[22:35] i want to create a GUI for dumpgenerator [22:47] :-O [22:47] emijrp, isn't it already easy enough to make it explode? [22:48] people dont like command line (windowsers mainly) [22:48] * Nemo_bis worriedly looks at his toolserver instance with almost 1 GB RAM pending killing and hundreds hours to wait just for the titles list [22:48] titles what? [22:49] the titles the tool then has to download (for werelate) [22:51] why do you use toolserver? [22:51] emijrp, just for the titles, as an experiment [22:52] I didn't have free CPU in that moment and thought it would have been faster. [22:52] But it's too busy now. [22:56] i think that if we create a GUI, all the guys using wikifarms will create a backup of their own wikis [22:57] oh, now I understand, it's terrorism [22:57] we cant backup all wikis in the internet, we have to grow our users community [22:57] Actually I think that more robust tools would make it possible to backup most of them automatically with a few servers. [22:58] Just using API and improve error handling, finding a way to automate lists and metadata... [22:58] yep, but i think that is is more difficult [22:59] also, i think that our ission is to show people that making backups is good [22:59] we dont have to be the last backup service in the web [22:59] again, we cant backup all [22:59] Yes. If you don't include the upload in the GUI, though,you'll increase backups but not actual preservation. [23:00] Because people who don't know about command line won't be able to upload it to IA (or at least won't bother). [23:00] i was thinking in adding a option to upload the backup to a server, and paste the link anywhere [23:00] so, we can download it [23:01] Then you need a way to attach metadata to it. [23:01] ? [23:01] If it's just an archive you won't understand where it's from. [23:01] Not to mention license etc. [23:02] license is in index.php [23:02] which is saved [23:02] It mostly is not. [23:02] (in it) [23:02] I almost always had to dig a lot to find license. [23:03] And the API is unclear, too, because there are too many was to store license info. [23:05] what is the problem with license? those wikis without license in the mainpage can be assumed as All Rights Reserved [23:05] emijrp, no. [23:05] The license is often in a copyrights page, or in the terms of use page, or in the edittools-sort messages, etc. [23:06] anyway, i dont see the relation with creating a GUI [23:06] we dont care about copyright when archiving [23:06] Some of us do, and metadata is important. If you separate download from upload too much, then it's harder to get it. [23:07] Anyway, pretty easy to solve, just ask some info (maybe optional) in the GUI and save it in the simplest possible way. [23:08] Like domain, wiki title, main page URL and/or index.php URL, API URL, license, additional details. [23:09] are you going to trust the data? license is complex [23:09] not only, text, also images [23:09] Well, about images you can't do much besides downloading descriptions. [23:10] Asking about license is educational as well: we can archive everything, but a (public) wiki whose content is not free doesn't really work well as a wiki. [23:12] emijrp, I'm completing the Commons dump till June 2005 soon. [23:21] well