Possible Way to Detect Copies of Tracks

I was just surfing on the forum at 1 in the morning, reading over comments and such about various topics when suddenly the topic of copied tracks was brought up. I remember way back when FRHD was in its earlier stages of beta when people were wanting some kind of system to detect copied tracks. Many people suggested just simply comparing the codes to see if they are the same, but all a copier would have to do is draw a line in a random place in order to bypass that. I couldn't help but think, however, that a way to compare the tracks would be by splitting the tracks up into sections, and comparing those sections. At first I thought that maybe dividing a track up into multiple segments would be a little much, but then it occurred to me that this has already been accomplished in the track editor, as you can put the grid on and only certain sections of the track (20x20 units, I believe) will have a grid background.

Is it possible to use this grid section method to compare two tracks in order to determine if they're copied? I'm not entirely sure how the code behind all of the tracks works, but there must be a way to take each line, put it into a section, and then compare that section to the corresponding section on another track.

Of course, if this were to be conducted every time a track is uploaded, it would be very inefficient to go through ever line of every track in the database. To save some efficiency, the system can maybe just investigate the first 4 segments surrounding the rider, and then only move on if 90% of the lines are identical to the other track? There are definitely numbers to play around with in order to get the right efficiency. Also, maybe the system can only be conducted if a certain number of players flag the track as copied? That may make the most sense.

Another bug in the system would be lines crossing between two sections within a track. (Such as the default line the rider sits on - it crosses evenly between two grid sections). I guess the system can just reject these lines, if it comes down to it. Assigning these lines to a section would become confusing.

Lastly, the biggest bug is outside of the system. Which track came first? Tracks still don't have a publish date on them. What's up with that? Hopefully the date of each publish is kept somewhere inside the database.

I'm just spitting out every idea that pops into my head at the moment. I'm sure I have more ideas to come, but it's 1 in the morning and I'm so very tired. I'm also tempted to come up with some prototype of this system as a personal project - but I will save that for tomorrow.. Or later today.. Whatever.

This thread is obviously mainly directed towards my homies all up in Vancouver, but I'm definitely open to suggestions, comments, questions, clarifications, violent concerns, etc. from everyone else.

 

Almost all of the stolen tracks were made before this site, a publish date would mean nothing if the original author never published the track here.
And let's say he did, but he took a long time to do it. Would his track be refused because someone posted his track before he could?

 

YOU ARE A GENIS!!!!!!!!!!!!!!

 

Sono, you are wrong here. there is a track # on the web-address bar: http://www.freeriderhd.com/t/53323-dirt-lab/r/pancakes345

3r4nd0n does have a point though

 

Ahh.. Didn't notice that. I guess that would work. But then again, 3r4nd0n has a point.

 

Track code is split in 3 sections for lines scenery and powerups and if a copier added a line then it would go to the end of the line section in the track code and leave the rest of the track code unaffected, so a track code comparing system would work if it took a snippet of code at the start of the line section and looked for that same snippet of code in other tracks (by doing the equivalent of ctrl + f using that snippet to look at other track codes). I guess it might be a problem if a copier deleted lines so maybe this method could be repeated by comparing several snippets of track code.

 

Very interesting.. So I guess my idea could still be used when it comes to comparing several snippits of track code? It's just whether or not using the 20x20 sections method would be the easiest way to approach the situation.

 

NOBODY CARES WHAT YOU THINK SONO! YOU JUST NEED TO SHUT UP IF YOU HAVE NOTHING CONSTRUCTIVE TO SAY

 

Ohhhh someone is bored, I see.

 

we need a dev to comment on this as they know how the game works so should know what the best method is.

 

lol thanks for using my track as an example

 

Ummm... why don't we do a 90% match check? If they draw one line it's still a 90%+ match. Colleges use this to prevent copying of papers.

 

This could still have flaws.. All you need is the copier to go off really far away from the track and draw a whole bunch of lines.
I like the way you are thinking, however.

 

Not necessarily. I'm talking about the code. I know how the code works. It just "line1,line2,line3..." etc. So if someone copies it and adds a line. It would just be a few characters away from the original ("line1,linea,line2,line3" vs "line1,line2,line3").

 

Sono is gonna be mad at you.

 

What happens when they draw enough lines to bypass your 90% rule? That was my main argument in my last post. Moving away from the track would do nothing but just hide the lines from the player.

 

Moving away from the from the track wouldn't make it less of a match. It would just add it to the code.Lines are listed in the code order of creation. Thus if you if you add any lines it will just append code, leaving the original code string in its entirety, we just search for a match.

 

What happens when the copier deletes lines from the original track? Lets say, for arguments sake, that the copier deletes the first line that was placed in the original track. What then?
I'm sure you could come up with something that would run through a bunch of sections of the track to compare, but that isn't efficient at all.

 

They delete the first line, but the rest is there. Again back to the college technology. There is technology out there to see if people copied other peoples papers in their papers. We can use this sort of technology with the tracks. We don't check sections. We check the enitre code vs the entire code and see how much is different minus the added lines outside the track.

 

I see how this would work, but this strategy (along with every other possible strategy) just seems way too tedious.
The best bet for now is to just have human moderators, but we all know that won't happen any time soon.