00:06:11  * _sorensenquit (Ping timeout: 255 seconds)
00:16:13  <rowbit>Hourly usage stats: [developer: 3, free: 29]
00:26:05  * shykeschanged nick to zz_shykes
00:33:24  * AvianFlujoined
00:41:48  * dbransquit (Quit: Page closed)
00:48:59  * st_lukejoined
01:09:38  * tilgovijoined
01:14:09  <rowbit>SubStack, pkrumins: These encoders are STILL down:
01:16:13  <rowbit>Hourly usage stats: [developer: 7, free: 39]
01:17:22  * jibayquit (Quit: Leaving)
01:47:27  * zz_shykeschanged nick to shykes
01:49:24  * shuaibquit (Quit: Computer has gone to sleep.)
02:06:09  * tphummelquit (Quit: tphummel)
02:15:21  * shuaibjoined
02:16:13  <rowbit>Hourly usage stats: [developer: 0, free: 19]
02:22:42  * AvianFluquit (Remote host closed the connection)
02:28:49  * shykeschanged nick to zz_shykes
02:30:13  * AvianFlujoined
03:14:09  <rowbit>SubStack, pkrumins: These encoders are STILL down:
03:16:13  <rowbit>Hourly usage stats: [developer: 0, free: 28]
03:20:10  * mike-dquit (Read error: Connection reset by peer)
03:20:31  * mike-djoined
03:38:00  * shuaibquit (Quit: Textual IRC Client: http://www.textualapp.com/)
03:59:24  * captain_morganquit (Ping timeout: 260 seconds)
03:59:43  * captain_morganjoined
04:16:13  <rowbit>Hourly usage stats: [developer: 2, free: 25]
04:27:30  * captain_morganquit (Ping timeout: 276 seconds)
04:43:30  * zz_shykeschanged nick to shykes
04:51:52  * jiang-plusjoined
05:14:09  <rowbit>SubStack, pkrumins: These encoders are STILL down:
05:16:13  <rowbit>Hourly usage stats: [developer: 2, free: 33]
05:38:21  <AvianFlu>OMG HI LOUDBOT
05:53:45  * tphummeljoined
05:59:06  <rowbit>/!\ ATTENTION: (default-local) susan@....au successfully signed up for developer browserling plan ($20). Cash money! /!\
05:59:06  <rowbit>/!\ ATTENTION: (default-local) paid account successfully upgraded /!\
06:02:31  * tilgoviquit (Ping timeout: 245 seconds)
06:16:13  <rowbit>Hourly usage stats: [developer: 4, free: 12]
06:42:28  * tphummelquit (Quit: tphummel)
06:44:01  * shykeschanged nick to zz_shykes
06:50:16  <isaacs>ITS LIKE, THE ROYAL WE, MAN
07:16:13  <rowbit>Hourly usage stats: [developer: 1, free: 34]
07:16:19  * AvianFluquit (Remote host closed the connection)
07:31:40  * AvianFlujoined
07:40:33  * AvianFluquit (Remote host closed the connection)
08:16:13  <rowbit>Hourly usage stats: [developer: 4, free: 29]
08:36:05  * st_lukequit (Remote host closed the connection)
08:57:29  * jiang-plusquit (Quit: Lost terminal)
09:16:13  <rowbit>Hourly usage stats: [developer: 5, free: 49]
09:34:29  * saijanai_quit (Read error: Operation timed out)
09:43:07  * saijanai_joined
09:52:27  * doolsquit (Ping timeout: 260 seconds)
09:53:10  * doolsjoined
09:54:19  * purrquit (Ping timeout: 260 seconds)
09:54:40  * purrjoined
09:55:35  * purrquit (Client Quit)
09:55:56  * purrjoined
10:14:54  * dominictarrquit (Ping timeout: 240 seconds)
10:16:13  <rowbit>Hourly usage stats: [developer: 0, free: 115]
10:46:15  * dominictarrjoined
11:00:32  * jibayjoined
11:00:33  * dominictarrquit (Read error: Connection reset by peer)
11:02:56  * LOUDBOTquit (Ping timeout: 245 seconds)
11:03:22  * py1honquit (Ping timeout: 246 seconds)
11:03:29  * py1honjoined
11:04:31  * LOUDBOTjoined
11:16:13  <rowbit>Hourly usage stats: [developer: 0, free: 32]
12:08:00  * AvianFlujoined
12:16:13  <rowbit>Hourly usage stats: [developer: 1, free: 42]
12:49:15  <rowbit>/!\ ATTENTION: (default-local) contolini+browserling@... successfully signed up for developer browserling plan ($20). Cash money! /!\
12:49:15  <rowbit>/!\ ATTENTION: (default-local) paid account successfully upgraded /!\
13:16:13  <rowbit>Hourly usage stats: [developer: 8, free: 45]
14:16:13  <rowbit>Daily usage stats: [developer: 26, free: 397]
14:16:14  <rowbit>Hourly usage stats: [developer: 3, free: 39]
14:24:31  * jiang-plusjoined
14:42:59  * st_lukejoined
14:58:11  * wiwilliajoined
15:05:51  * jiang-plusquit (Quit: Lost terminal)
15:06:01  * _sorensenjoined
15:11:11  <rowbit>SubStack, pkrumins: At least 5 people waiting in the queue for free servers! (Waiting: 5)
15:16:13  <rowbit>Hourly usage stats: [developer: 6, free: 47]
15:20:21  * st_luke_joined
15:21:18  * st_lukequit (Read error: Connection reset by peer)
15:21:21  * st_luke__joined
15:21:32  * st_luke__changed nick to st_luke
15:24:39  * st_luke_quit (Ping timeout: 244 seconds)
15:36:58  * tphummeljoined
15:38:11  * tphummelquit (Client Quit)
15:39:04  * tphummeljoined
15:49:40  * zz_shykeschanged nick to shykes
15:52:08  * dominictarrjoined
15:54:06  <dominictarr>substack, just got some important patches into scuttlebutt today, about persistence.
15:54:28  <dominictarr>although, I want to change it so it just uses the same duplex protocol to save to disk.
15:59:20  * AvianFluquit (Remote host closed the connection)
16:09:00  <substack>nice
16:09:35  <substack>dominictarr: I met somebody last night who was complaining about dynamo and how big table is fundamentally better
16:10:03  <dominictarr>what was their reasoning?
16:10:13  <dominictarr>I havn't read the big table paper yet...
16:11:11  <substack>something about how you can build transactions on top of it
16:11:23  <dominictarr>bah
16:11:39  <dominictarr>although that you can have two keys is interesting...
16:12:10  <substack>http://www.google.com/events/io/2009/sessions/DesignDistributedTransactionLayerAppEngine.html
16:12:23  <substack>is the guy
16:13:00  <substack>oh right global vs local transactions was the big thing
16:13:42  <substack>but it seems like building partitioning on top of scuttlebutt wouldn't be so impossible
16:14:03  <substack>unsure about how to translate that into local transactions though
16:14:38  <dominictarr>you'd have to either send a query out to all instances (impossible, in general)
16:15:21  <dominictarr>or have a central server that adjudicates over each transaction.
16:15:47  <dominictarr>I'm much more interesting in seeing how far we can push transaction-less applications.
16:16:11  <substack>anyways he's a researcher at ucsf doing distributed computing for genomics research
16:16:13  <rowbit>Hourly usage stats: [developer: 8, free: 70]
16:16:32  <substack>this is the paper that goes along with that video http://arxiv.org/abs/1106.3325
16:16:34  * tphummelquit (Quit: tphummel)
16:16:35  <dominictarr>I will watch his talk.
16:17:14  <dominictarr>it sounds like it might be talking about a transactional memory approach
16:18:18  <dominictarr>basically, you queue up all your stuff, and then you check that nothing has changed in the mean time.
16:19:52  <dominictarr>substack, where did you meet him?
16:20:30  <substack>hackers and founders in oakland
16:20:57  <substack>at the uptown
16:27:46  * AvianFlujoined
16:54:19  <dominictarr>ah, this transactional stuff is boring. it's just the old way.
16:55:16  <dominictarr>eventual consistency is interesting territory.
