no support, i don't see the point of removing them.

codeman1044 wrote:

I know that they seem broken for you, but from my experience, if something seems broken, it's way more often the coder's fault than the computer.

I go to cloud data history pages and there's nothing there on projects many people have played. It's definitely broken.

I think it only breaks when too many people use it. It works on projects where few people have used the cloud data for me. Maybe the ST should remove the option to create them until they fix them enough to work on all projects.

No support, cloud data is useful despite its limitations.

No support, cloud variables are really good and have never broken for me. If you want to report a broken cloud variable, check the Bugs and Glitches forum.

miniepicness wrote:

Sometimes cloud data doesn't work. It doesn't seem to work at all for me. Cloud lists are already ruined, why not get rid of the option to make cloud variables so people don't try to use something that doesn't work. I might have just messed up how my project loads the cloud data, but it seems to always work wrong on any of the cloud data that I use.

Edit: it seems to break on projects where many different people have used the cloud data

TERRIBLE idea.

Cloud data was implemented back at the launch of 2.0. Since launch it has been mostly doomed since the start. Check out this clip from a video showcasing the features of 2.0

https://www.youtube.com/watch?v=VPGmV8ejWP4

The major cloud data feature showcased is cloud lists, which as we know now only existed for a small time during the alpha/beta and never made it to launch.

Regularly, the cloud service simply stops working, and requires attention from the Scratch Team to get it working, due to stress on the cloud servers and issues with saftey of Scratchers, cloud data is severely limited by forcing them to be 256 digits or less. This makes them very difficult to be used outside of simple “high score counters”.

Often, not even highscore counters work, leading to a generally bad Scratch experience, especially for new Scratchers who might place the blame of something not working in their projects on themselves. Which could lead to them simply not thinking they're “good enough” at coding, even though the issue lies in Scratch. If this happens, it's probably safe to say that Scratch failed at one of it's goals of educating people about computer science.

One increasingly common use case for cloud data is modifying them using external scripts. While this is fun and educational, it also opens the door to dangerous projects that could, for example bypass cloud restrictions to do stuff like saving a picture from the video sensing feature to an externally accessible database or maybe a web browser that could be used to access community guidelines breaking web pages.

I propose that cloud data is removed for the following reasons:

• It requires regular maintenance from the Scratch Team, which eats away at time and resources that could go towards other projects like working on the Scratch editor or website..
  
• It is a safety risk, exposing users to unmoderated chat rooms, external scripts doing dangerous things and any variables simply being easily externally modified due to the nature of how it works
  
• It is a security risk, leaving a higher attack surface on the Scratch websites. A service with no features is a service with no vulnerabilities.
  
• Many users don't even know it exists, due to it having to be limited to full Scratchers.
  
• It can't do many of the exciting things it was planned to be able to do, the way it's presented to users makes it only viable as a high score counter without lots of work going into converting digits to strings and back. Some projects even have to use compression algorithms to get around the digit limit.
  

It would be sad to see such a powerful feature go, but in it's current state, Scratch would be better without it.

Obviously, removing features and not compensating with other things isn't cool, so some kind of local variable storage (via localstorage or cookies) would be a useful replacement. While this suggestion isn't about that though, it would be a simple extension to add and would not require much maintenance as everything is just stored locally in the user's browser. It also wouldn't have any of the safety/security risks of cloud data outlined above.

Removing the cloud variables would be total chaos and i'm sure it would create a worse scenario than 3.0 studios. Multiplayer projects, highscores and more would be gone so it probably isn't a good idea to make this decisinon.

Jeffalo wrote:

This makes them very difficult to be used outside of simple “high score counters”.

*ahem* 1 2 3

Jeffalo wrote:

it also opens the door to dangerous projects that could, for example bypass cloud restrictions to do stuff like saving a picture from the video sensing feature to an externally accessible database or maybe a web browser that could be used to access community guidelines breaking web pages.

If that is such a common issue, why isn't the scratch team even fixing the new scratcher cloud data bug? They can fix that instead of pulling down an entire aws websocket server.

Very no support, they can be used to make soo many projects, just imagine how many projects would be broken if this was implemented???

Jeffalo wrote:

It is a safety risk, exposing users to unmoderated chat rooms, external scripts doing dangerous things and any variables simply being easily externally modified due to the nature of how it works

The scratch team is there to moderate those projects, if it was like this, then all of scratch should be removed since anyone can upload any image or link to any website inside a project. so i suggest removing projects just because of this I say this isn't a valid reason.

AntonL1kesPotato wrote:

Removing the cloud variables would be total chaos and i'm sure it would create a worse scenario than 3.0 studios. Multiplayer projects, highscores and more would be gone so it probably isn't a good idea to make this decisinon.

Yes, me and other people who love cloud variables would cry. Do you know how many unshared projects I have that use cloud vars? All my months of hard work, imagination and debugging would be gone. :'(

edit - turbowarp will help! but there isn't a community I can share this with, so yeah, I'd stick with scratch, and I'd stick with my no support on this suggestion.

i'd like to know what motivated you to make such a long suggestion about removing such a feature, which would obviously cause controversy more than what the new studios created.

[removed—read my newest post]

First I had a heart attack reading this but after reading your points I do have to agree.

