did that right. imgur has been failing hard at this for ~5+ years.

Their design is to control how the uploaded images are accessed and minimize the load on the backend.

They don't want hotlinking so the image data doesn't have a permanent url. They only want the backend to be responsible for sending image data. This then requires an image viewer written in javascript. This is not a bad design. This is fulfilling business requirements successfully and efficiently.

EDIT: efficient for the business not for your browser

It doesn't just require Javascript to load the image, it requires Javascript from 13 different domains (thanks, NoScript.)

- imgur.com

- btloader.com

- ccgateway.net

- cloudflare

- facebook

- google analytics

- media-lab.ai

- sascdn.com

- scorecardresearch

- stretchsquirrel

- amazon adsystem

- assembly exchange

- run.app

3 of these, imgur, cloudflare, and medialab appear to be the minimum to load the image, and also seems to make everything on the page work (upvoting, commenting, etc.)

I don't think it's efficiency even on their end they're looking for, its tracking and revenue maximization. Which is the reason I block Javascript.

EDIT: actually it seems like you just need imgur and medialab unblocked, so just 2 of the 13.

> I don't think it's efficiency even on their end they're looking for, its tracking and revenue maximization.

These are not mutually exclusive. In fact both these things are the whole point.

> EDIT: actually it seems like you just need imgur and medialab unblocked, so just 2 of the 13.

Yeah exactly.