For all who are unaware, there’s a phone number filter in scratch that will block all strings of numbers that are 10 numbers long and replace them with Xs.
At some point, Scratch project URLs will be 10 numbers long, and I’m concerned that this’ll lead the project links being blocked by the phone number filter.
What will Scratch do when the time comes? Remove the phone number filter? Whitelist scratch project links?
Semi-related, but there have been a few times where scratch project links trigger the filter, but it doesn’t seem to be because of the phone number filter. It might be that the numbers somehow resembled a bad word? But I dunno.

This is a good question. People could use this to bypass the filter in the future, so this definitely needs to be considered.

they could scan if its a actual callable number then put it for review if it is

randomguy3513 wrote:

they could scan if its a actual callable number then put it for review if it is

Yes, but this would still result in a lot of false positives.

EDIT: I have an idea to fix this. The ST could just skip over all of the projects that are phone numbers, right? More phone numbers will be made, but it will still be somewhat effective. For example, if my phone number is 123-456-7890, then that project ID gets skipped over so that no project of it exists.

I asked @thisandagain this exact question way back in the day. He indicated that changes to the filter would need to be made to ensure that Scratch project links don't get blocked by the filter. There is no doubt that the Scratch Team is aware of this, but I imagine they're not really concerned about finding a solution until we get a little closer to that milestone.

musicROCKS013 wrote:

This is a good question. People could use this to bypass the filter in the future, so this definitely needs to be considered.

Imagine someone making a project, and the project number just happens to be their phone number. Like it wasn’t even intentional, the project would just be their phone number by pure chance.

Eg let’s say my phone number is 12345678910 (it’s not btw), and the project I made just happens to be the 12345678910th project ever. It’d be really weird.

And someone could theoretically disguise their phone number as a scratch link, too… saying “my number is scratch.mit.edu/projects/12345678910” and they’d get away with it since it’s a scratch project link. Hmm.

M1mikyu wrote:

And someone could theoretically disguise their phone number as a scratch link, too… saying “my number is scratch.mit.edu/projects/12345678910” and they’d get away with it since it’s a scratch project link. Hmm.

Yes, that’s what I was talking about. I added an update of how to work around this issue. I think that it would be unlikely you would get the project, but you wouldn’t even need to get the project to use the ID in a link.

I did a little calculation…

I made a new project, and the ID was 703343278, which is 9 digits, and I set the variable X to this.
Then, I did X + 40, which is my estimate of how many projects are being made every second.
Now, I need to find out how many seconds it's gonna take for it to become a 10 digit number/1 billion projects, so now I have X + 40 x S (for seconds).
Then I did a little plugging in numbers and experimenting with them, and I came up with the number 7416418.05, and that made exactly the number 1 billion.
I searched up how many days 7416418.05 was, and this was the answer: 85.838171875.
This is kinda specific, so I'm just gonna round it to:
85 and 4/5 days left until the 1 billionth project!

Maybe I'm doing the calculation wrong, and if I am, then someone please correct me!

It will be about 469 days.

RED-001-alt wrote:

Maybe I'm doing the calculation wrong, and if I am, then someone please correct me!

You should check out this project.

EDIT: According to it, it will be about 469 days until the 1,000,000,000th project.

RED-001-alt wrote:

85 and 4/5 days left until the 1 billionth project!

Maybe I'm doing the calculation wrong, and if I am, then someone please correct me!

That seems a little too close. The Notes and Credits on project 600000000 seems to indicate that it was created sometime before January 19th 2022. That was about 5 months ago. Recall that project 700000000 was created on about June 1st 2022. Continuing this trend, I would expect project 1000000000 to be created sometime in summer 2023 — much longer than 85 days (~3 months).

Note also the relatively long time it took to even reach 600000000 projects — almost 15 years! I don't know how much traffic will be on Scratch in the next year (is the pandemic over yet?), but I wouldn't be surprised if it takes even longer than a year to reach 1000000000 projects.

You ever spend half an hour attempting to use regressions on Scratch statistics to accurately predict when 1,000,000,000 projects will be made, only to realize that the statistics only count shared projects at the very end? I know I have.

well then it seems we will just have to deal with the fact that people might know that someone with a random phone number happens to be on scratch

if someone doesnt want to see or hear nsfw things dont click the link

musicROCKS013 wrote:

randomguy3513 wrote:

they could scan if its a actual callable number then put it for review if it is

Yes, but this would still result in a lot of false positives.

