Greasy Fork

New feature: find derivatives of your scripts

To help script authors find unauthorized derivatives of their scripts, I've added a "Derivatives" tab visible to script authors. This will look for scripts posted by others with similar names or identical namespaces.

I would like to also search for similar code, but couldn't find an algorithm that could handle the length of the strings involved. Anyone have any ideas?

Just a suggestion - should the derivative search only search for scripts that run on the same pages as your scripts?

Personally I don't mind people make mods on my scripts. However this is a very nice feature.

Just a suggestion - should the derivative search only search for scripts that run on the same pages as your scripts?

One of the things this is supposed to check for is people using the name of your script with completely different (e.g. malware) code. I'm open to more suggestions on what this should check for, though.

Usually the description of the script is somewhat similar.

Example:
http://userscripts-mirror.org/scripts/show/152165 (Official Release)
http://userscripts-mirror.org/scripts/show/168708 (Malware)

I did a google on plagiarism and came up with this page "Source Code Similarity Detection Tools".

IMHO, if people want to plagiarise your source code, they will always find a way around like posting it on other sites. You really have to depend on peoples' ethics.

Thanks for the suggestion, I'm trying out CPD - https://github.com/JasonBarnabe/greasyfork/issues/228

Just like spam, you don't need to stop everything, you just need to stop enough to make your effort worthwhile.

I've added a "Derivatives" tab visible to script authors.

Why only script authors? As a user, I'd like to know if I'm looking at the original, a duplicate, or a fork. Example:

Original, outdated
https://greasyfork.org/scripts/1483-viewtube
https://greasyfork.org/scripts/1484-savetube
https://greasyfork.org/scripts/1485-linktube

Duplicate, up-to-date (fetched from the above author's site)
https://greasyfork.org/scripts/2984-viewtube
https://greasyfork.org/scripts/2983-savetube
https://greasyfork.org/scripts/2985-linktube

This doesn't really tell you if something is original, duplicate, or fork. All it says is that there are other scripts that are similar in various ways, without any info of which came first or if it's authorized or what.

My point was just that this would be a useful feature for the general public, not just something for script authors and their own scripts.

You may have noticed I do have a code duplication checker running now. I still need to work on it a bit as it's pretty hard on the server, so I don't want it to run unattended.