[01:50] I can't believe there's a band that put out their music as 01.rar [01:50] and each of the tracks is 01.mp3, 02.mp3, etc. [01:51] Luckily they at least had the foresight to use id3 tags. [01:54] heh nice [01:54] how friendly of them [01:54] Yeah [01:55] First chance I get, that's coming off my drive [01:55] Also noting that 640GB is likely not going to be enough for /media/music [01:55] I'm already at 13%, and haven't really started ripping discs. [01:55] /dev/sdb1 615319352 70835964 513226924 13% /mnt/music [01:56] yay /mnt/! ;) [01:56] sorry, had to [01:56] Yeah, delete me once, shame on me [01:56] delete me twice, I find you and play maracas with your nuts. [01:57] delete you twice, never delete you no more? [01:57] Delete me once. Delete me twice. I'll undelete at any price [01:57] heh [01:57] I need this data like i needed it then [01:58] I'll alway say "backed up again... someday" === widox_ is now known as widox [11:49] g'morning [11:58] yessir [12:29] says who? [12:39] says me, party party [12:41] bah [12:41] they locked our thermostat and set the temp to "uncomfortably high". It's definitely not a good morning [12:42] ugh, I get that a lot here to [12:42] I have a themometer I show to my boss [12:42] "see, it's 80...now 80 is beach weather. I should just head to the beach" [12:42] Yeah, I threatened to turn the steel-toes in for sandals, and wear shorts [12:43] but of course nobody that I can complain to has any control over it... [12:43] the union guys are saying "work slower. It's too hot, and we don't want anybody getting a heart attack" [12:44] lol, awesome [12:44] The sad thing is, it's just the offices. The plant floor is still nice and cool. [12:44] so I may go for a walk with my laptop and not come back [12:58] heh [12:59] GOod morning, internet folks [13:02] Wolfger: it took me 3 weeks of 80+ degrees in this office and complaining to various facilities/management/etc people to get our temperature fixed. Apparently, there is one thermostat on the 2nd floor (I'm on the 4th) that controls this entire wing [13:11] greg-g: gotta love that [13:11] our therm is in a conference room down the hall [13:12] well of course that small closed room is a totally different climate than the big cube farm we're in [13:14] no kidding. We now have 5 desktops and 8 22" monitors in here, it can get warm if you let it :) [13:15] but don't you wear sandals, shorts, and no shirt to work? [13:20] I wish [13:20] for a while I was wearing sandals, but now my boss is back in town [13:22] darned bosses [13:22] no kidding, right? [13:27] In this Flask app that I'm writing I use a list to keep some data that changes regularly. I that list the same across threads if i deploy to mod_wsgi? https://sourceforge.net/p/waznexserver/code/ci/4028742f72bea7570941752ee42beaaec721fb47/tree/main.py [13:31] brousch: hey hey hey, no talking about real work when we're bitching about the weather. [13:31] well technically it's a side project, so it's fun [13:31] Heh [13:31] Also see that Dave got you using the new beta. [13:31] yeah, i like it [13:32] Yeah, it's quite nice. Lots of good ideas in there. [13:32] i wish SF had private repos so i could host non OSS stuff there [13:32] Not likely to happen any time soon [13:32] that's what he said too [13:33] i should go to a database backend instead of a list and eliminate any worry [13:34] That would be my first recommendation [13:37] brousch: so when you deploy mod_wsgi, apache spins up several workers of your app [13:37] i think so [13:37] now, any requests that go the same worker that would be constant [13:37] but not across the apache workers [13:37] each of those is separate [13:37] damn [13:37] as apache spins up more instances to hanlde load, you'd end up with a bunch of different versions of that list [13:38] you need a persistant store, sqlite, memcache, etc [13:38] single points like that don't let you