Experimental IRC log happs-2007-07-19

Available formats: content-negotiated html turtle (see SIOC for the vocabulary)

Back to channel and daily index: content-negotiated html turtle

These logs are provided as an experiment in indexing discussions using IRCHub.py, Irc2RDF.hs, and SIOC.

10:16:28<shapr>yarr
10:18:35<twb>Aah! A pirate
10:19:08<twb>Release the region-free DVDs!
10:22:55<kaol>ACTION has fun with type errors
10:25:55<kaol>I still feel that it might've been a wiser decision to turn to Erlang to do web serving right, but I just find Haskell a nicer language
10:26:38<shapr>hiya twb
10:26:51<twb>kaol: haskell's CML-style concurrency doesn't cut it?
10:26:59<shapr>Cover My Leg?
10:27:04<twb>Concurrent ML
10:27:06<shapr>ah
10:27:14<twb>I assume that's what it's based on, anyway
10:27:27<twb>It's not like I write Haskell regularly
10:27:41<shapr>I don't know CML.
10:28:55<kaol>being able to hot swap code sounds like a nice feature
10:29:10<shapr>Yeah, mnesia is another nice feature.
10:30:06<shapr>But Erlang doesn't have compile-time type checking for example..
10:30:28<kaol>and erlang's distribution support is second to none.
10:31:47<Saizan>in an event-driven model it shouldn't be difficult to hot-swap even in haskell
10:33:01<Saizan>just disconnect the handler and load the new one
10:34:56<shapr>I think Erlang is also strict and impure.
10:35:03<shapr>I'd miss monads very much.
10:35:16<kaol>me too
10:35:21<twb>It doesn't look as nice, either
10:36:11<kaol>and #erlang wasn't nearly as nice as #haskell is
10:36:15<shapr>heh
10:36:22<shapr>Yeah, community counts.
10:36:36<twb>There's a guy in #emacs that IRCs from erlang
10:37:49<shapr>#haskell high water mark is 358, and it has 339 right now. That's pretty impressive.
10:38:51<twb>Bah.
10:39:26<twb>Waah, that's more than #lisp!
10:39:33<twb>Something has gone wrong.
10:42:03<shapr>How so?
10:42:29<twb>Because Lisp is better than Haskell, duh
10:42:40<twb>>duck<
10:43:02<shapr>Well, maybe something is going right then :-)
10:43:06<shapr>Since ...
10:43:17<twb>ACTION braces for figlet output
10:43:22<shapr>heh, nah
10:43:37<shapr>I don't have to be insistent, just consistent...
10:43:54<shapr>I have a fulltime job writing BSD3 licensed Haskell.
10:44:45<twb>I'm too stupid and inexperienced to manage that
10:44:55<shapr>I doubt that.
10:45:11<twb>I did get paid to write some scsh, but not much
10:45:17<shapr>Lots of Haskell companies looking for more employees.
10:47:53<twb>Seems to be 99% PHP/Java/Python around here.
10:49:26<shapr>Well, I upgraded from VB to Java to Python to Haskell.
10:49:33<shapr>In only seven years!
10:49:55<twb>Part of the problem is me not wanting to just start with Haskell
10:50:01<twb>Or Lisp, or Scheme
10:50:57<shapr>Why not?
10:51:04<twb>Lack of experience
10:51:09<shapr>That's circular.
10:51:10<kaol>ACTION wouldn't mind a Haskell job either
10:51:11<twb>It's like wanting to start your career as CEO
10:51:29<shapr>How do you get experience if you don't want to start with something?
10:51:35<shapr>Or did I misunderstand?
10:51:54<shapr>I got about four hours of sleep, so confusion is my first name today.
10:51:57<twb>Because nobody wants to hire crap Haskell programmers, but you can easily get a job doing crap PHP
10:52:16<shapr>Oh, I just wrote code in my language of choice after hours.
10:52:26<shapr>And found ways to use my language of choice at work as well.
10:55:57<shapr>twb: So jump in!
10:56:07<twb>Well, I've tried
10:56:20<shapr>You seemed to be doing just fine when you were hanging out on #haskell
10:56:20<twb>I tried writing a replacement editor for Emacs
10:56:35<twb>And a replacement WM for Ratpoison
10:56:51<shapr>yi, xmonad?
10:56:56<twb>I didn't start trying to replace TeX, because I couldn't work out how.