17:01:35  <substack>aha
17:02:58  <dominictarr>I mean "correctness" ? BORING!
17:05:46  * mikealjoined
17:07:18  <dominictarr>like in his talk he mentions that you need transactions to make a bank ... but then there is bitcoin.
17:08:07  <dominictarr>I read this thing once about an alternative interperation of computers that differed from the turing model
17:10:11  <dominictarr>the difference what the a computation did not need to halt to give you an answer... instead it gives you an provisional answer, but can callback later with improved answers...
17:12:04  <dominictarr>ah here it is http://en.wikipedia.org/wiki/Super-recursive_algorithm#Inductive_Turing_machines
17:12:12  <dominictarr>inductive turning machines
17:12:37  <dominictarr>that idea seemed to me to fit RL better.
17:13:11  <dominictarr>you never really know if a given answer idea is correct, and people are used to changing their mind, and dealing with changing information.
17:16:13  <rowbit>Hourly usage stats: [developer: 2, free: 14]
17:27:41  <substack>that probably maps better to how people think
17:28:38  * ehdjoined
17:46:10  <dominictarr>exactly!
17:47:31  <dominictarr>and sometimes you get situations where like, some one accidentally walks in on you on the toilet.
17:47:52  <dominictarr>people just deal with it.
17:48:06  <dominictarr>mostly, they use forgiveness rather than permission.
17:48:19  <dominictarr>transactions are about permission.
17:48:40  <Raynos>this is crazy talk
17:48:46  <Raynos>its interesting though
17:48:55  <Raynos>dominictarr: have you solved streaming UIs yet?
17:49:43  <dominictarr>Raynos, "solved" is rather loaded. I have what I want to make easy UI stuff for crdt.
17:49:55  <Raynos>I see
17:49:59  <dominictarr>it will be useful in a few other cases too.
17:50:15  <dominictarr>I'm not trying to solve every problem,
17:50:41  <dominictarr>instead it's better to make some compromises that make a particular problem easy to solve.
17:51:29  <dominictarr>it should also work well for streaming lists, like search results, or chat rooms, etc.
17:52:03  <Raynos>I see
17:52:07  <Raynos>that might be my problem
17:52:14  <Raynos>I keep trying to solve the entire problem
17:52:19  <Raynos>instead of making compromises
17:52:38  <dominictarr>you should build a boat. that will teach you about compromise.
17:52:46  <dominictarr>compromise in design.
17:53:35  * tphummeljoined
17:54:10  <Raynos>:P
18:02:31  * AvianFluquit (Remote host closed the connection)
18:09:36  * dominictarrquit (Ping timeout: 245 seconds)
18:12:15  * ehdquit (Ping timeout: 260 seconds)
18:16:13  <rowbit>Hourly usage stats: [developer: 0, free: 39]
18:39:50  * AvianFlujoined
18:40:19  * ehdjoined
18:40:34  * dominictarrjoined
18:44:35  * tilgovijoined
18:49:33  * ehdquit (Ping timeout: 252 seconds)
19:03:34  * dbransjoined
19:03:42  <dbrans>howdy
19:04:28  <dbrans>Is there a way to pass url parameters to a testling browser?
19:06:21  <dbrans>Oh, didn't see the argv url parameter. Perfect.
19:14:06  * st_lukequit (Ping timeout: 276 seconds)
19:16:13  <rowbit>Hourly usage stats: [developer: 0, free: 45]
19:17:47  * ehdjoined
19:26:57  * ehdquit (Ping timeout: 252 seconds)
19:27:34  * ehdjoined
19:28:09  * mikealquit (Quit: Leaving.)
19:31:56  * mikealjoined
19:32:30  * ehdquit (Ping timeout: 268 seconds)
19:39:45  * mikealquit (Quit: Leaving.)
19:40:00  * tilgoviquit (Remote host closed the connection)
19:58:57  * ehdjoined
20:02:32  * tilgovijoined
20:09:53  * mikealjoined
20:10:39  * ehdquit (Ping timeout: 276 seconds)
20:16:13  <rowbit>Hourly usage stats: [developer: 1, free: 48]
20:20:17  * mikealquit (Ping timeout: 255 seconds)
20:20:54  * mikealjoined
20:38:42  * ehdjoined
20:39:02  <dominictarr>tanepiper, ping?
