Edit: Disagree? Please comment. It's attackable via the DMCA because the word/brand is in the file in it's copyrighted form. If you encode it such that it isn't in that form, it's protected from that approach.
which is the ASCII _encoding_ of the forbidden string.
Just because these bits are an EBCDIC cp500 encoding of that string doesn't make it less forbidden than an ASCII encoding:
0b100001100b101001000b100101010b100000110b101000110b100010010b100101100b100101010b100000010b100100110b100000110b100100110b100000010b100101000b10010110b100000110b100101100b10010100
Here's how I made both.. shows reasonably (IMO) that encoding is.. nothing.
> ''.join(bin(c) for c in 'functionalclam.com'.encode('ascii'))
> # vs this:
> ''.join(bin(c) for c in 'functionalclam.com'.encode('cp500'))And reduced discoverability that your url is on my page in the first place. Even though a url isn't something you can copyright.
The DMCA complaint is patently invalid in this case. A counter notice would almost certainly eventually succeed, it's just a lot of work and potential spend in legal. Encoding it is just removing the ability to use it inappropriately in the first place.
That's why I don't consider this underhanded. As you mention, it's not copyrighted in the first place. Obscuring it is just closing the loophole that's being exploited.
I'm not talking about obscuring long passages of copyrighted text. Just obscuring one plaintext url that's on a page with others.
If you put an encrypted blob on each line instead of a plaintext url, the DMCA notice will contain "My copyrighted work appears at <link to github line number> which decrypts to <my url> when using <decryption method>." It's exactly the same as the original claim. If you feel they're not technically sophisticated enough to file a claim like this, I think we'll have to agree to disagree.
This is from experience. Scrapers that change one or two words in a long passage of text getting away with it because it doesn't match. Or very lightly cropping 1000's of original photos. I will cede that it depends on the provider. Google and many others do some rudimentary checking for sanity before allowing a DMCA request. They would never have passed this one, where just a short piece of text was claimed as infringing.
1. Can IP be encoded as a way of dodging the DMCA? a: no.
2. Will encoding the source material allow you to dodge DMCA trolls?
I think the answer is again no. The tradeoff in complexity isn't worth the marginal gains. Imagine not being able to browse the code on github, for example, since it's encoded.
A better response is to fight. If the DMCA has been abused, it's up to us to meet each claim with energy and on the same battlefield. It's tiring work, but this is the compromise America has reached circa 2017.
So the encoding isn't skirting anything.
It is just removing an easy, incorrect loophole. They are getting away with the DMCA complaint filing because most providers will allow them if you can find the "matching infringing text" on the page, without further justification.
Encoding it forces the DMCA filer to more comprehensively explain what is being violated. Basically it just raises the bar to a point where frivilous complaints can't clear it.
You can argue that's unfair, but we don't have a justice system. We have a legal system. And these are the rules.
Encoding it forces the DMCA filer to more comprehensively explain what is being violated. Basically it just raises the bar to a point where frivilous complaints can't clear it.
I think the crux of the disagreement is that I've seen how tenacious DMCA trolls are. They're smart. We like to think of them as dumb, but that's just flattery.
If you're trying to apply technical tricks to get around the legal system, you generally lose. The way to win in this case is to meet force with force on the same terms: file a counter-notification.
I think you're underestimating how valuable it is to have the content un-encoded on github. Anyone can see it. Flip through their commit logs. They have dozens of contributors that add or remove URLs almost daily. That wouldn't happen if it were encoded.
Github is a very above the table solution for distributing a list of things. But if it gets to expensive to host it there, due to frivolous lawsuit threats, I expect other solutions to distributing these lists start to pop up.
The best advantage of encoding or even encrypting this stuff would be obfuscation. It would make is much more difficult for someone to even become AWARE that they are being blocked, if the info is hidden.
Not everybody has the resources to fight off trolls, even if what they are doing is completely legal (as in this case). Sometimes technical solutions are your only option.
Regarding your point about obfuscation, I feel like people aren't thinking this through. It's on Github. The commit messages are public. Finding whether you're blocked is as simple as grepping the logs for your domain. And if you feel the people filing the DMCAs aren't technically capable, you may be underestimating them.
The advantage of having it on github far outweighs the negatives. How many contributors do you feel they gained from this publicity? Possibly at least one, which is one more than EasyList would've had otherwise.
Obscuring it makes the DMCA filer have to better explain the why in the request. Something more than just "this short piece text is on this page and I don't like it".
This is false if the IP being enforced is trademark (which I'm not even sure is covered by the DMCA).