zlacker

The implementation here is somewhat interesting.

- Video IDs are spit out onto a Socket.io connection. (Another person claims it’s synchronized, which seems likely.)

- While one video plays, another player is in the background buffering the next video. Making it quite seemless.

- The code is from 2011, apparently, and it feels like it. You have code in script tags and plain old unminified JS, not to mention jQuery. Nothing wrong with that, but it’s almost nostalgic at this juncture.

So many of the videos it was pulling up had IMG/MOV/DCS in the title that I wondered if that was the strategy for finding unwatched videos, but I don’t think so, it must just be a consequence of many people uploading videos directly from camera files.

One remark I do have is that it seems to not be picking the most recent videos. There might be good reason for that (maybe waiting filters out bad content, or content that will have views?)

>>jchw+(OP)
"These videos come from YouTube. They were uploaded in the last week and have titles like DSC 1234 and IMG 4321. They have almost zero previous views."

That is from the initial page load. So it would seem that the title pattern that you observed is intentional

>>dhmill+hi
There are plenty of videos without that format. Those titles are just sequential file names of many cameras.

>>dhmill+hi
It's probably a strategy to find videos that were recorded IRL by real people.

There's a ton of content on YouTube that's generated automatically, as well as marketing videos, screencasts, etc. but those are not going to be nearly as interesting as something that someone recorded and uploaded by hand.