Strip out the ref URL search parameter

[robinmetral]

Prerequisites

• I checked the documentation and found no answer;
• I checked to make sure that this issue has not already been filed;
• This is not an ad/bug report.

Problem description

I use the AdGuard URL Tracking Protection list via uBlock but I've noticed that it doesn't strip out the ref URL search parameter by default. The list strips it out on specific domains only (see for example specific.txt at L22, L47, etc.)

Is this intended? Have you noticed that stripping out ref by default breaks too many websites? I've tried to find issues or docs about this but couldn't (expected to see it under Limitations of tracking protection).

Proposed solution

*$removeparam=ref

Additional information

No response

[Yuki2718]

Probably too risky

[robinmetral]

Is there any way we can test this assumption? I've been browsing with *$removeparam=ref for a while and haven't noticed any issues yet. If it's more commonly used for URL tracking than for necessary functionality, maybe it makes sense to strip out the param and whitelist it when needed?

[krystian3w]

This might start by grouping all ref deletions in same place while keeping their source comments.

This seems a bit against the new up rule, but has anyone checked recently whether websites have stopped using this parameter voluntarily?

[Yuki2718]

It's better to do in a personal list.

[krystian3w]

I only "wanted" to move the rule lines one below the other with a copy of their comments (so that line 22 has a smaller spread than the second occurrence in position 47 and already somewhere around 24), and not advertise risky filters without URL path and domain= delimiters in my own repo.

As I am not familiar with merging differences in git repositories while maintaining such sorting, I can not offer a mirror that would be more readable files for others.

[Yuki2718]

Sorry, I replied to @robinmetral, not you.

[robinmetral]

@Yuki2718 I have it in my personal list, as mentioned above, but this isn't the point.

The purpose of hosted lists like AdGuard URL Tracking Protection is to help users prevent URL tracking without them having to maintain personal lists. As a user, I'd like to be able to trust the list to protect me against URL tracking. However, many websites use ref in URLs in the same way they use utm_source, and the fact that AdGuard doesn't strip it out is going against the best interest of its users.

I believe that stripping out ref and whitelisting cases where it's used as necessary functionality is more maintainable and privacy-friendly that just ignoring the parameter. I understand it's a "risky" change, but anything that breaks will be reported and can be whitelisted—isn't this better than failing to remove a well-known tracking parameter?

That being said, if you're not willing to consider this, I'll be happy to report websites that use ref for URL tracking, so they can be blocked on a domain-by-domain basis. It just seems like the wrong way to do things.

[Alex-302]

Yep, it may cause many problems.

[robinmetral]

What problems? "May cause problems" is inherent to URL alteration, which is exactly what this list is for, for the benefit of privacy. If we don't know what the impact would be, why not publish a beta version that users can try, so actual problems can be found? Am I missing something?

Very sad that you're not even willing to engage and discuss. But 🤷

[Alex-302]

We have repeatedly added this rule for certain sites, and have had to remove it many times.

[robinmetral]

Still not answering my point about the fact that stripping out ref ought to be the default (i.e. using a whitelisting approach), since ref is a well-known tracking parameter. Don't you agree?

All I'm hearing is that it's a risky change, along with vague statements about "many" potential problems but no facts. And if we agree that ref ought to be stripped out by default, then I don't see why the risk should block this improvement, as long as it's done with proper steps in place to avoid as many issues as possible (such as a beta release).

Anyway—since this is clearly an unpopular request, I'll stop bothering you. If others find themselves here, I'd recommend stripping ref out in your custom filters and maintaining your own whitelist.

Thank you again for your work on this list, I still find it very helpful! ❤️