scale beyond a single process really [13:38] unless you want to get into the joys of interprocess communication [13:39] no. i want to keep it simple [13:39] I seem to recall that the scoreboard might be useful for something like that [13:39] sqlite it is then [13:39] but I may be remembering wrong [13:39] the scoreboard? [13:39] well, except that it doens't like multiple users at once without locking issues :/ [13:39] Bah, install PostgreSQL [13:39] be a man [13:39] dreamhost frowns on that [13:39] heh [13:39] couchdb then? :) [13:40] obviously i need mongo for this [13:40] now you have two problems [13:40] now you have three problems [13:40] (mongo counts for two problems) [13:40] hey, I like couch [13:40] I like couch too [13:40] * Wolfger likes couches [13:41] time to dive into sqlalchemy [13:41] Also loving the news that Dropbox unlocked everyone's accounts? [13:41] now would be a good time for a tutoria, if only there was one coming up :) [13:41] * rick_h_ goes back to work...too depressing to think about the world's security [13:41] http://hardware.slashdot.org/story/11/06/21/045228/Dropbox-Password-Goof-Let-Any-Password-Work-For-4-Hours [13:41] rick_h_: Sorry, man [13:42] our podcasts might have been compromised [13:42] * snap-l doesn't keep anything in there, really [13:43] all my purchased books [13:43] scanned day care receipts [13:43] you mean my purchased books! [13:43] meh, could be worse I guess [13:43] "Oh shit, they might access my coupons and my FUDGE books that I'm reading on the iPhone" [13:43] And our logo / podcasts. [13:43] i'm afraid to ask what a FUDGE book is [13:44] http://fudgerpg.com/ [13:44] Unfortunately their site is a hot mess, but it's a good system. [13:44] http://fudgerpg.com/goodies/fudge-files.html [13:45] Developed on the Internet back when folks used Usenet [13:58] Folks don't use Usenet anymore? [14:02] i use the net every day [14:05] har har [14:05] Not like they used to use Usenet [14:05] it was the Internet's biggest forum [14:05] I really miss Usenet some days [14:06] then I look at other message boards and think "I'm glad I'm not paying for the privilege of having this piped directly into my house" [14:07] You get free broadband? :-o [14:07] or do you mean s/not paying/not paying extra/ [14:15] God, Windows with Bluetooth is shit [14:15] Paying Extra, I mean [14:19] ubuntuone needs to jump on the dropbox security woes [14:20] And get their own? :) [14:22] ouch [14:33] greg-g: what building are you in that has such poor env controls? [14:35] jrwren: It's a library building. :) [14:35] Monitored by the finest in library sciences gestappo agents [14:35] you realize there are over 100 libraries at umich? [14:36] jrwren: You realize there's no such thing as library sciences gestappo agents [14:36] that's at least 98 too many [14:37] snap-l: how about a library sciences gazpacho? === smoser` is now known as smoser [14:39] jrwren: shapiro library [14:42] finding the Bitcoin story fascinating [14:42] Possibly of interest: Free (CC-BY) icons. http://www.fatcow.com/free-icons [14:42] nice! [14:43] Also a free font [14:48] Thanks, snap-l. Now I have that song stuck in my head. [14:51] greg-g: you work in the UGLi? [14:53] brousch: I work on the 4th floor of it, yep :) [14:53] getting all specific now... [14:54] Are you facing the Atlantic or the Pacific [14:54] ? [14:55] pacific [14:56] what are your GPS coordinates? [14:59] <_stink_> a/s/l? [14:59] oooh shiny: https://github.com/Pita/etherpad-lite/tree/182477af93eaab667a5c7db03888788ea0f47434 [15:03] Two months ago we migrated from MD5 hashing to freeBSD MD5 salted hashing.?The unsalted user accounts in the wild are ones that haven't been accessed in over 2 months and are considered idle. Once we are back up we will have implemented SHA-512 multi-iteration salted hashing and all users will be required to update to a new strong password. [15:04] seriously, a