[03:00] https://twitter.com/#!/lewiscollard/status/185916679757365248 [03:27] http://i.imgur.com/RxyBm.png [03:28] bonus points for guessing where I started downloading mobileme [03:28] :D [03:36] probably someplace you shouldn be? [03:36] I meant where in time [03:37] especially when it quadruples usage [03:37] but that might be true too [04:40] lets do this [04:41] Welcome [04:41] how many mobileme instances would be ideal? [04:42] 42000 [04:42] haha [04:45] i'll do 50 for now [04:45] alard: where's that script again? [04:45] i think the app i have laying around is out of date [04:49] " Didn't get a client id. Sleeping for 60 seconds…" :( [05:05] http://techcrunch.com/2012/03/30/jordan-mechner-creator-of-prince-of-persia-finds-original-source-code-in-his-dads-closet/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+Techcrunch+%28TechCrunch%29 [05:23] Oh, that thing I'm flying to LA to work on? [05:52] SketchCow: you're actually flying out there to help? [05:52] That's cool [06:53] OT but great: http://serverfault.com/questions/293217/our-security-auditor-is-an-idiot-how-do-i-give-him-the-information-he-wants [07:06] (That link) WIN! [12:18] ShoutWiki wikifarm is probably dead http://shoutwiki.com/ (no response in the last month). #WikiTeam (nemo_bis) took a copy in August 2011 http://archive.org/details/shoutwiki.com. [12:25] * BlueMax salutes the Wiki Team [12:38] Knol, one month to the shutdown. [13:37] i got scientific tuesdays from rev3 [13:37] hd ones too [13:37] thats about 3.1gb [13:42] http://hackaday.com/2012/03/31/floppy-autoloader-takes-the-pain-out-of-archiving-5000-amiga-disks/ [13:42] ^_^ [14:40] So, I have various music stuff i was involved with when I was a teenager. Rock 'n roll type stuff mostly. I was thinking of dumping it into IA. I'm kind of confused where these should go? I'm not sure if it's really considered a "netlabel" [14:50] human39: Technically, it's "open source music" [14:50] Is it a lot of music? [14:50] Or, I should say, items? With an item being an album of music. [15:16] SketchCow: How is it best to deal with recorded music that's probably still within copyright but extremely obscure to a point where the original authors/producers no longer exist and that won't get reissued probably ever? [15:16] Copy it, upload it [15:17] Is this the case with scores too? [15:18] My idea was a nice system using mediawiki to keep everything organized [15:18] But ofc that's more work [15:36] If you make it public, you're going to get torn a new asshole probably in due time [15:36] As in a web site or so [15:38] balrog_ph: http://en.wikipedia.org/wiki/Orphan_works it's a fun fun fun fun legal agea [15:38] area [15:38] tef: I know [15:39] In this case it's obscure enough that I don't think it's THAT dangerous [15:41] balrog_ph: label them as such, make sure you have contact information for rights holders to contact you and request a takedown [15:41] should be enough to cover your butt but IANAL [15:41] Yeah I know [15:41] cool :-) [16:28] kennethre: Hooray, looks like Readibility is not going to jail [16:28] SketchCow: hah, new news? [16:29] Shame it took an asshole like Gruber to fix that. [16:29] Yeah, the highjacking of links made gruber-pushed news (Benjamin Brooks, too) [16:29] So they switched things [16:29] yeah i was all over that yesterday [16:29] Naturally, people are complaining because it broke a ton of things. [16:29] he railed into them [16:30] My thoughts were... [16:30] http://utf-8.jp/public/aaencode.html *head explode* [16:30] "Readability isn't run by scumbags, but I do sincerely believe that the concerns people are raising people are very real." [16:30] "They're fantastic guys with great intentions, but they're moving too fast. Everything's a reaction. The blog post today is a testament that." [16:30] "You have to take time to do things properly. To reflect on what you're building and why. It's why I left the company." [16:30] "Transparency." [16:31] Oh man, I hadn't seen that Anil Dash has his nutsack in all this. [16:31] i think he's typically buddy buddy with them [16:32] He's called a "Readability advisor" [16:32] ah yeah [16:32] so glad i left :) [16:32] This was a big year for me - three times people I have fucking roasted publically didn't even TRY to engage me at events. [16:32] Anil was one of them. Saw me, walked right past [16:32] \o/ [16:33] Hahahaha [16:33] you're too quick with your tongue to engage [16:34] (in a good way) [16:34] Well, beating people down verbally is a specialty [16:35] Don't worry, the kind of people who hand-wring themselves to death over a long pause have taken time to tell me I catch more flies with honey [16:35] I point out you catch even more with a flamethrower. [16:35] hahah [16:36] readability is almost purely in the business of having press [16:37] we almost actually waited several hours to push a release to production so someons' buddy at CNN could run their story on it [16:37] i suspect they won't recover well from yesterday, but we'll see I guess [16:39] Sorry we stol ur moneys [16:40] Fact. If you hold money 'for' someone, you hold it FOR YEARS [16:40] And then LATER, you DETERMINE THE FUTURE OF IT [16:40] You don't decide, oh, 365 days? That's a random number we decided now represents free monies we get to keep [16:41] You're already getting interest at the bank [16:41] Which can be significant [16:41] And best yet, you say so. [16:41] I feel like the whole fiasco could have been avoided by a simple bulleted list [16:42] "We do X with unclaimed money." [16:44] "When you share a link, this hapens." [16:44] My new favorite phrase when talking to someone is "Oh, make no mistake, you're going to go to jail" [16:44] Amazing how that changes the tenor of the conversation [16:44] how often can you fit that into conversation? [16:44] Anyway, yeah, I was delighted that people have figured out not to engage. [16:45] I talk to a lot of people who are going to jail [16:45] lmao [16:45] Seems I missed something interesting... [16:45] So, Amazon S3 for free guy is running into static/friction [16:45] So we'll see if that changes [16:45] What's going on? [16:46] Free S3 stuff [16:46] Not going so well? :/ [16:46] Well, it's not a "going" thing [16:46] He's going to ask and get an answer [16:46] I see [16:46] I wouldn't expect anything...unfortunately [16:47] SketchCow: link? [16:54] http://cl.ly/2I471r433P3q082r0z12 [16:54] sigh [17:43] and that ShoutWiki copy is up-to-date, because the wikis were put in read only... [18:59] SketchCow, starting out, one comp, so one item for now. I have a number of others, but I don't think they amount to more than a couple of handfuls of "items". [19:11] Woot, what is this talk about Readability? [20:10] kennethre: when you're around, I'd like to talk to you re: httplib.BadStatusLine and robust ways to debug the problem [20:10] I figure you're the biggest Python person here [20:16] human39: "community audio" is kind of the grab bag [20:17] ersi: http://daringfireball.net/linked/2012/03/30/readability [20:17] yipdw: honestly that's something i don't run into very often. what's causing it now? [20:21] kennethre: I'm really not sure; I've just managed to track the exception as resulting from here -> https://github.com/ArchiveTeam/ffnet-grab/blob/client/retrieve.py#L165-167 [20:22] it could be the tracker [20:22] yipdw: http://python-requests.org :) [20:22] yipdw: takes a lot of that stuff into account [20:22] yeah :P [20:22] I saw that after I wrote most of retrieve.py [20:22] the API seems, uh [20:22] less overengineered [20:22] should be pretty easy to switch [20:22] haha yes :) [20:22] I think I will [20:22] (that's my library) [20:22] yep :P [20:23] so quick answer [20:23] make sure you're sending a host header [20:23] aaaaactually [20:23] it could very well be the tracker [20:23] I'm running with a prerelease version of Rubinius and Puma [20:24] i don't often use non-released code in production, but when I do I do it throughout my whole fucking stack [20:24] might as well switch the client to requests, if only for fun [20:24] i encourage this :) [20:26] yipdw: It's also good to add a "Content-Type: application/json" header. [20:26] to the request? [20:26] oh hmm [20:26] I am indeed not adding that, oops [20:26] kennethre: oh I found a 'bug' in requests but I fixed it by not being stupid [20:26] tef: what was it? [20:27] you can make a GET request with a body attached [20:27] which is invalid, but many servers won't complain and drop the body [20:27] iirc on the invalid part [20:28] could be wrong mind, but tbh I was expecting an error locally saying 'no body allowed on get requests' or silently dropping the data part [20:28] tef: it's actually not invalid, but the default api doesn't provide a 'data' argument [20:28] tef: but if you use requests.request('GET', …) directly it'll let you [20:28] ah yes I am doing the latter [20:29] yeah that api makes no assumptions [20:29] like I said, I fixed 'the bug' by not being stupid [20:29] you can send a ARCHIVE request if you want :) [20:30] I like that you can do that via XMLHttpRequest [20:30] at least in Gecko, haven't tried it elsewhere [20:30] one day I will write a webapp that uses profanities for methods [20:48] guys, can I have your opinion on something? -- the current draft of the DiscFerret file format: www.discferret.com/DFIFileFormat_20120331.pdf [20:48] I'm basically looking for suggestions on what it should store, and how. [21:09] 1.1M tweets of superbowl http://r-shief.org/dan/superbowl_twitter.zip [21:26] philpem: seems good so far [21:27] philpem: I've got a book called "magnetic recording" by Lowman; suggest you buy a copy if you're going to be doing much work with tape or disk [21:27] philpem: is DFI also going to be usable for tape archiving? [21:28] I wasn't intending it to be used for tapes, but it could be extended to do so [21:28] you'd end up splitting things into "blocks" rather than "tracks", "heads" and "sectors" though [21:29] tapes have tracks and heads and blocks, which is kind of like tracks and sides and sectors. [21:29] hm, ok, then in that case you could store it -- and add a metadata block to identify it as being a tape image [21:29] anyway, if you allow for more than 2 sides, the format becomes more widely useful with little change needed [21:29] yes [21:29] "sides" [21:30] it already does allow for more than two sides -- the identifiers are 32-bit. [21:30] how do you plan to encode the data? [21:30] ah, excellent! [21:30] flux transition timing, as seems to be the best current method? [21:30] exactly [21:30] hmmm [21:31] it stores two blocks of data: the Timing Delta Block (which contains the transition timings) and the Index Offset Block (which stores the location of any index pulses which were detected) [21:31] kennethre: converted that script over to requests, seems to be going fine so far [21:31] TDB is compressed, IOB is not [21:31] aye [21:31] yipdw; :D [21:31] well, let me rephrase that, TDB "can be" compressed. [21:31] I like how it eliminated code in a 4:1 ratio [21:31] or more [21:32] generally bzip2 but lzma, lzo and gzip also have compression format IDs [21:33] philpem: right. [21:33] bzip2 and gzip are mandatory, lzma and lzo are optional. [21:33] how is this format different from the DRAFT system being designed by the Kryoflux people? [21:33] DRAFT doesn't exist [21:33] it was announced a long time ago [21:33] mine actually has publicly available documentation :) [21:33] not at all? [21:33] their current STREAM is a joke [21:33] I haven't seen it [21:33] not seen it == vaporware/nonexistent [21:33] well, balrog_, yes STREAM is not an archival format [21:33] ok [21:34] well then! [21:34] fuckem [21:34] * philpem grins [21:34] yipdw: that's the goal :) [21:34] I mean seriously, how long does it take? :/ [21:34] chronomex^ [21:35] balrog_: to publish a file format? nfi. [21:35] to design and implement and publish [21:35] one file per track is clunky [21:35] :< [21:35] kennethre: though, I am bit uncomfortable with the fix, because I don't really understand why I was getting that BadStatusLine exception in the first place [21:35] depends on the problem you're solving! [21:35] the tracker logs show no errors server-side [21:35] so I guess it's a client thing, but hell if I know why [21:35] yipdw: probally missing `host` or `accept` headers [21:35] hm [21:36] requests has taken a lot into consideration :) [21:36] weird exception to raise :P [21:36] chronomex: an unpacker solves that [21:36] I mean, I'd expect something like MissingAcceptHeader or whatnot [21:38] philpem: I would enjoy working on this with you. [21:38] yipdw: nah, different servers do drastically different things [21:39] chronomex: freenode, #discferret :) [21:39] freenode? :| [21:39] yes, freenode, what's wrong with it? [21:39] it's not efnet [21:43] heh [21:48] efnet doesn't have chanserv, which means miscreants can take over too easily :/ [21:49] efnet is my home [21:50] buy raspberrypi and have xchat running 24/7 [21:50] wrong solution. use shellserver, irssi. [22:23] hmm [22:23] 87 examples, 0 failures, 7 pending [22:23] Finished in -98664708.0478 seconds [22:23] I think something may be wrong [22:23] maybe [22:23] nah, looks good [22:26] yeah, my superluminal computer is awesome [23:31] I've got a tarball that seesaw-s3 is having trouble uploading-- curl's getting 500 errors from batcave and it's down to 33 retries. Known issue? [23:52] Completely unrelated to the above, but I'm curious. SketchCow, how resilient are Apple II floppies in terms of data recovery? [23:55] The Prince of Persia thing is amazingly cool, but at the same time, I'm given the impression that twenty-five years is a long time for magnet-coated plastic. [23:55] it is [23:56] (Though isn't tape pretty much the same thing? What makes that more-suitable for archival?) [23:57] nothing, it just historically has a cost advantage [23:57] Aah [23:57] best format for long term storage is holes punched in tyvek tape or cards