EDIT: I have an idea to fix this. The ST could just skip over all of the projects that are phone numbers, right? More phone numbers will be made, but it will still be somewhat effective. For example, if my phone number is 123-456-7890, then that project ID gets skipped over so that no project of it exists.

that wont work because people are going to hate scratch a whole lot if 10 people get a annoying call from scratch every second to see if its a active phone number or not

randomguy3513 wrote:

they could scan if its a actual callable number then put it for review if it is

According to this Wikipedia article (which probably isn't accurate, I know), there are about 8 billion phone numbers in the world. I'll use a much smaller number, just in case that number is bloated (which it probably is), so 2 billion. There are going to be about 9 billion projects that could have a project ID with 10 digits (1,000,000,000 - 9,999,999,999), so the chance of a project also being a phone number is around 22%. Of course, there are other complications, such as specific area codes not existing, so we can reduce that number to 10%.

10% is pretty bad, considering 1 in 10 project links will need to be put up for review. A simple solution would be just to see if https://scratch.mit.edu/projects/ comes before the link. Then, there's no need for searching for active phone numbers, but the ST is going to have to improve the filter a lot to combat trying to get around it.

Za-Chary wrote:

Note also the relatively long time it took to even reach 600000000 projects — almost 15 years! I don't know how much traffic will be on Scratch in the next year (is the pandemic over yet?), but I wouldn't be surprised if it takes even longer than a year to reach 1000000000 projects.

Although the pandemic isn't over, most places are going back to in-person learning, like the US. However, that doesn't seem to have impacted the number of project creators or monthly project shares.

EEP020 wrote:

that wont work because people are going to hate scratch a whole lot if 10 people get a annoying call from scratch every second to see if its a active phone number or not

Well it would be annoying, but it would only need to happen once per number, so it wouldn't be spamming one person over and over.

musicROCKS013 wrote:

EEP020 wrote:

that wont work because people are going to hate scratch a whole lot if 10 people get a annoying call from scratch every second to see if its a active phone number or not

Well it would be annoying, but it would only need to happen once per number, so it wouldn't be spamming one person over and over.

It would cause Scratch many legal problems if they were to spam call every person with a phone to see if they would pick up. Not to even mention that it's probably going to take an absurd amount of time to do so.

The live count project isn't accurate

960times10_5is106044 wrote:

The live count project isn't accurate

If you're talking about the project count on the Statistics page, that page only shows the number of shared projects, as far as I know. However, you can find the total number of projects easily by clicking on the create button at the top of the page.

The number in the link is how many projects came before the one you made.

RL1123 wrote:

randomguy3513 wrote:

they could scan if its a actual callable number then put it for review if it is

According to this Wikipedia article (which probably isn't accurate, I know), there are about 8 billion phone numbers in the world. I'll use a much smaller number, just in case that number is bloated (which it probably is), so 2 billion.

I don't think the number would be that durastically inaccurate. I know Wikipedia isn't always reliable, but people often exaggerate how inaccurate it is. A difference that large for the total number would require nearly the entire article to be vandalized, which almost definitely would have been caught.

As @Za-Chary said, changes will have to be made to the filter in order to maintain the balance of fairness and safety we already have.

Speaking of which…

Za-Chary wrote:

That seems a little too close. The Notes and Credits on project 600000000 seems to indicate that it was created sometime before January 19th 2022. That was about 5 months ago. Recall that project 700000000 was created on about June 1st 2022. Continuing this trend, I would expect project 1000000000 to be created sometime in summer 2023 — much longer than 85 days (~3 months).

However, that's just a hypothesis.
We don't know when project number 700000000 will be created shared, nor do we know when project number 1000000000 will be created.

Za-Chary wrote:

Note also the relatively long time it took to even reach 600000000 projects — almost 15 years! I don't know how much traffic will be on Scratch in the next year (is the pandemic over yet?), but I wouldn't be surprised if it takes even longer than a year to reach 1000000000 projects.

First of all, yes. It's true that it will take so long for the 1000000000th project to be created.
However, one thing we do know is, there will be enough room in the MIT servers for that amount of projects to be created.

Second, noooooooo, this pandemic isn't over.

I hope someone suggests this in time to come.

k0d3rrr wrote:

Za-Chary wrote:

That seems a little too close. The Notes and Credits on project 600000000 seems to indicate that it was created sometime before January 19th 2022. That was about 5 months ago. Recall that project 700000000 was created on about June 1st 2022. Continuing this trend, I would expect project 1000000000 to be created sometime in summer 2023 — much longer than 85 days (~3 months).

However, that's just a hypothesis.
We don't know when project number 700000000 will be created

Huh? As of right now, the project ID is about 703,491,000. So, the 700 millionth project has already been made. Because the real 700 millionth project isn't shared, we can use a project that's kind of close. In this case, I found a project with ID 699,999,904 that was shared on June 2 EST.