Hacker News Reader

I thought the same thing. If you look at the crawled page, it's only one of like 20 images that survived.

So either the crawl got lucky and saved the only relevant image, or there is survivor bias.

Then again, the actual crawl might be triggered precisely because the image was linked.

What do you mean there is survivorship bias? That the only image used is the one that survived? Or it survived because it was used?

Something I noticed was that all other jpgs in this site have a lager number in the filename, for example: www.hobbytownoshkosh.com/Dsc00348.jpg

So maybe the crawler that saved this webpage had a limit on how many suburls it would capture and it sorted by name and then stopped at around Dsc00161.jpg, which is the name of the image in question. Though there is a Dsc00164 that is lost so it seems kind of unlikely...

That might have been the 161st image taken by an old Sony camera.

Cyber-shot model names use a DSC prefix, which is an initialism for "Digital Still Camera".

https://en.wikipedia.org/wiki/Cyber-shot

The filename for sure comes from the camera. My point was that the crawler stopped there and did not pick up the other images on the page because all of them have a higher number in the name and it stopped at an artificial number of sub urls

Reply on news.ycombinator.com