Checking changes made by new users to stop vandalism

Hi, I’ve noticed that there’s been a lot of vandalism recently - particularly the addition of very long lines all over the map.
That’s why I’m proposing, if it’s technically feasible, that the first X changes a user makes should be approved before being published, to avoid having a map full of vandalism almost all the time.
There’s already a “ask someone to check my changes” button, but it’s obvious that vandals aren’t going to tick it.


“ask someone to check my changes” is only a flag at the changeset. Nevertheless the data goes directly live in the database. At the moment, the only flag to hide data is the visible flag used to mark deleted data.

visible “true” or “false” Whether the object is deleted or not in the database, if visible=“false” then the object should only be returned by history calls.

How about limit new accounts?


I had the same idea but the vandals can work around this by creating good changesets until the limit is reached.

Different level restrictions would make it harder for vandals but still feasible with time or bots.

I want to point out that your “but” is not something I’d see as actual criticism to the proposal of limiting new accounts.

Your objection is like saying it makes no sense to give people driving tests as bad ones could just study hard for one day and still be crappy drivers.

It’s not entirely wrong, but it is not right either. Having those investments of time and effort are most definitely known to have a positive effect. For starters it will massively limit the inflow of accounts that can do vandalism.

The linked proposal additionally states we can limit the amount of actual damage that can be done by having some system that rewards good behavior with more power. At the highest level of power comes the ability to approve new changesets. So a bot will not be able to do much damage and will not get their changesets approved which is needed to start doing more damage.

You curb most of the damage right there.


Limits on new accounts may help to prevent accidental and incidental damage.
The current repeted vandalism attacks are something else entirely. Limits are easy to avoid, all it takes is some patience and maybe a bunch of harmless edits. The gates of OSM are wide open for vandalism. We’re lucky the damage was crude and easily detectable, next time they probably will get more inventive, less obvious.
We’ll probably need AI to detect abuse patterns, but even then the damage will already be in the database, propogated all over the OSM-cosmos.


You should read the actually linked to proposal, as I think most people disagree with your assesment.

They are actually not possible to do on a new account and it would take serious actual human effort (not bots) to get a mature account.

Ask yourself this; why are new accounts given full access to the entire database of OSM with basically no limitations?

Detecting damage is what the dwg has been doing for a year or more, it is not actually solving anything. We revert and block. While there are still hundreds of thousands of accounts out there that are capable of doing the same tomorrow.

I think that if a new account doesn’t actually have rights to edit a large area, does not actually have rights to make more than X new nodes (or changes) per day, you actually solve it at the core. Damage like this week is impossible without spending weeks per account actually doing real valuable work.

Sure, someone may still do the damage after being productive for weeks, but then the gain is still greater than the loss to OSM. :man_shrugging:

Take a look at the initial suggestion for more on how this isn’t just a “bot can do the work” kind of approach. How about limit new accounts?


There’s an assumption here that someone would do weeks of good work before doing the vandalism, therefore a net gain. What if they do weeks of bad work, with lots of small deletions or degrades before the big one? How do you catch the small stuff? While not so obvious, it could be far more problematic long term.


They already found an (obvious) workaround for the existing edit limit (and they are not even using that to the full extent possible).

There is no panacea for this, at best we make things slightly more difficult, but in no way impossible.


The saying goes: How high should the wall be that surrounds your belongings? Answer: Higher than the one of your Neighbours. It is about the amount of effort, kind of a proof of work in crypto currency. Whether those are

or just garbage does not matter much. Certainly trash easier to create, but also might get spotted early on?

Which would probably be spotted, reverted & reported by the always vigilant community :+1:, resulting in them being warned, & if they persist, blocked.

1 Like

You are assuming that there are many volunteers in every region of the world who are willing to review new account changesets (“thumbs up” as per the proposal). This assumption might only hold true in a few countries with a large base of active mappers.


For example, in my area I am the only active contributer and if someone else begins I probably will not check their edits

I was thinking little things like minor building shape changes, misspell a business name, noleft turn instead of no right. All over the place. How are those one offs spotted? Just for my understanding really, i can understand sytematic analysis of edits to spot behaviours, but can individuals really spot every deliberate error?

So the original proposal made a year ago still is not implemented (lack of developers or focus, I don’t know).

The solution of limits alone isn’t ever going to be good enough, as many here point out. Also as the vandals show.

The original idea was to combine limits with the social aspect. Let real people judge the quality of a contributor. An unjudged or new contributor has lower limits and they will thus be able to do much less damage. Less changes per day, smaller areas only.

The beauty of using the social aspect (which is the hard part to get OSM to accept) is that actually experienced mappers are supposed to help inexperienced mappers anyway. Simple handholding, a pat on the back when doing things well. And it then also becomes really easy to recognize mappers that have more experience and knowledge and thus approve their changesets and effectively give them higher limits.

Limits can rise based on a new mapper being actually more productive and a net positive. If their changes are not that, they will not get higher limits. As simple as that.

A lot of new mappers would love to get an experienced user saying they like the change. Or a quick correction if not. And moving in that direction is in my opinion the real solution to the vandalism that we’ve been experiencing for far too long now.


Comrades, we need to get real here. This is not the 1990s. Unless proper protections are put in against cyberattack and the creation of automated vandal acccounts, OSM will be destroyed.


People determined to damage will find a way. The best we can do is put reasonable roadblocks that won’t turn off those here to be responsible. I’m sure the are other procedures thay could be used to prevent and flag vandalism.

Any proposal will go nowhere unless someone is willing and able to implement it.


I dont know of anything formal. Though I could suggest some type of change profiling system. Something can flag edits out character for a bound box. Development Seed just demoed some related tools at this year’s StoM US.
It should be able to notice drastic changes to more established areas. Especially things like
name changes to places or streets. Multiple additions and deletions of ways in a single changeset might be sogn of vandalism. It could just warrant a message to a new mapper explaining why the practice is considered a bad habit.