20:41:11  <substack>I find myself needing a module to convert streams of one encoding to another encoding
20:41:25  <substack>for instance signed 16 bit integer audio into float arrays
20:41:43  <substack>in order to feed microphone data into an fft library
20:41:53  <substack>that I can use to decode radio teletype
20:42:00  <substack>and for other experiments
20:42:22  <dominictarr>"MAD SCIENCE"
20:42:29  <dominictarr>in other words.
20:43:06  <substack>I already have the shortwave receiver set up and piping data
20:43:24  <substack>and already tuned to a strong rtty signal
20:45:33  <dominictarr>does anyone still use radio teletype?
20:45:58  <substack>yes
20:46:05  <dominictarr>who?
20:46:14  <dominictarr>don't say "BANKS"
20:46:16  <substack>you can still pick up radio teletype signals on shortwave in most places
20:46:16  <substack>I have no idea who
20:46:32  <substack>that's partly why I'm writing a decoder
20:46:35  <substack>to figure out what this signal is actually carrying
20:47:29  <dominictarr>probably people stuck on deserted islands with obsolete equipment that no one has rescued because they don't have radio teletype.
20:47:56  * ehdquit (Ping timeout: 245 seconds)
20:48:06  <substack>packet radio rocks
20:49:55  * mikealquit (Quit: Leaving.)
20:54:32  * jjjohnnnyjoined
20:55:54  <hij1nx>dominictarr: why you not reg your nick?
20:56:21  <hij1nx>tried to memoserv you but couldnt
20:56:23  <dominictarr>hij1nx, I think i tried, but then I forgot my password or something
20:56:30  <hij1nx>lol
20:57:30  <substack>dominictarr: forwarded correspondence with daniel
20:57:52  <dominictarr>substack, just writing a reply with the links to relevant papers right now
20:57:57  <substack>sweet!
20:58:21  <dominictarr>oh, discovered this today too
20:58:21  <dominictarr>http://regal.lip6.fr/spip.php?article135
20:59:26  <dominictarr>gonna email that guy, even though I am a "predoc" since I am basically doing that already.
21:01:08  <hij1nx>woah, why does someone want a postdoc, wouldnt they just want someone who understands the subject and has exp?
21:02:49  <hij1nx>i guess if you dont know the subject you would need to trust something like a cert
21:02:59  <substack>indeed
21:03:06  <substack>you don't need organizations to do research
21:03:09  <substack>just do it
21:04:06  <substack>"candiates must have an excellent academic record"
21:04:09  <substack>WHY!
21:08:40  <isaacs>substack: because it's a lot cheaper to look at your transcript than it is to find out if you know stuff.
21:09:03  <isaacs>substack: and they're not interested in being fair to all applicants, they're interested in finding *one* valid applicant in the search space.
21:09:19  <isaacs>substack: if saying "must have academic stamps" removes more invalid candidates than valid ones, it's a wise move.
21:13:37  <substack>if there is more supply of applicants than demand
21:13:47  <substack>not true in programming generally
21:14:44  * ehdjoined
21:15:31  <dbrans>substack: testling tunnel unexpected response from server: Error: Error: EACCES, open '/home/testling/projects/testling-master/web/lib/../db/tunnel/dbrans@gmail.com.json'
21:16:13  <rowbit>Hourly usage stats: [developer: 2, free: 40]
21:19:06  <dominictarr>hij1nx, this is a new subject.
21:19:24  <dominictarr>that is the guy who's name is on pretty much all the papers on this stuff.
21:20:22  <dominictarr>maybe they think that only academics are interested in this stuff...
21:20:31  <dbrans>anyone else getting this when you "curl -u dbrans@gmail.com testling.com/tunnel/open"?
21:20:40  <dbrans>Error: Error: EACCES, open '/home/testling/projects/testling-master/web/lib/../db/tunnel/dbrans@gmail.com.json'
21:25:07  <dominictarr>substack, incidentally this is a research position, not a corporate position.