site that is based on cryptographic strong hashes for currency was using MD5 salted passwords? [15:14] who? [15:14] bitcoin stuff? [15:19] Yeah [15:19] ok, pythonistas, why would one do "newlist = mylist[:]" instead of just "newlist = mylist" ? [15:19] newlist = mylist would copy the pointer, so they'd be the same [15:19] newlist = mylist[:] would copy the data [15:20] <_stink_> shit, i didn't know that trick. [15:20] * snap-l verifies this [15:20] <_stink_> i would just use deepcopy or something. [15:21] In [1]: mylist = [1,2,3] [15:21] In [2]: mylist [15:21] Out[2]: [1, 2, 3] [15:21] In [3]: newlist = mylist [15:21] In [5]: newlist.append(4) [15:21] In [6]: mylist [15:21] Out[6]: [1, 2, 3, 4] [15:21] wow, thanks! [15:22] wouldn't [:] copy the pointers to any objects in the list? [15:22] In [10]: mylist = [1,2,3] [15:22] In [11]: newlist = mylist[:] [15:22] In [12]: newlist.append(4) [15:22] In [13]: mylist [15:22] Out[13]: [1, 2, 3] [15:22] so the objects in the new list would be same as in the old? [15:22] <_stink_> i think [:] would just copy the first level. [15:22] yeah, that's what i was asking [15:22] <_stink_> like if the list held dicts, the new list would hold the same dict pointers [15:22] _stink_: ++ [15:23] so deepcopy still has a purpose [15:23] <_stink_> right [15:23] (this was a question for someone who works at CC and saw it in some code but didn't know exactly what it was. He's the sysadmin) [15:23] <_stink_> hey, learned something today. and it's only 10:30. [15:24] _stink_: time to go home, then [15:24] <_stink_> damn right [15:55] I'm not sure you'd want to deepcopy by default [15:55] but yea, [:] is a new list from a slide [15:55] /slide/slice [15:59] http://scoopertino.com/exposed-the-imac-disaster-that-almost-was/ [15:59] The letter from Sean Connery to Steve Jobs is priceless. [16:01] https://twitter.com/#!/Pinboard/status/83187358165499904 ouch [16:02] Wow [16:03] Man, FF5 already [16:03] looks like this coming back to bite him: https://twitter.com/#!/GDgeek/status/72671736897875970 [16:04] Apparently Firefox got Chrome version number envy [16:04] https://twitter.com/#!/DavidAndGoliath/status/72736927589208065 [16:04] lol [16:04] and I look forward to FF2112 [16:04] yea, FF on the frazy [16:04] Should drop any day now [16:04] I got an email from an automated check for the Bookie FF extension on FF6 already [16:05] That DavidAndGoliath post doesn't make any damn sense [16:05] wtf. they were hosting 3 TB of data for $35/mo? [16:06] yes [16:06] a while back he was trying to find the absolute cheapest dedicated server hosting he could [16:06] do people pay for pinboard? [16:06] which I thought odd since he tries to keep a small number of boxes [16:06] yes [16:06] Either they got the sweetest deal on the planet, or someone is hosting from their garage [16:06] I did [16:06] i suggest a heapin pile of flaming poo bags on his porch [16:06] well that's where that comment is from, people warning him that at that price, dont' expect a lot [16:07] I pay $24.95 for Linode, and I expect it to be there, damnit. :) [16:07] and now he's been down for 9 hours and trying to setup a new db server elsewhere to get back up [16:07] once people start actually paying you for your service you should really upgrade your hosting [16:08] Or have a way to spin up S3 hosting quickly [16:08] he did an interview and prides himself on his 3 boxes repliating around them to each other [16:08] /s3/ec2 [16:08] it [16:08] Sorry [16:08] s the db server that's the thing [16:08] you can't just spin up db servers with TB of disk [16:08] anyone tried http://sparkleshare.org/ ? [16:08] Though I wonder how one gets 3TB of bookmarks [16:08] krondor: Not yet. Looks interesting [16:09] maybe it indexes like bookie? [16:09] he does have a 'storage' account you get for bookie like caching of content [16:09] but I've not figured out if it's just screenshots, or what [16:09] Still, that's archive.org sizes [16:11] https://twitter.com/#!