10:57:04<twb>shapr: this was before they existed
10:58:35<shapr>Sounds like you are already qualified to hack on Yi and xmonad then.
10:58:42<twb>You don't get it
10:58:45<shapr>Should make it easier to jump into those projects.
10:58:46<twb>I failed. Miserably
10:58:55<shapr>That's not a crime :-)
10:59:19<shapr>Failing to complete a piece of fun software teaches you lots.
10:59:19<twb>Because ncurses and xlib are deeply procedural
10:59:55<shapr>Well, x11-extras is used in xmonad.
10:59:56<twb>And I got real bogged down trying to use various FFIs
11:00:32<shapr>One good way to learn how to do something is to try and fail to do that thing many times.
11:00:41<twb>I guess.
11:00:43<shapr>Humans learn more from failure than from success.
11:01:08<shapr>So, keep on jumping in and trying new things.
11:01:16<shapr>And keep track of what didn't work, and why it didn't work.
11:02:51<twb>Heh, it.seek.com.au lists just two vacancies that refer to Haskell.
11:03:04<twb>(For all of .au)
11:03:12<shapr>Sounds like monster.com or dice
11:03:27<shapr>But if you're in the #haskell community, there are many more jobs available.
11:03:33<twb>It some thing that seems to be mostly used by headhunters
11:03:51<shapr>Most Haskell companies don't use headhunters.
11:03:53<twb>i.e. places where hiring is done by H.R. rather than clued people
11:03:55<twb>Right.
11:07:00<shapr>Ok, back to sleep for me.
17:00:49<msouth>is there a backup strategy for persistent data?
17:02:12<msouth>I am just playing around with the apps in the tutorial, and I see all the files it is creating to maintain state on the filesystem. Is there a way to copy those files off at some interval to have a backup in case that disk fails?
17:07:50<shapr>msouth: I'm not sure if you can copy off the data while it's running, but I know you can shut down, copy, and restart.
17:08:49<shapr>msouth: Lemmih says that HAppS only appends data, so you can copy while it's running.
17:09:10<shapr>Lemmih also says that the separate archive directory could be backed up to have a continuous history.
17:15:25<twb>msouth: rsnapshot?
17:16:04<twb>You can use LVM to snapshot running databases, which means you only need to shut down the app for a few seconds, rather than the whole rsync time
17:17:02<twb>When doing this with scalix, you can say to scalix "hey, don't write to the database for the next 60 seconds please", and then once you've made the LVM snapshot (which probably takes like 5 seconds) you say "OK, go again"
17:22:24<msouth>suppose I had two webservers with a mount of the same filesystem, and one is a read-only mount. Could it point to those files, derive the state, and serve read-only requests?
17:22:39<msouth>I apologize if I'm asking annoyingly newbie questions.
17:23:07<msouth>I'm hoping to get this considered in our new service-oriented architecture
17:23:34<msouth>and I'm going to get hammered with questions about backups, reliability, etc.
17:23:45<twb>I don't know; I'm new here.
19:45:46<jmelesky>does the recent lull in commits mean we're approaching the next release?
19:46:38<shapr>Lots of commits in the fork
19:46:42<shapr>And the fork should be public soon
19:47:00<shapr>With big notices included that this it's not a real release.
19:47:53<jmelesky>it's just the reorg prior to pushing for the release?
21:12:07<msouth>something that would be really nice: heavy comments in the stuff in examples/
21:13:19<kaol>do all of them even compile with the current version?
21:13:34<jmelesky>yeah, i'd settle for working examples
21:13:59<jmelesky>iirc, the ones on the wiki compile, but not the ones in the repository
21:14:06<jmelesky>it's been a few weeks since i've tried, though

Back to channel and daily index: content-negotiated html turtle