| --- Log opened Fri Feb 03 00:00:23 2012 |
| 00:00 | < celticminstrel> | Is that the forum I think you're talking about? |
| 00:00 | <&McMartin> | Very unlikely. |
| 00:02 | <&McMartin> | Most of these people I didn't get to know until we splintered of the one he describes to a privately-controlled one which was itself mostly a subset of Nightstarites. |
| 00:02 | <&McMartin> | The causality gets tangled. |
| 00:05 | <@Alek> | celtic: probably. |
| 00:06 | <@Alek> | mcm: erh. I can count, oh... maybe a dozen people here, offhand, if not more, who have had a presence either on the forum or on the associated channel. |
| 00:07 | <&McMartin> | Yeah. But that's not the one he thinks we're talking about. He thinks we're goons~ |
| 00:07 | <@Alek> | ... pfft. |
| 00:07 | < celticminstrel> | I was thinking of CRFH. |
| 00:07 | <@Alek> | although by this point in time, more than a decade since joining, it's hazy as to who I recall from there, and who I recall from elsewhere. >_> |
| 00:07 | <&McMartin> | Yeah, it gets tangled |
| 00:07 | <@Alek> | celtic: well, yes. |
| 00:07 | <&McMartin> | Especially so since a lot of it may have been "the old Keenspot forums generally" |
| 00:07 | <@Alek> | yeah. |
| 00:08 | <&McMartin> | Nightstar specifically I was more associated with the Avalon community for some time, before it moved on. |
| 00:08 | <@Alek> | well, I only hung on one other keen forum, and that only for a very short time and in brief spurts. |
| 00:08 | < celticminstrel> | I don't think I met anyone here through there, except Alek. |
| 00:08 | <@Alek> | celtic: and yet, a LOT of the people you see here, I first met there. XD |
| 00:08 | <@Alek> | or in the associated channel. |
| 00:08 | <&McMartin> | Or its spinoff channels |
| 00:08 | <@Alek> | but only a very few for the latter. |
| 00:09 | <&McMartin> | Or the channels for other KeenSpot comics at the time, etc. |
| 00:09 | < celticminstrel> | A couple of the people here I first met in its channel. |
| 00:09 | <@Alek> | heh. I haven't been in the space one in years. XD |
| 00:09 | <@Alek> | for some reason, don't exactly feel welcome there. D: |
| 00:09 | <&McMartin> | It happens. |
| 00:10 | <&McMartin> | The splinter channels are all community channels and if you don't fit in, you don't fit in. |
| 00:10 | <&McMartin> | That's basically what happened with the Avalon diaspora |
| 00:10 | <@Alek> | mmh. |
| 00:10 | <@Alek> | I liked that comic, too. |
| 00:10 | <&McMartin> | Happily, I still got to reuse the Josh Phillips art for my first major IF project~ |
| 00:11 | <@Alek> | and actually, by "not feeling welcome", I mean I've been forcibly ejected from there at least once. and locked out. |
| 00:11 | <@Alek> | apparently partly due to some misunderstanding over someone using my name in vain at one gathering. |
| 00:11 | <@Alek> | way back when. -_- |
| 00:14 | <&McMartin> | "Being welcoming" isn't a goal when you're a splinter community~ |
| 00:18 | < RichyB> | I know I arrived at Nightstar via Schlock Mercenary. |
| 00:19 | <&McMartin> | Schlock, CRFH, and Avalon were the three main influx points, yeah. |
| 00:19 | < RichyB> | and I think I stuck around specifically because of the "Code Monkeys" forum, hee. :) |
| 00:19 | <&McMartin> | Yeah, #code is the only "open" community I'm in on IRC, actually. |
| 00:19 | <&McMartin> | Three splinters, two alumni networks, and two special-purpose project channels. |
| 00:20 | < RichyB> | Huh. |
| 00:20 | <&McMartin> | And #code. |
| 00:20 | <&McMartin> | So yeah, for me IRC is basically Glorified IM~ |
| 00:20 | < RichyB> | I hang out on a load of channels on freenode just for the sake of repairing my bruised karma. |
| 00:20 | < RichyB> | (Help random newbies, acquire grudging pseudo-respect.) |
| 00:20 | < RichyB> | Ah. For me, IRC is Corporate IM. |
| 00:21 | < RichyB> | As in I installed an ircd on a box in my employer's office and now it's the official company internal messaging system. |
| 00:21 | <&McMartin> | Nice |
| 00:21 | < RichyB> | Yeah. They were using iChat before - all based on Bonjour and Zeroconf or whatever it is. |
| 00:22 | < Reivles> | #Code is one of our more successful attempts at a general-purpose channel~ |
| 00:22 | < RichyB> | Multicast packets that go flying off in the 164.254.x.x range and have TTLs that limit them to one ethernet segment. |
| 00:22 | < RichyB> | Terribly, *terribly* unreliable. |
| 00:22 | | * Reivles salutes Chalcy and the Interesting Problem that started it all. |
| 00:23 | < Reivles> | (To whit: She needed Interesting Programming. She was in one set of channels, Vorn another. I made Code to bridge the gap... and it ended up bridging all of them.) |
| 00:23 | < RichyB> | IRCds on the other hand have been around for more than 20 years now, and it's real ordinary damned TCP. So it's absolutely rock-solid. :) |
| 00:23 | < RichyB> | Reivles: ooh ooh STORY TIME STORY TIME is it can it plz be story time now? |
| 00:24 | < Reivles> | snrk |