/curbednetwork/status/83171089429168128 ?! [16:11] oh fuck [16:12] I hope he spent some of that internet cash for a PO box [16:12] because I'd hate for people to show up at his doorstep [16:13] https://twitter.com/#!/Pinboard/status/83050440689844224 [16:13] so looks like worst case is a new install brought up from s3 backup [16:13] which might take a while based on size [16:14] Someone needs to tell him about gzip. :) [16:15] http://thuktun.wordpress.com/2011/01/03/an-interview-with-maciej-ceglowski-of-pinboard/ [16:16] down there has a server setup description [16:20] "Too much of the site now runs in screen;" lol [16:27] I say that boring hosting is a feature because in times of crisis you find yourself in a very comfortable and familiar setup. It’s also easy to bring people in and explain what you’re doing. And you get the benefit of a lot of documented online experience with similar tools, by people smarter than you are who write about it well. [16:28] rick_h_: A Liberal Arts major that knows how to use screen? My God [16:30] at least he used screen [16:30] instead of wondering where the job went when he logged off. ;) [16:30] Milyardo: :P [16:31] Indeed, I'm actually Impressed [16:31] What's funny is how many services we use that we think are bulletproof turn out to be scotch-taped together. [16:32] _stink_: jcastro snap-l you guys know of any local college I could look at taking something on machine learning? stuff like this: http://www.flickr.com/photos/azaraskin/5765482763/sizes/o/in/photostream/ ? [16:32] not seeing anything in oaklands course work [16:33] man, no clue [16:33] actually very afraid to see cobol and visual basic [16:33] rick_h_: Shit, that has math on it [16:33] yea, it does :/ [16:33] Oakland wouldn't have that. ;P [16:33] I used to be good at it [16:33] wondering if I could cobble enough together to grasp it [16:33] rick_h_: University of Michigan [16:33] pinboard is $10??? WTF? [16:33] <_stink_> rick_h_: there are faculty in Computer Engineering that do this stuff, but i'll have to look more about courses [16:34] <_stink_> at WSU [16:34] <_stink_> i mean [16:34] Wayne might have something like that [16:34] _stink_: ok, wondering if one school is known more for higher tech stuff [16:34] Hope College would have that [16:34] <_stink_> hmm, not sure WSU is 'known' for anything [16:34] <_stink_> bums? [16:35] WSU is known for being the university for people who don't mind parking tickets and random shit getting stolen from their car [16:35] Oakland is for people who don't mind not learning shit (at least in the Engineering department) [16:35] EE department, rather [16:35] ok, well oakland is closest so started there [16:35] but yea, turning up empty [16:36] I would guess U-M [16:36] jwren, that $10 is totally worth it. [16:36] yea, that's a hike [16:36] I have my Theory of Computing book if you want to borrow it, rick_h_ [16:36] snap-l: yea, I'm debating. Learning on my own would take a lot of time I think because I'm rusty on the math side [16:37] wondering if a class would help make it faster worth the $$ [16:37] snap-l: my future BIL is a cop for WSU :P [16:37] <_stink_> rick_h_: http://classschedule.wayne.edu/sections_new.cfm?subj=CSC&course=5800&campus=NOSELECTION&instr=NOSELECTION [16:37] rick_h_: look at OU's expert systems course. [16:37] geeze, crazy to think my stats class was that long ago [16:37] or AI. [16:37] Machine Learning is just an AI specialization [16:37] * _stink_ shrugs [16:37] brousch: At least he'll have his work cut out for him. :) [16:37] <_stink_> rick_h_: and other 5000-level courses on that site [16:37] <_stink_> fwiw [16:37] _stink_: ooh, thanks [16:38] <_stink_> np [16:38] that looks like a definite good start [16:39] http://www.amazon.com/Theory-Computation-Languages-Automata-Complexity/dp/0805301437 [16:39] rick_h_: ^^ [16:39] Mine is so old, it's a hardcover. ;) [16:40] My Father had that book, started reading it once, never finished it. [16:41] Shit, i should re-read this book [16:41] http://www2.oakland.edu/gradstudy/pages_cfm/course_detail.cfm?cterm=201130&subj=CSE&numb=513 [16:41] Sits on my shelf like the Knuth books [16:41] jrwren: cool, thanks [16:42] http://www2.oakland.edu/gradstudy/pages_cfm/course_detail.cfm?cterm=201130&subj=CSE&numb=581 [16:42] I must have been in the undergrad class list [16:42] snap-l: is brookshear any better than others/ [16:42] jrwren: It's what we used in my upper-level Comp-Sci classes [16:42] snap-l: why are you lking to comp theory for machine learning. [16:42] VERY diffrent fields. [16:43] snap-l: yeah, we used a different text, but it is not the cononical. [16:43] I think Chomsky's work is the more canonical text. [16:43] rick_h_: this one seems MOSt relevant: http://www2.oakland.edu/gradstudy/pages_cfm/course_detail.cfm?cterm=201130&subj=CSE&numb=616 [16:44] the whole list is here: [16:44] http://www2.oakland.edu/gradstudy/pages_cfm/catalog3.cfm?cterm=201130&subj=CSE&action=printable [16:44] jrwren: Was linking to Computational Theory as part of the building blocks for Machine Learning [16:44] oh, well, I don't know that it is really. [16:44] i mean, it is in as much as it is the foundation for all comptuer science [16:44] but beyond that, I dont' know that it does. [16:45] machine learning seems to be built on basic algorithms. [16:45] and stats [16:45] like Bayes and Markov [16:45] right. [16:46] rick_h_: don't forget about http://itunes.stanford.edu/ [16:46] Heck, even MIT Courseware [16:46] right. [16:47] http://see.stanford.edu/see/courseinfo.aspx?coll=348ca38a-3a6d-4052-937d-cb017338d7b1 [16:47] yea, I like the look of CSE 581 as a starter [16:47] http://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-867-machine-learning-fall-2006/ [16:47] only pre-req is db which I can handle no problem [16:48] http://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-867-machine-learning-fall-2006/ [16:48] but the neural net/etc is stuff I'd want to get to [16:48] oh, lol. [16:48] jrwren: Beat you. :) [16:48] snap-l: we must have googled the same stuff. [16:48] nice [16:48] snap-l: JYNX! U OWE ME A COKE! [16:48] heh [16:50] grumble. their class search for the schedule has no CSE section at all, just CIS [16:51] who? OU? [16:51] yea, all the drop downs of decipline have CIS vs cse [16:51] are you in business school instaed of engineering school or something? [16:52] no, I think the main site only looks at undergrad [16:52] could be [16:53] yea, got it, can find the listings now [16:53] just checking if there's an offering schedule/etc [16:53] if you think you need the focus of a classroom instead of online coursewear... we could organize something here and with CHC. [16:53] cool thanks guys, lots of good links [16:53] ya know, take the online thing together and tehn meet at CHC to discuss and review that weeks course. [16:53] and be on teh same schedule [16:53] jrwren: I'd love that [16:53] i've done that with book reading group before and it was very successful IMO [16:53] hah, forget a book club [16:53] "machine learning club" [16:54] It's been a while since I've worked on any of the CompSci stuff [16:54] right [16:54] online coursewear club [16:54] cool, will have to check it out [16:54] I've just had a bug recently about some of the big data type learning/recommendation stuff [16:54] rick_h_: You are hitting the wall. :) [16:54] and never got that high in school things, but it's a bit out of my casual reading/self learning branch [16:54] big data type? [16:54] and are looking for something to climb [16:55] loved the pycon talk from the bit.ly tech [16:55] big data [16:55] all the social networks and their big data [16:55] hopefully bookie and some larger datasets [16:57] oh, well, I