Jeffalo wrote:

ften, not even highscore counters work, leading to a generally bad Scratch experience, especially for new Scratchers who might place the blame of something not working in their projects on themselves.

I have done this. Before in platformer games, I have made, I have got frustrated with how on and off the high score system was. Sometimes it worked - other times did not. I thought it was not working because of me and that made me stop on that platformer game.

-
I also agree that cloud data gives a lot of work on the server, causing lag to the site, especially projects that use an external server, which causes issues itself Seeing it go would put some ease on the delicate servers, and the ST is already implementing updates to help site performance so this could be another step to that plan.

I did recently found out how much matinence scratch puts on this one feature. Removing this, as you said, would free up resources, which may mean quicker bug fixes and updates. I do believe these new updates will make up for removing this feature.

Jeffalo wrote:

Often, not even highscore counters work, leading to a generally bad Scratch experience, especially for new Scratchers who might place the blame of something not working in their projects on themselves. Which could lead to them simply not thinking they're “good enough” at coding, even though the issue lies in Scratch. If this happens, it's probably safe to say that Scratch failed at one of it's goals of educating people about computer science.

I have to agree with this, when code is correct, it doesn't work. Which is sad, and I say that scratch should improve cloud data
Whilst improving the system would be great, how can we improve it whilst also making it easier on the servers? This will also take up even more resources for one feature.

But, cloud data is not just used for high score counters, it is used for something else:

MULTIPLAYER GAMES
With this, multiplayer games will not work. Slither IO and other impressive projects that show what you can do with scratch would break. Cloud projects do show what you can do with scratch, and tonnes of projects have used cloud data in different ways. This goes with the scratch philosophy of wide walls and high ceiling. Removing cloud data would affect the “scratch room”. However, the new features could regain this.

Have you thought about adding more restrictions to cloud data once again? More ease on servers and cloud variables would still exist. Rules with cloud variables could be used, like projects that use external servers would be banned

Overall, I think this is a sacrifice that could be avoided for now, maybe with more restrictions, but if stuff gets really bad for the server, it could be worth taking so scratch can have less stress on servers plus more resources.

These new resources should run client-side instead of server-side, and this removal would make all scratch projects run 100% client-side. So features that had no cause to lag and are useful to the editor, restoring the wide walls and scratch room, is the best replacement

Jeffalo wrote:

Often, not even highscore counters work, leading to a generally bad Scratch experience, especially for new Scratchers who might place the blame of something not working in their projects on themselves. Which could lead to them simply not thinking they're “good enough” at coding, even though the issue lies in Scratch. If this happens, it's probably safe to say that Scratch failed at one of it's goals of educating people about computer science.

I have to agree with this, when code is correct, it doesn't work. Which is sad, and I say that scratch should improve cloud data

Chiroyce wrote:

Jeffalo wrote:

Often, not even highscore counters work, leading to a generally bad Scratch experience, especially for new Scratchers who might place the blame of something not working in their projects on themselves. Which could lead to them simply not thinking they're “good enough” at coding, even though the issue lies in Scratch. If this happens, it's probably safe to say that Scratch failed at one of it's goals of educating people about computer science.

I have to agree with this, when code is correct, it doesn't work. Which is sad, and I say that scratch should improve cloud data

it's either improve or remove, and the scratch team has shown that it makes more sense to focus resources on the rest of the site, rather than improving cloud data.

RIP the appel leaderboards.

But actually, support as the cloud data system has been creating a lot of downtimes, and causing annoying stuff [I'd like to see scratch use reliable code for this (idk whats causing downtimes) and open sourcing the clouddata system so bugs can be easily fixed if they actually accepted suggestions for cloud].

Jeffalo wrote:

it's either improve or remove, and the scratch team has shown that it makes more sense to focus resources on the rest of the site, rather than improving cloud data.

Yeah maybe, but straight up removing it isn't a good solution.

9gr wrote:

But actually, support as the cloud data system has been creating a lot of downtimes, and causing annoying stuff. I dont think the scratch cloud data server is even under fastly anyways.

Nope, Amazon Web Services, Jeffalo has the link to it but I lost it.

Chiroyce wrote:

9gr wrote:

9gr wrote:But actually, support as the cloud data system has been creating a lot of downtimes, and causing annoying stuff. I dont think the scratch cloud data server is even under fastly anyways.

Nope, Amazon Web Services, Jeffalo has the link to it but I lost it.

yeah just checked, it defo looks like it

9gr wrote:

Chiroyce wrote:

9gr wrote:

9gr wrote:But actually, support as the cloud data system has been creating a lot of downtimes, and causing annoying stuff. I dont think the scratch cloud data server is even under fastly anyways.

Nope, Amazon Web Services, Jeffalo has the link to it but I lost it.

yeah just checked, it defo looks like it

If you search “elasticbeanstalk” AWS comes up too

support, i think at this point that it makes sense. cloud variables are meant for highscore, not for what griff has been doing with them. notice anything about my post Jeffalo? *hint hint*

No support. A lot of projects will no longer work and you cannot do cloud projects. lnstead of removing them because of the issues, they could fix them.

It will also be spammed in all the forums to bring them back, if it happened.

Support, I like cloud data but I like a transformation extension even more and that is what will probably be implemented if this was removed