| 00:24 | < Reivles> | That was, um, pretty much it |
| 00:24 | | You're now known as TheWatcher[T-2] |
| 00:24 | < Reivles> | With a minor segue when TF saw the channel get regged and lept in with "OH FUCK YES", a little subtle plugging to get the Code Talk into here (And for the benefit of everyone else, /out/ of the splinter channels), and...~ |
| 00:25 | < RichyB> | Well it was a good story. |
| 00:25 | < RichyB> | Thank you. |
| 00:26 | <@ToxicFrog> | I actually ended up helping Chalcy with some data visualization stuff, too. |
| 00:29 | | You're now known as TheWatcher[zZzZ] |
| 00:30 | | Attilla [Obsolete@Nightstar-ffe96ba6.as43234.net] has quit [Ping timeout: 121 seconds] |
| 00:40 | < Rhamphoryncus> | RichyB: IRC.. and rocksolid? Well, if it's just one I suppose. Spanning trees are anything but. |
| 00:43 | < RichyB> | Yeah, okay. |
| 00:44 | < RichyB> | If it were actually distributed then I'd have troubles with that. But if we were big enough to have that as a problem then I'd solve it at layer 3 or 4, not layer 7. :) |
| 00:44 | <@Namegduf> | You can't. |
| 00:45 | < RichyB> | Redundant networking kit and lines, rather than trying to find a chat system that's actively failure-resistant at the application level. |
| 00:45 | <@Namegduf> | There's more to it. |
| 00:45 | < RichyB> | Namegduf: what am I missing? |
| 00:45 | <@Namegduf> | Fundamentally what the IRC servers are maintaining with links, in addition to message relaying, is maintaining identical copies of a database. |
| 00:45 | <@Namegduf> | You can't solve distributed databases at layer 3 or 4 |
| 00:46 | <@Namegduf> | Servers go offline and you need to ensure everyone is still synced |
| 00:46 | < RichyB> | IRC already solves that, albeit in a way that doesn't scale particularly well. |
| 00:46 | <@Namegduf> | No one got changes others didn't, that kinda thing. |
| 00:46 | <@Namegduf> | Not in a way that could avoid bursts. |
| 00:46 | | Attilla [Obsolete@Nightstar-5b15dd03.as43234.net] has joined #code |
| 00:47 | <@Namegduf> | When a server drops, the burst resyncs the in-memory data. |
| 00:47 | < RichyB> | Not something that I'm going to worry about until we get at least 200 employees here. |
| 00:47 | < RichyB> | and two locations that actually have enough people to need separate IRC servers. |
| 00:47 | <@Namegduf> | That's fine. |
| 00:48 | < RichyB> | Given that we've (AIUI) never grown beyond ~50 people, I'm not panicking about it. |
| 00:48 | <@Namegduf> | I would recommend not worrying about it at all. |
| 00:48 | <@Namegduf> | It's just been an area of interest for me and I found there wasn't really any straightforward way to solve it. |
| 00:48 | < RichyB> | The memory overhead of the completely replicated state is... irrelevant, and has been since RAM hit ?16/gigabyte a few years ago. |
| 00:48 | < RichyB> | Oh yes, I do understand the CAP theorem and I am aware of the tradeoffs that IRC makes. |
| 00:48 | <@Namegduf> | Yeah, that's not the problem. |
| 00:49 | <@Namegduf> | The problem is agreeing on changes. |
| 00:49 | < RichyB> | It's just that the constants are so tiny that I can get away with ignoring them for quite a long time yet. |
| 00:49 | < RichyB> | We will have to ditch IRC because of the O(n**2) messages a long, long time before we'll have to ditch it because of the CAP theorem. |
| 00:50 | <@Namegduf> | CAP theorem's not the big deal. |
| 00:50 | <@Namegduf> | Drop the P, bang, you're golden. |
| 00:50 | < celticminstrel> | What do you mean by O(n^2) messagess? |
| 00:50 | <@Namegduf> | IRC provides neither of the other two, though |
| 00:51 | <@Namegduf> | Consistency is the big thing it can't provide without rebursting, and potentially having to change state on both sides and revert changes |
| 00:51 | < RichyB> | celticminstrel: if we got to 100 people in a single channel then we'd have to ditch IRC because every time anybody speaks in any channel, it distracts 100 other people. |
| 00:51 | <@Namegduf> | Split the channel? |
| 00:51 | < RichyB> | The bottlenecking resource being _human attention_ rather than cycles, bandwidth or memory. |
| 00:51 | < celticminstrel> | It distracts what now? |
| 00:52 | <@Namegduf> | What I'm talking about is a design problem, though, not a resource problem |
| 00:52 | <@Namegduf> | IRC fundamentally doesn't obtain consensus on changes in a reliable way, so you can't really solve the problems of spanning tree needing to relink at the network layer. |
| 00:53 | < RichyB> | Employees will either stop paying attention to IRC (bad, defeats the point of its existence as corporate IM) or they'll get unduly distracted by it. |
| 00:53 | | eckse [eckse@Nightstar-e77ea8ce.dsl.sentex.ca] has joined #code |
| 00:53 | | mode/#code [+o eckse] by ChanServ |
| 00:53 | <@Namegduf> | In the case of a dropped server I don't think that's even possible, really |
| 00:53 | < RichyB> | Namegduf: eh, isn't just sending the entire state reliable enough? |
