Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Store a hash of the URLs instead of the URLs themselves.


I don't see how this would work when the URL lists are full of regexps etc.

If you could make it work, you'd would wind up with some sort of source file of domains, parsed at build time to generate list of hashes that actually ship with your plugin. That source file would belong in a repo, and would still be DMCA-able in this way.

Also, this would make ad-blockers even more CPU intensive than they are.

We shouldn't have to pursue a technical solution. It's my computer, it's my internet connection. If I'm not allowed to control my own property, then the concept of property becomes meaningless.


>I don't see how this would work when the URL lists are full of regexps etc.

I guess you could distribute a compiled finite automaton instead of a list of hashed values. It would make searching GitHub for "infringing lists" much harder (even harder than hashes).

Nevertheless, I agree with "We shouldn't have to pursue a technical solution". There's no point in trying to act like a Mr. smarty-pants in legal situations, as most technical solutions might not work out as expected in court.


>Also, this would make ad-blockers even more CPU intensive than they are.

Although they may be CPU intensive in terms of browser add-ons, I'm under the impression ad blockers are usually less CPU intensive than loading all the scripts required to display the ads, at least on particularly heavy pages. Does the CPU cost of blocker vs ads ever favor ads?


Agreed with seeking a legal solution instead of just relying on a techincal one, but if the problem is the string of characters, just hash those instead without hashing the whole host:

[HASH:1AB543.124A3CC4.1AB543]/ads


Good idea, had the same idea, but we need (the community/users need) to be able to determine what the URLs are, for transparency. So I suggest ROT-13 instead.


URLs are matched via a regex or wildcard type syntax, to eliminate files served out of /ad/ paths (yes, people still do this).

This makes hashes somewhat less useful. And a distinct hash of a URL could likely be considered to be the URL itself for the purposes of a legal action.


Regexp could still work by only hashing the domain name and keeping the regexp separate.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: