HN Explorer

11 comments

[ 3.2 ms ] story [ 33.0 ms ] thread
forgotpwd16
What's this wizardry?
tomschwiha
It's a PoC for HN related to this submission: https://news.ycombinator.com/item?id=33127285

Basically I guessed the id of the submission (HN has a global counter for each comment or submission).

forgotpwd16
>Basically I guessed the id of the submission

That's... kinda of letdown. Was expecting a more intrigued approach but apparently HN is slow enough for guessing to work.

tomschwiha
I actually was first thinking about writing a small script to check if an ID is in use and then auto create a submission.. however I first checked how often the ids actually change.. and in the moment I checked it was about one new comment/submission every 30 seconds.

I guess an timezone advantage .. when there is more traffic I guess the automation would win.

8n4vidtmkvmk
you can reference a YouTube video id in the video itself. that's much harder to do.
varun_ch
Here's a short write up about doing this. https://varun.ch/video-id :)

It works because there are two requests associated with creating a YouTube video. First, the metadata is sent (file size, video title, etc.) and it returns a video ID. Then the actual video content is sent. I cancelled the upload request and replaced the video with a new one I rendered quickly.

manfre
This is one of the reasons I prefer uuids for any public facing IDs.
lgreiv
Back when image boards had their run, using this technique to get posts with „rare“ combinations of digits [1] (e.g. „dubs“) boasting about it in that very post was a popular pastime. Some folks even built serious analytics models around it to achieve a pretty good hit rate. I haven’t been around for about a decade, so I do not know whether it is still a thing.

[1] https://knowyourmeme.com/memes/get

capableweb
Had the same idea yesterday (see my failed attempts at https://news.ycombinator.com/submitted?id=capableweb, unable to delete them after a few tries) but both the website + API were too slow to update for me to be able to guess the right idea. When the website showed that one ID existed, there were already multiple ones created afterwards...

Any clues to how you did this?

tomschwiha
I was basically just checking the latest ids (you can estimate it based on the /newest submissions and then increase until you hit the "No such item."-response.

In my case I guess I was lucky that there was not much traffic. However I was on mobile when I started observing the ids but finally switched to PC as I felt on PC I can react much quicker.