| 00:54 | <@Namegduf> | Having to reburst and relink is the mentioned problem of spanning tree.# |
| 00:54 | <@Namegduf> | But that's not gaining consensus, that's syncing things already desynced. |
| 00:55 | <@Namegduf> | You can pull off some fun bugs on modern IRCDs, anyway, where you set a mode on two servers at once and the mode changes cross on the wires |
| 00:55 | <@Namegduf> | And they end up in an inconsistent state |
| 00:55 | <@Namegduf> | (Until split/relink, but they don't notice) |
| 00:55 | < RichyB> | Heh. |
| 00:56 | <@Namegduf> | Fundamentally each server just pretends it is king and makes changes and tells everyone else, and where two changes happen at once that conflict, anything can happen. |
| 00:57 | <@Namegduf> | The dropping-message aspect could be improved some, so long as it wasn't the client's own server which went down. |
| 00:58 | | * Alek snerks. |
| 01:00 | <@Alek> | would be nice if there was some sort of function that let you select a subset of the users to speak with, maybe something that expanded on creating a side channel and inviting all the people you want, to make it more painless. |
| 01:01 | <@Alek> | it'd go some way to fixing the 0(n^2) problem. |
| 01:01 | <@Alek> | of course, eventually you'd end up with a fuckton of separate channels, yes. but maybe there's ways to trim that too. |
| 01:01 | < RichyB> | I think you end up reinventing twitter. |
| 01:01 | <@Alek> | pfft. |
| 01:02 | <@Alek> | fuck the character limit. |
| 01:02 | < RichyB> | No really. Correctly used, hashtags work sorta kinda almost like channels. |
| 01:02 | <@Alek> | so I gather. |
| 01:02 | <@Alek> | not that I really know, as I have had nothing whatsoever to do with twitter, so far. |
| 01:02 | < RichyB> | Well yes, you don't use *twitter* but you make something that resembles it. |
| 01:02 | <@Alek> | mnyeah. |
| 01:03 | < RichyB> | For starters, hashtags on twitter are almost useless because a critical mass of people don't have the barest understanding of how they're supposed to actually work and misuse the damn things in entertaining stupid ways. |
| 01:03 | < RichyB> | So your first step is user re-training. ;) |
| 01:03 | <@Alek> | what I was thinking was, you define a "channel" that's basically a "mailing-group" - any hashtags you send to there are seen by a specific subset of people. or so. |
| 01:04 | <@Alek> | defined by yourself, mind. |
| 01:04 | <@Alek> | of course, then you run into the mirroring problem. |
| 01:05 | <@Alek> | which is where I like my irc analogy - you create this "channel", and it automatically invites the needed people in - then any reply by them in there is seen by just those. |
| 01:05 | < RichyB> | UI issue I have with that is "oh which window did I just type that in?" |
| 01:06 | < RichyB> | Might not actually be a problem. |
| 01:06 | <@Alek> | twitter is just, from what I can gather, a simplified and modified version of irc for mobile phones and suchlike devices. |
| 01:06 | <@Alek> | richy: yeah... |
| 01:06 | <@Namegduf> | Twitter's got the "subscribe to who you want to hear" thing. |
| 01:06 | <&McMartin> | Twitter direly needs @mutec |
| 01:08 | < RichyB> | @mutec? |
| 01:08 | <@Alek> | nameg: yeah, but what I meant is the other direction. |
| 01:08 | <@Namegduf> | Yep. |
| 01:08 | <@Alek> | send only to those people you want to hear it. |
| 01:09 | <@Alek> | as in, you control who's subscribed. |
| 01:09 | <@Alek> | and you're the one who adds them. |
| 01:09 | < RichyB> | One of the models that Twitter supports quite well is retroactively joining a channel (by opening a search window for a particular hashtag) |
| 01:09 | <@Alek> | we wouldn't want this public, mind (*shudders* spammers), but in a private or business environment it'd be awesome. |
| 01:10 | < RichyB> | If I were redoing twitter for corporate use I'd bump the message length to 1200 bytes (no point being too much smaller than ethernet MTU ^_^) and use tagging as an alternative to channels. |
| 01:11 | <@Namegduf> | One of the models it doesn't support is private channels. |
| 01:11 | <@Alek> | mnyeah. |
| 01:11 | <@Namegduf> | And back-and-forth on it is extremely awkward |
| 01:11 | < RichyB> | and I'd ditch following *people*. I'd implement following *channels* instead. |
| 01:11 | <@Namegduf> | And involves totally ruining your feed |
| 01:12 | <@Namegduf> | I also think combining all your channels into one feed would be horribly painful, but that's just me. |
| 01:12 | <@Namegduf> | I understand some IRC users do that. |
| 01:12 | <@Namegduf> | Aside that it would basically work. |
| 01:12 | <@Namegduf> | I don't think it's a replacement today, though. |
| 01:12 | < RichyB> | Then the only reason for peeking at someone's feed is to peek at what channels they're talking on. |
| 01:13 | <@Namegduf> | Viewing, too, would need to be channel-by-channel with a UI for easy switching between them and notifications, including when the window isn't open, of activity. |
