Lately, I've been on a lost media frenzy recovering old things I made from years ago.
Though things like the Way-back machine have helped me recover some Scratch projects of mine, most are still very much missing.
People say that it's impossible to recover Scratch projects that have been emptied from your trash, but from what I've experienced, that isn't entirely the case.
The thing is, all of my missing projects, even the ones I've recovered, have been deleted for well over 3 years at this point. I've made great progress in recovering them, but now my search has hit a dead-end. There are no more archives that contain the project IDs of my projects, and I have no way of knowing the IDs of the projects.
That was until I thought about a theoretical solution, though it is well above my level of expertise and knowledge in programming.
If there was a way to make a tool that can skim through thousands, even hundreds of thousands, of Scratch project URLs that could find projects made by a specific user, and if they are not unrecoverable, then this could be huge for archival, not just for my own projects but for many projects out there that have been deleted with no hope of restoration. You'd have to be the owner of the project yourself to see the projects, given that they're deemed deleted + private.
I've thought about this for about a year now, and I have no idea where to look for people who know how to program at this level.
If there is anyone with knowledge in programming who knows what I'm talking about, please let me know if this is possible.
Even if the answer is no, thank you for letting me know.

(P.S: This is from the Request forums, I thought I should post it here given its more inquisitive nature.)

There are more than 1.2 billion projects on scratch at this point. It'll only take few thousand requests before you get either automatically or manually IP banned for spamming the servers.
And assuming you didn't, and you have fast internet, and you keep program running 24/7, it'd still take years to scan everything that exists so far.

I remember how in anticipation of 700000th project, someone made a live countdown that was using binary search and spamming servers with requests to determine which project exists, and they actually got banned in an unusual way - they lost ability to view any projects, including their own, or save any projects, but could still use the rest of the site.

Vadik1 wrote:

There are more than 1.2 billion projects on scratch at this point. It'll only take few thousand requests before you get either automatically or manually IP banned for spamming the servers.
And assuming you didn't, and you have fast internet, and you keep program running 24/7, it'd still take years to scan everything that exists so far.

You don't need to scan all projects though, only ones made after your account creation.

If there is a specific pool of URLs that you can scan from that'd be better (Ex: IDs from 160,000,000 to 160,001,000).
I didn't think about the spamming aspect of it though. I wonder if there is a way to do it that doesn't involve doing that.
Thank you for responding!

TheHappyGabeShow wrote:

If there is a specific pool of URLs that you can scan from that'd be better (Ex: IDs from 160,000,000 to 160,001,000).

You'd probably only need to scan from 600,000,000 to present (around 1,300,000,000)
That would still be pretty impossible. Even going at ~10 requests/second (which is still pretty high and may get you banned), it would take over 2 years.

I see.
Could there be any way around that?
Again, this is above my level of expertise, so I wouldn't have too much knowledge about this sort of thing.
If there were a specific pool of URLs I'd search from, it'd be from January 1st-November 1st, 2022.
Those would contain all of the projects I'm searching for.
That would also contain a massive amount of URLs. I wish that maybe someone could find a way to do this that doesn't involve spamming the servers.

Is there a way to see what URLs a specific user has made, that are already attributed to the user?
I've found a google service called “Google Search Console”, which appears to scan URLs of either a domain or a specific prefix.
I'd scan but I'm unsure if it will spam the servers.

Vadik1 wrote:

I remember how in anticipation of 700000th project, someone made a live countdown that was using binary search and spamming servers with requests to determine which project exists, and they actually got banned in an unusual way - they lost ability to view any projects, including their own, or save any projects, but could still use the rest of the site.

Kind of off topic, but that is a really weird way to get banned.

IS THERE ANYONE THAT HAS A COPY OF THIS https://scratch.mit.edu/projects/1051186106/fullscreen/ IF SO PLEASE SHARE IT WITH ME ON SCRATCH

Importedtaken wrote:

IS THERE ANYONE THAT HAS A COPY OF THIS https://scratch.mit.edu/projects/1051186106/fullscreen/ IF SO PLEASE SHARE IT WITH ME ON SCRATCH

please do not create off topic posts