On Wednesday 28 September 2005 23:55, Christian Biere wrote: > Well it's not as evil as in my bogus example. However, if you change > to filenames to something meaningless, people won't be able to > find the file at your site. ... > But as Gnutella user you and your shared files are part of a global > network. Its sole purpose is to share files world-wide. ... > If I download some music file with Gnutella, it ends up in my > download directory and "mplayer -shuffle *" takes care off my > needs.
That shows that you treat your files differently than I do. I understand that you are concerned about correct "round-tripping" of downloaded files, so that they, from a network perspective, don't vanish into a black hole in the namespace after being downloaded. But in my case, completely downloaded files won't be automatically round-tripped anyway. In this sense, I treat partial file sharing and sharing of local files completely different. In case of partial file sharing, you are right. There must not be a black hole. Still-downloading files should really be re-shared automatically, that's part of the protocol and is the only way to keep the network sane. But partial files are requested by SHA-1 anyway, so this will not be affected. But once a file is completely downloaded, it becomes a local file, and it should be my decision if or how I want to share it. In practice, after a file is completely downloaded, I first review its content and delete it if I don't consider it worth keeping. Then I give it a new meaningful filename according to my policy. Then I edit the meta tags of the file to give most valuable information, so the SHA-1 of the file changes. Then I move the file to my data folders. And then, if I consider the file to be interesting and high-quality and shareable, I perhaps re-share it. I currently share about 1000 high-quality, interesting, hand-picked files that surely enrich the net and are worth the bandwidth. But most of them have nothing to do with the files I download. They come from completely different sources. All that being said, it should be clear that there is simply no automatic round-tripping going on anyway. Now back to our initial subject. If the name of a downloaded file contains characters that don't make sense in my on-disk charset, I of course change or delete these characters manually. The important information is stored in the meta tags of the file anyway, which are, at least in the case of ogg files, locale-independent, i.e. defined to be always utf-8. So I rename "À" to "ä", which should be done by GTKG automatically without losing information. And I delete "æ¥æ¬èª", even if I know that it is probably supposed to mean "日本語", which can't be represented in my charset. I don't want to see "æ¥æ¬èª" in my file browser. So I won't return query hits on searches for "日本語" anyway. Perhaps I will name it "nihongo" or even "japanese" instead, which are both just as reasonable choices. Now, in the light of how I handle my files, you see that automatically converting names to my locale, even if some underscores are needed, is just not evil if I choose so. The "no black holes" policy only applies to partial file sharing, or maybe also to freshly completed downloaded files (whose original name is still internally known by GTKG), but nothing more. Hauke ------------------------------------------------------- This SF.Net email is sponsored by: Power Architecture Resource Center: Free content, downloads, discussions, and more. http://solutions.newsforge.com/ibmarch.tmpl _______________________________________________ Gtk-gnutella-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/gtk-gnutella-devel