| 01:13 | <@Namegduf> | It's possible but it hasn't been done. |
| 01:14 | < RichyB> | Fun. I should actually write the thing that is in my head right now as software. :) |
| 01:21 | | Derakon[AFK] is now known as Derakon |
| 01:34 | < maoranma> | RichyB: Sex software? |
| 01:47 | | Rhamphoryncus [rhamph@Nightstar-5697f7e2.abhsia.telus.net] has quit [Client exited] |
| 02:04 | | gnolam [lenin@Nightstar-202a5047.priv.bahnhof.se] has quit [[NS] Quit: Z?] |
| 02:20 | <@Eri> | Hmm, that's weird. |
| 02:21 | <@Eri> | I'm looking at a pdf document. Everywhere I expect there to be pi, I get an inequals sign. When I highlight it, it becomes p. |
| 02:21 | <@Eri> | I saw the same thing in a spec-sheet, as well, where microamps became infinite-sign amps |
| 02:21 | <@Eri> | Wonder if something's screwy in my localization, or something |
| 02:22 | | Kindamoody[zZz] is now known as Kindamoody |
| 02:33 | <@himi> | You may have a screwy font |
| 02:54 | | Attilla [Obsolete@Nightstar-5b15dd03.as43234.net] has quit [Ping timeout: 121 seconds] |
| 03:42 | | kazrikna [kazrikna@4CA975.462904.08E9A2.992A2D] has joined #code |
| 03:58 | <~Vornicus> | Okay. B-spline calculations set up. |
| 04:04 | <~Vornicus> | Now to do /calculus/. |
| 04:15 | | Eri [Eri@Nightstar-3e5deec3.gv.shawcable.net] has quit [Ping timeout: 121 seconds] |
| 04:15 | <~Vornicus> | (fuck yeah, calculus.) |
| 04:17 | | Eri [Eri@Nightstar-3e5deec3.gv.shawcable.net] has joined #code |
| 04:20 | | Alek [omegaboot@Nightstar-efc8dc53.il.comcast.net] has quit [Ping timeout: 121 seconds] |
| 04:24 | | Alek [omegaboot@Nightstar-efc8dc53.il.comcast.net] has joined #code |
| 04:24 | | mode/#code [+o Alek] by ChanServ |
| 04:39 | < maoranma> | So, in Transcendence, the mining laser does 2.5HP x5 at 2 shots/s |
| 04:39 | < maoranma> | But the starter gun does 4.5hp x1 @ 3.8/s |
| 04:40 | < maoranma> | 25 dps on the laser... |
| 04:40 | <~Vornicus> | There's damage reduction; higher damage values punch above their apparent weight. |
| 04:41 | < maoranma> | 17.1 dps on the gun |
| 04:41 | < maoranma> | I see |
| 04:41 | <~Vornicus> | COnsider the situation where each hit gets reduced by 1 point. |
| 04:41 | < maoranma> | 15 DPS |
| 04:41 | < maoranma> | lol |
| 04:42 | <~Vornicus> | The gun still does less damage in this particular example, but it's gained efficiency. |
| 04:42 | < maoranma> | True |
| 04:42 | < maoranma> | So my aim would matter |
| 04:42 | <~Vornicus> | let me check... |
| 04:42 | <@Alek> | your notes? |
| 04:43 | <~Vornicus> | No, my calculator. |
| 04:43 | <~Vornicus> | At damage reduction above 1.27 per hit, the starter gun is doing more damage. |
| 04:43 | < maoranma> | That would be 127%? |
| 04:44 | <~Vornicus> | No, what i mean is |
| 04:45 | <~Vornicus> | On each hit, the damage taken is reduced by 1.27 hit points, to a minimum of 0. |
| 04:45 | < maoranma> | Well, the armor is listed in % though |
| 04:46 | <~Vornicus> | I could've sw... oh, that's why. |
| 04:47 | | * Vornicus hunts around for the system documents on transcendence. |
| 04:49 | <~Vornicus> | What laser are you using? |
| 04:49 | < maoranma> | Mining laster |
| 04:49 | <~Vornicus> | rather, the non-mining-laser gun? |
| 04:49 | < maoranma> | laser* |
| 04:49 | < maoranma> | Nope, it's totally for mining |
| 04:50 | < maoranma> | MINING PEOPLE OUT OF THEIR HULLS PEW PEW |
| 04:50 | <~Vornicus> | (the, uh, starter gun, I mean.) |
| 04:51 | < maoranma> | A recoilless cannon |
| 04:54 | <~Vornicus> | Ah, I see. Okay, the recoilless is in fact a lower tier gun (1 instead of 2); however, it's got lower power use, and longer range. |
| 04:54 | <~Vornicus> | oh, and some knockback. |
| 04:55 | < maoranma> | I got dead |
| 04:55 | < maoranma> | It's hard to do those starter missions with anything but the figher ship >_< |
| 04:55 | <~Vornicus> | I never had any trouble in the freighter. |
| 04:57 | < maoranma> | Was in the yacht |
| 04:59 | < maoranma> | Wow |
| 04:59 | < maoranma> | This place is selling a turbo laser cannon |
| 05:17 | < maoranma> | Frak |
| 05:18 | < maoranma> | This stupid yacht has the worst sheilds |
| 05:19 | <&Derakon> | Should've gone for the smartyacht. |
| 05:19 | < maoranma> | No, the iYacht was more expensive |
| 05:20 | <~Vornicus> | CALCULUUUUUUS |
| 05:20 | < maoranma> | Is there a way to fine tune aiming? |
| 05:58 | <~Vornicus> | I should do some more work on my piecewise thing; right now it doesn't do any matching of domains, so I can't tell if (for instance) I should keep ends of certain intervals closed when doing derivatives. |
| 06:04 | <~Vornicus> | (situation: A piecewise function may be C1 continuous because one of the two intervals is closed at that point, and they both limit to the same value and derivative, but currently my derivative method on it makes dropout discontinuities at every such location instead.) |
| 06:05 | | * Alek is tempted to \o/-meme, but refrains. XD |
| 06:05 | <~Vornicus> | whut |
