youtube-dl and youtube through free software [Was: Fwd: Re: praising our moderators]
Alexandre François Garreau <[email protected]>
| Newsgroups | gmane.org.gnu.discuss |
|---|---|
| Message-ID | <[email protected]> |
Hi, I received this message from g-m-d, and I’d like to know your opinion about it. First of all (as last suggestion) he suggests using free-JS “frontend” to youtube, which likely can only at best themselves work by running the less non-free software for downloading videos from youtube… isn’t that as wrong as using it directly? I mean, it has even the added downside of looking like SaaSS and requiring an added intermediary and being a centralized solution… what do you think about it? Also about youtube-dl apparently it’s like I thought and it only executes a branchless (hence turing-incomplete) subset of javascript to compute a signature for the video to be downloaded… isn’t that the Right Thing? how is that different from automated reverse engineering? Because: the work-load-heavier alternative would be to look at the extracted function alone, and rewrite it from that small subset of js into python by hand (what youtube-dl apparently already does, but in an automated manner)… but this subset is automatically generated, and has no meaning… it’s like a programming-language-based encryption… so isn’t that like doing manually what is done automatically? since this is automatically generated, and has no capabilities (other than hiding information), doesn’t it have “no meaning” as a program? doesn’t this sort of sub-“evaluation” is something akin to “decrypting”?
message suivi
(message/rfc822, 5 KB) - not displayed