think big data is yet another field :) [16:57] I got high in school things. [16:57] big data? [16:57] see my masters degree? [16:58] large data [16:58] I tried to get a master, I just can't stand school that much. I'm a practical person I think [16:58] right. [16:58] you are too smart. [16:58] the only reason this is coming up is that I want to try out doing practical stuff with it [16:58] school is too slow for rick_h_ [16:58] i'm dumb enough taht I didn't figure out how dumb school was until I got far into my masters. [16:58] but a few classes from those that know more than me on a subject I'm interested in might work [16:59] vs a whole set of courseware [16:59] mapreduce is really the #1 answer to large data. [16:59] if you can't solve it with map reduce, I think you are immediately back into research land that hasn't been practically proven. [16:59] yea, but there's the math on relating items [16:59] you need to know what to tell map reduce to do to get meaningful stuff out [17:00] right. [17:00] that is the art part of the computer science [17:01] you are basically saying "how do I knwo what map and reduce function to use" [17:01] and sometimes that reduction part can be VERY challenging [17:01] so my immediate practical thing is with bookmarks [17:01] for bookie? [17:01] "you're viewing this bookmark, and here are 5 like it via tags/content [17:01] right [17:01] but also at work, we've got a lot of stats data that might be interesting [17:02] "people that like car XXX also like car yyy" [17:02] the "like it" sounds more like clustering. [17:02] I'm not 100% on it, but the idea's been in the back of my head [17:02] http://en.wikipedia.org/wiki/K-means_clustering [17:02] right, well 'like' can be scored based on popularity (views) actual content, age, etc [17:02] yea, there's a ton of ways to put the things together [17:02] but then you need to define your distance function, and that can be challenging [17:03] you have to walk through them and see what tends to help get your better scores/results [17:03] yup [17:03] I think all that is the art part of what we do. [17:04] yea, I mean these days, customer purchase info, etc all kinds of ways to need/get invovled in this [17:04] yup. [17:04] amazons "customers viewing this also viewed" is a good example. [17:04] right [17:05] and you can also s/view/purchase/ [17:05] and get a different dataset [17:05] and when you are talking business application it is likely that both are important [17:05] or... i mean, multiple functions are important [17:06] greg-g: is vermont in MI? [17:07] nope, neither are the smokies ;) [17:07] ok, sorry read that too fast [17:07] should have put a "..." in there [17:07] got thinking maybe I went 'too far' [17:07] IMO application frameworks like MapRduce are often symptomatic of overuse of relation database models over object oriented or heirarchial databases [17:07] disagree... :) [17:10] greg-g: check out the NCT, they have some great maps [17:10] cool [17:10] we've driven a few scenic routes that ran through it last year checking it out [17:10] walked a bit down a trail head, it's on our todo when the boy gets older [17:11] my grandmother's house in VA is like 5mi from a trail head [17:11] oh crap, nvm, NCT was the one we did by manistee river [17:11] that's the app trail [17:12] yeah, there is some good lookign NCT up by Duluth, though. But a 12 horu drive away :( [17:12] yea, VA is as well [17:12] how far is the smokeys? [17:13] well 10-12hr [17:14] ish, yeah [17:38] Time to pounce on the mail-dude. [17:38] uh oh, what toy is in the mail this time? [17:44] Just a few more ebay winnings. [17:45] I'm thinking about upgrading from my G1, anyone have any recommendations? The G2 seems underpowered, so I was looking at the Nexus S or Desire HD at the moment [17:45] got some shrinkwrapped GURPS stuff for $1.99 [17:45] My sis in law is thrilled with her G2 [17:45]