| 06:05 | <~Vornicus> | I do not have a cat that is also a mathematician. |
| 06:18 | | * ToxicFrog gives enceladus a cookie |
| 06:24 | | Vash is now known as Vash[Sleeping] |
| 06:27 | < maoranma> | Wow, that sucked, the entire first system had nearly zero fuel |
| 06:33 | | Derakon is now known as Derakon[AFK] |
| 06:37 | | gnolam [lenin@Nightstar-202a5047.priv.bahnhof.se] has joined #code |
| 06:42 | | eckse [eckse@Nightstar-e77ea8ce.dsl.sentex.ca] has quit [Client closed the connection] |
| 06:43 | <@ToxicFrog> | Fuck windows in hell forever |
| 06:43 | <@ToxicFrog> | Fixed version of steam2backloggery released |
| 06:43 | <@ToxicFrog> | Enceladus updated |
| 06:43 | <@ToxicFrog> | Windows binaries of steam2backloggery and categories released |
| 06:45 | <&McMartin> | Grats |
| 06:47 | <@ToxicFrog> | Enceladus still needs some more polishing, but it's good enough for me to use it for releases now. |
| 06:48 | <@ToxicFrog> | (in particular it now supports cross-embedding, which was the big thing it was missing) |
| 06:58 | < celticminstrel> | Within a Deep Forest beaten. |
| 07:03 | <@ToxicFrog> | Woot |
| 07:03 | <@ToxicFrog> | high0 |
| 07:10 | < celticminstrel> | ? |
| 07:10 | < celticminstrel> | high0? |
| 07:10 | < maoranma> | He lost his fingers in the war, insensitive much? |
| 07:11 | < maoranma> | He types with his nose and tongue |
| 07:19 | <@ToxicFrog> | celticminstrel: well, as a sphere, you have no fingers... |
| 07:21 | | celticminstrel [celticminst@Nightstar-5d22ab1d.cable.rogers.com] has quit [Ping timeout: 121 seconds] |
| 07:30 | | himi [fow035@D741F1.243F35.CADC30.81D435] has quit [Ping timeout: 121 seconds] |
| 09:47 | | You're now known as TheWatcher |
| 10:08 | | Rhamphoryncus [rhamph@Nightstar-5697f7e2.abhsia.telus.net] has joined #code |
| 10:14 | | RichyB [mycatverbs@Nightstar-3b2c2db2.bethere.co.uk] has left #code ["ERC Version 5.3 (IRC client for Emacs)"] |
| 10:15 | | RichyB [mycatverbs@Nightstar-3b2c2db2.bethere.co.uk] has joined #code |
| 10:19 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has joined #code |
| 10:19 | | mode/#code [+o himi] by ChanServ |
| 10:36 | | Attilla [Obsolete@Nightstar-fe7a3ed4.as43234.net] has joined #code |
| 11:14 | | Kindamoody is now known as Kindamoody|out |
| 11:16 | < gnolam> | Kindamoody: So. Get anything done at the game jam? |
| 11:16 | <@Kindamoody|out> | Some, but I g2g |
| 12:14 | | Number3 is now known as AnnoDomini |
| 12:14 | | AnnoDomini is now known as ShellNinja |
| 13:42 | | Stalker [Z@Nightstar-3602cf5a.cust.comxnet.dk] has quit [Ping timeout: 121 seconds] |
| 14:13 | | Stalker [Z@Nightstar-5aa18eaf.balk.dk] has joined #code |
| 15:21 | | Vash[Sleeping] is now known as Vash |
| 15:28 | | RichyB [mycatverbs@Nightstar-3b2c2db2.bethere.co.uk] has quit [Connection closed] |
| 15:28 | | RichyB [mycatverbs@Nightstar-3b2c2db2.bethere.co.uk] has joined #code |
| 16:01 | < gnolam> | Hmm. That was way too unpainful. Something must be wrong. |
| 16:01 | < gnolam> | This is VHDL. There must be constant, headdesking pain. |
| 16:05 | < Rhamphoryncus> | Are you using git with it? |
| 16:12 | | cpux|2 [cpux@Nightstar-c5874a39.dyn.optonline.net] has joined #code |
| 16:14 | <@ShellNinja> | VHDL isn't so bad. |
| 16:14 | <@ShellNinja> | But then again, I play DF and Nethack and EU3 for fun. |
| 16:14 | | cpux [cpux@Nightstar-c5874a39.dyn.optonline.net] has quit [Ping timeout: 121 seconds] |
| 16:18 | < gnolam> | Rhamphoryncus: No? |
| 16:19 | < Rhamphoryncus> | gnolam: well apply some git and the constant, headdesking pain will surely follow |
| 16:20 | < gnolam> | I'm on Windows. Wouldn't dream of using git. |
| 16:22 | < gnolam> | (Well, that's actually one of /two/ reasons I'm avoiding git. The other is that its advocates speak of it in /religious/ terms rather than arguing actual merits, which is the warning flag of all warning flags in IT.) |
| 16:22 | < gnolam> | Any particular reason why git and VHDL doesn't mix? |
| 16:26 | < Rhamphoryncus> | Nothing to do with VHDL. Git itself is more than enough for that reaction. |
| 16:39 | < RichyB> | Except for the Windows support (which IMHO is a completely reasonable problem to have with it) I don't really get the distrust. |
| 16:39 | < RichyB> | Git's a single data structure (DAG of repository states, with each repository state having a vector of pointers to all of its ancestors) which is directly exposed in the UI. It's not all that complicated. |
| 16:43 | < gnolam> | Yes? I don't really care what makes my VCS tick. I'd be happy with the answer "gnomes", as long as it works. |
| 16:47 | <@Tamber> | RichyB, it's easier to simplify to "gnolam hates everything" ;) |
| 16:50 | < RichyB> | gnolam: the irony is that making a statement like "fuck git" in public is just going to invoke the ire of mouthbreathing Giterati like myself and get you bombarded with even more verbiage about why you need to CONVERT NOW YOU HERETIC |
| 16:51 | < RichyB> | Even "I don't like Git" will get a similar reaction, though with less brimstone attached. |
| 16:51 | < kazrikna> | heh. |
| 16:52 | | * kazrikna prefers Mercurial to Git, but will use either if really needed. |
| 16:54 | < kazrikna> | they're fairly similar, I just prefer things originally prototyped with a high level language instead of ones prototyped with Perl and shell scripts. ;) |
| 16:54 | < gnolam> | Well, that's the thing. The argument tends to be "$VCS_YOU_ARE_USING sucks [no explanation given]! So you should use Git instead, because... Because. If you're not using Git, you're an idiot." |
| 16:55 | <@jerith> | You should be using VCS. |
| 16:55 | < RichyB> | You should be using RCS. |
| 16:56 | < kazrikna> | I've already moved on from rcs... heh. |
| 16:56 | <@jerith> | But not svn. Or CVS. Or anything that predates CVS. |
| 16:56 | < kazrikna> | I used that way back in college, before learning svn. |
| 16:56 | < RichyB> | If you're not using RCS then you must be an idiot because you're incapable of completing entire projects solo and you can't do all your work in a perfectly linear function without ever branching or trying experiments. |
| 16:56 | < kazrikna> | heh. |
| 16:57 | < RichyB> | Real Programmers don't cooperate, they set landmines for that other bastard. |
| 16:57 | < kazrikna> | rcs is pretty antiquated. I do like branching, which is why I switched from SVN to Mercurial. |
| 16:57 | <@jerith> | RichyB: Yeah. Like themselves, next week. |
| 16:58 | < kazrikna> | I used to do a lot of branching in svn, but the merging was more painful than it needed to be. |
| 16:58 | <@Tamber> | jerith, but real programmers will never encounter said landmines, because they'll just rewrite from scratch because they have a better way of doing it. ;) |
| 16:58 | < RichyB> | kazrikna: now, at the moment I'd prefer to be mixing Git or Hg and SVN. |
| 17:00 | < RichyB> | Centralised VCS still has a really important use case - for things textures and 3d art assets where the files are huge, there is no sensible merge mechanism even in principle and you actually want to be able to lock files so that no one else wastes their time trying to edit them concurrently. |
| 17:00 | < kazrikna> | At work, we still use SVN for the big archives of binary files that the field techs and QA people generate. |
| 17:00 | < Rhamphoryncus> | RichyB: the theory of git may be simple but it has no end of pain |
| 17:00 | < gnolam> | I used to do a lot of branching. Then I took an arrow in the knee. |
| 17:01 | < kazrikna> | But for text based code, it's purely mercurial. |
| 17:01 | < Rhamphoryncus> | caused me* |
| 17:02 | < Rhamphoryncus> | hg has been just as bad to me |
| 17:02 | < RichyB> | Rhamphoryncus: what materials were you following trying to pick it up? |
| 17:02 | < Rhamphoryncus> | Various things over the years |
| 17:03 | | * kazrikna just used hginit.com, that was enough. |
| 17:03 | < Rhamphoryncus> | Note, I wasn't starting my own private repository. It's always one of an existing project that I'm just trying to create a local branch of (or 2) and track |
| 17:04 | < Rhamphoryncus> | Many of which are gitsvn mirrors or the like |
| 17:05 | <@froztbyte> | http://imgur.com/Rk9XV |
| 17:06 | < Rhamphoryncus> | heh |
| 17:07 | < RichyB> | gitsvn isn't (AIUI) a pleasant way to get introduced to git. |
| 17:08 | < RichyB> | Saw a bloke run a Git-HOWTO lightning talk at 2011 PloneConf which was pretty good. He then gave a Git-SVN-HOWTO greased-lightning talk, which was way faster on account of being one word long. (Namely, "Don't.") |
| 17:10 | < Rhamphoryncus> | heh |
| 17:10 | <@froztbyte> | lulz |
| 17:10 | < RichyB> | The standard *pleasant* intro to Git as a committer these days is to try committing to someone else's project on Github. |
| 17:10 | | * Vornicus pokes vaguely at his piecewise function code. |
| 17:10 | < Rhamphoryncus> | I've had the same kind of results with hg and I believe at some point an ordinary git repository |
| 17:12 | < RichyB> | If you're getting stuck trying to use Hg then there's a serious pedagogical issue. Hg is not supposed to be difficult. oO |
| 17:13 | < Rhamphoryncus> | Not supposed to be |
| 17:14 | <~Vornicus> | jerith: I got de Boor working. Next is derivatives and then subdividing and then crazy toys like curvature. |
| 17:14 | < Rhamphoryncus> | Then I'll get an odd error message, be unable to get past it, and spend an hour on here getting 5 people to give suggestions trying to fix it |
| 17:16 | < RichyB> | Bizarre error messages absolutely quality as serious pedagogical issues. :/ |
| 17:16 | | RichyB [mycatverbs@Nightstar-3b2c2db2.bethere.co.uk] has left #code ["ERC Version 5.3 (IRC client for Emacs)"] |
| 17:16 | | RichyB [mycatverbs@Nightstar-3b2c2db2.bethere.co.uk] has joined #code |
| 17:17 | < RichyB> | Okay, I found the dumbest way to fuck up my IRC client yet. |
| 17:17 | < Rhamphoryncus> | Let's see if I can remember some recent hg issues.. the docs (manual and --help) mentioned a --rev option and that it took a revision number, but didn't say you could use a tag or whatever too |
| 17:17 | < RichyB> | ERC hates it if I turn on a different major mode in a channel buffer. :) |
| 17:18 | < Rhamphoryncus> | heh |
| 17:19 | < Rhamphoryncus> | And (might be related to hgsvn) there's no labelling of branches. They clearly exist: you can pull multiple branches into one repository, merge them, etc. They don't label though. "tip" is the only label it applies for me |
| 17:20 | < Rhamphoryncus> | So if I want to compare to upstream I have to fire up thg, browse to the right revision number, apply a local tag, then quit thg and do diff --rev tagname |
| 17:21 | < Rhamphoryncus> | It's silly. I can't believe that's how it's supposed to be done. So much so that I have a hard time not trying to find the "right" way |
| 17:22 | < Rhamphoryncus> | Sure, some times I have to accept the workaround, but in the past that includes multiple tarballs of git repositories to protect against it imploding (which it did) |
| 17:24 | < RichyB> | Hang on no no no. |
| 17:24 | < RichyB> | You should never be using git-svn or hgsvn. |
| 17:24 | < RichyB> | They're meant for people who have some kind of crippling addition to (git|hg) but also need to interact with an svn repo. |
| 17:24 | < Rhamphoryncus> | <RichyB> You should never be using git or hg. <--- ftfy. That's what the choice often means :P |
| 17:25 | < RichyB> | Well, if that's your constraint then yes. |
| 17:25 | < Rhamphoryncus> | I want to use version control. If I don't have commit access then using svn means I'm not using version control |
| 17:25 | < Rhamphoryncus> | Which I've done plenty. It's stupid. |
| 17:25 | < RichyB> | Yeah, lame situation. |
| 17:26 | < RichyB> | You're doomed from the very beginning if you're using something with a completely different data model than the upstream developers you're trying to work with. |
| 17:26 | < Rhamphoryncus> | yup |
| 17:27 | < RichyB> | Given your constraints, I'd either: |
| 17:28 | < RichyB> | a) use move files back and forth between $DVCS and svn manually, mostly using "svn export". This is a somewhat horrible way of working but at least you're using both tools with their own native interfaces. |
| 17:29 | < RichyB> | b) whine until I get commit rights and carry on using svn. |
| 17:30 | < Rhamphoryncus> | b isn't an option for me. I just won't do it. |
| 17:30 | < RichyB> | c) say "hell with those people" and either fork or go find another product with a more cooperative community. |
| 17:30 | < Rhamphoryncus> | c I'm pondering |
| 17:30 | < RichyB> | Technically "fork upstream" is a special case of "go find another product with a more cooperative community" where your means of finding a better community is starting one. ;) |
| 17:31 | < Rhamphoryncus> | I've submitted one cleanup patch to openttd before. One person made a couple dickish comments on the tracker, otherwise it's been ignored. |
| 17:32 | < Rhamphoryncus> | Sorry, by "before" I mean like a week ago. I'm working on another much bigger patch now |
| 17:33 | < Rhamphoryncus> | I consider the devs to be on trial, to determine if it's worth my time |
| 17:37 | | celticminstrel [celticminst@Nightstar-5d22ab1d.cable.rogers.com] has joined #code |
| 17:37 | < Rhamphoryncus> | This was after an IRC discussion where I said there was a lack of visible leadership and no way to expect consideration of any work you do. One of the people on IRC said he'd consider my patch (the main one, not the one I already sent) |
| 17:37 | | * RichyB just dropped a penny. |
| 17:38 | < RichyB> | Rhamphoryncus: if I did the work to export openttd's svn history into a github repo this evening would you be interested in working from there? |
| 17:39 | | * ToxicFrog upreads |
| 17:39 | <@ToxicFrog> | git-svn has a great many issues directly traceable to the fact that it has to work with svn, but I've never had issues that bad with it. |
| 17:39 | < Rhamphoryncus> | No, sorry. I already have it running through an official hgsvn repository and it hasn't bitten me too hard *this time* |
| 17:39 | <@ToxicFrog> | I've never used hg-svn, though. |
| 17:39 | < RichyB> | I'm not trying to be altruistic or anything; I want to try playing some of the patches you've been talking about the last couple days because they sound fun. ;) |
| 17:39 | < Rhamphoryncus> | ToxicFrog: I seem to be a magnet for it |
| 17:40 | <@ToxicFrog> | And it does map svn branches and tags to git branches and tags, and vice versa, properly |
| 17:40 | < RichyB> | Eh, okay, fair enough. |
| 17:40 | < Rhamphoryncus> | http://git.openttd.org |
| 17:40 | < RichyB> | ToxicFrog: presumes that the svn repository is properly managed, right? |
| 17:41 | < RichyB> | As in trunk, tags/* and branches/* directories in the right place. |
| 17:41 | <@ToxicFrog> | Rhamphoryncus: what the fuck is that |
| 17:41 | <@ToxicFrog> | It appears to be a collection of git repositories generated from fragments of the svn repo |