21:25:56  * ehdquit (Ping timeout: 256 seconds)
21:29:40  * wiwilliaquit (Ping timeout: 268 seconds)
21:38:45  <rowbit>/!\ ATTENTION: (default-local) kusno@... successfully signed up for developer browserling plan ($20). Cash money! /!\
21:38:45  <rowbit>/!\ ATTENTION: (default-local) paid account successfully upgraded /!\
21:53:00  * ehdjoined
22:04:14  * ehdquit (Ping timeout: 252 seconds)
22:12:55  * tomshredsjoined
22:16:13  <rowbit>Hourly usage stats: [developer: 4, free: 13]
22:28:25  * mikealjoined
22:31:36  * ehdjoined
22:42:44  * ehdquit (Ping timeout: 252 seconds)
22:43:59  * tilgoviquit (Remote host closed the connection)
23:05:57  <dominictarr>substack, looks like that guy is impressed!
23:07:25  <substack>yeah!
23:11:20  * ehdjoined
23:16:13  <rowbit>Hourly usage stats: [developer: 3, free: 42]
23:21:32  * ehdquit (Ping timeout: 256 seconds)
23:31:12  * ehdjoined
23:42:19  <tanepiper>oh cool, browserify: http://www.youtube.com/watch?v=gkb_x9ZN0Vo&feature=g-all-lsb
23:43:49  <Raynos>dominictarr: so eventual consistancy is better then locking?
23:44:22  <dominictarr>Raynos, it's not "better", just cooler.
23:44:34  <Raynos>dominictarr: I have this resource which needs to have Xs created and creating an X has a side effect and takes like 20s. I would have thought of keeping a list of the resources that need Xs and locking them until an X exists
23:44:49  <Raynos>The problem is I dont want to create two Xs in parallel and be eventually consistent because of annoying side effects
23:44:55  <substack>tanepiper: oh sweet, I've been in correspondence with one of those folks
23:46:04  <substack>Raynos: when a node starts to create an X it should spam the network about it
23:46:20  <substack>so that other nodes know that somebody is taking care of creating the X
23:46:42  <Raynos>substack: what if two nodes start an X in parallel
23:46:43  <substack>and if 2 nodes accidentally try to both create an X they will know pretty immediately so you can just abort one of them
23:46:49  <Raynos>Oh I see
23:47:05  <substack>just have a deterministic algorithm for deciding who should abort
23:47:23  <dominictarr>basically, it all depends on the network, if you have predictable network performance (like you might in a fancy data center) then you can do locking
23:47:26  <tanepiper>I've been having a node rant the last few days
23:47:34  <tanepiper>I just want this project to end
23:47:42  <dominictarr>zookeeper, and some google stuff do a cool thing
23:48:11  <dominictarr>where when an update comes in, they know how long it would take to tell the rest of the cluster about it (tracking latency)
23:48:37  <dominictarr>so they know how long they need to wait for news from other servers before they act on it.
23:49:13  <dominictarr>in this way, all the nodes can apply updates yet be distributed, with bounded latency.
23:49:48  <dominictarr>however, this would only work in a data center, but not in a mobile/web app...
23:50:00  <dominictarr>or robots, or cool stuff...
23:50:49  <dominictarr>basically, transactions come down to having async operations that you can cancel.
23:50:57  <jjjohnnny>substack https://github.com/Benvie/live-arrays
23:51:15  <dominictarr>like you start something... and then say "nevermind", or the other guy does.
23:51:55  <jjjohnnny>https://github.com/Benvie/reified
23:55:19  * ehdquit (Ping timeout: 252 seconds)
23:55:23  <dominictarr>or you can say "okay, I'll do it", which means you give up your right to cancel, but are waiting for the final word.
23:56:19  <dominictarr>so say, you are booking a bunch of airline flights... you send a message to all the airlines, and they either say "I can do it" or "flight is full"
23:56:40  <dominictarr>if any say "flight is full", then you cancel all the others.
23:58:21  <Raynos>I think its easier to just use a transactional database
23:58:24  <Raynos>to keep the state
23:59:36  <dominictarr>Also, there is a problem here, if the caller crashes, because he will forget that the he has other people waiting on him, and then will have to go back and find them again...