| 17:41 | <@ToxicFrog> | RichyB: yes. |
| 17:41 | < Rhamphoryncus> | Looks odd, yeah |
| 17:42 | < Rhamphoryncus> | http://vcs.openttd.org/git/ |
| 17:42 | < RichyB> | ToxicFrog: good, that reduces the "magic" quotient in my eyes. ;) |
| 17:43 | < Rhamphoryncus> | On further inspect I'd say it's lameness on their part. It's a link to the server root dir, not the repository |
| 17:43 | <@ToxicFrog> | RichyB: if the repo is organized in a different, but consistent, manner, there's configuration options you can set to adapt to that (eg, --tags=mytags, --master=therealtrunk, --branches=lookattheprettytree) |
| 17:43 | <@ToxicFrog> | If it's not then you're fucked. |
| 17:43 | <@jerith> | Vornicus: \o/ |
| 17:43 | < Rhamphoryncus> | hg otoh shows a list of branches for both hg links provided |
| 17:44 | <@ToxicFrog> | Rhamphoryncus: this is inexcusably terrible |
| 17:44 | <@ToxicFrog> | And if their hg interface is set up the same way no wonder you're having trouble |
| 17:44 | < RichyB> | ToxicFrog: oh the repository that I'm thinking of is totally not even slightly consistent. I'm fully aware that I'd have to do masses of work to it in order to convert it to anything. |
| 17:44 | < Rhamphoryncus> | No, most of my comments have been historical. I've done very little programming in the last year or two |
| 17:45 | <@ToxicFrog> | (unless hg has a fundamentally different approach to branches which I guess it might; I've never used it) |
| 17:46 | <@ToxicFrog> | But yeah, the way that's meant to map to git - and the way git-svn does it - is that you get a single 'openttd.git' repo, and all of those branches/ and tags/ repos are branches and tags within it, and trunk/ is master. |
| 17:47 | | Kindamoody|out is now known as Kindamoody |
| 18:10 | | * Vornicus fiddles with pygame briefly. |
| 18:21 | | * Vornicus also finds himself considering openttd once again, and the thing he wanted to try doing with smoothing, has determined that he had originally done it wrong. |
| 18:26 | <~Vornicus> | (I was taking pieces and fitting a biquadratic; but it did not maintain certain necessary invariants, namely the location of the centers of edges.) |
| 18:27 | | RichyB [mycatverbs@Nightstar-3b2c2db2.bethere.co.uk] has quit [Ping timeout: 121 seconds] |
| 18:51 | < Rhamphoryncus> | Aaaaaaugh. Off by 1 has got me :P |
| 18:53 | < Rhamphoryncus> | Well, it may be harmless, but it's still a 1 sneaking in there. Specifically, openttd updates the vehicles one at a time, so you'll see one of them being 1 tick further ahead of the other |
| 18:54 | < Rhamphoryncus> | And.. it's not harmless. It's possible for that to accumulate |
| 19:00 | < maoranma> | Like phlegm? |
| 19:04 | < Rhamphoryncus> | ... yes, like phlegm |
| 19:04 | | Stalker [Z@Nightstar-5aa18eaf.balk.dk] has quit [Ping timeout: 121 seconds] |
| 19:14 | < Rhamphoryncus> | Yup, 668%74 == 2. It's accumulating |
| 19:23 | < Rhamphoryncus> | Now I won't say it works perfectly now (other than a couple known issues like the time phlegm), but I will say I'm between major bug discoveries ;) |
| 19:34 | | * Tamber giggles at "time phlegm" |
| 19:37 | < Rhamphoryncus> | It's basically like bus drivers swapping, it's 6 pm, they're supposed to arrive at 7 pm, one looks at his watch and sees 6:01 pm so he says "Got to be there in 59 minutes", the other looks at his watch and sees 5:59 pm so he thinks "Okay, 6:58 pm" |
| 19:38 | <@Tamber> | o.0 |
| 19:39 | < Rhamphoryncus> | That really is what it's doing :/ |
| 19:40 | < Rhamphoryncus> | There is a reason for this: time is entirely relative. There's no worry of overflowing the year, or what happens when the month rolls over, or whatever |
| 19:40 | < Rhamphoryncus> | And you can move the game forward or back, trivially |
| 19:41 | < Rhamphoryncus> | Of course you could just start the game at 0 and only translate that in to a year for the user :P |
| 19:44 | | celticminstrel [celticminst@Nightstar-5d22ab1d.cable.rogers.com] has quit [[NS] Quit: Kablammo! Celticminstrel is now unavailable for comment.] |
| 19:47 | | Stalker [Z@Nightstar-3602cf5a.cust.comxnet.dk] has joined #code |
| 19:55 | <@ToxicFrog> | Snrk |
| 19:55 | <@ToxicFrog> | "99 bugs in the legacy code, 99 bugs in the code. Take one down, pass it around. 102 bugs in the legacy code." |
| 19:55 | <@jerith> | Quite. |
| 19:56 | < Rhamphoryncus> | heh |
| 19:57 | < Rhamphoryncus> | Code is still broken. Methinks it's time to write a sanity check function that audits the linked list to be sure it's in the right order |
| 20:33 | | Kindamoody is now known as Kindamoody[zZz] |
| 23:16 | < Rhamphoryncus> | Now I'm pondering implementing a linked list sort. I may do a half-assed quicksort ;) |
| 23:48 | < maoranma> | while list != sorted: |
| 23:48 | < maoranma> | randomize(list) |
| 23:48 | < maoranma> | if list != sorted: endUniverse() |
| 23:48 | < maoranma> | Works every time |
| --- Log closed Sat Feb 04 00:00:40 2012 |