[02:05:06] ** bear_ has joined us [02:05:27] ** bear has left IRC (Read error: 110 (Connection timed out)) [02:28:28] bear_ is now known as bear [04:02:28] [connected at Sat Mar 12 04:02:28 2005] [04:02:28] <> *** Looking up your hostname... [04:02:28] <> *** Checking ident [04:02:29] <> *** Found your hostname [04:02:59] <> *** No identd (auth) response [04:02:59] <> *** Your host is kornbluth.freenode.net[kornbluth.freenode.net/6667], running version dancer-ircd-1.0.35 [04:02:59] [I have joined #peak] [04:02:59] ** kornbluth.freenode.net set the topic to http://dirtsimple.org/2004/11/generic-functions-have-landed.html [04:11:13] ** vlado has joined us [05:55:01] vlado is now known as vlado|lunch [06:46:11] ** bear has left IRC (Read error: 60 (Operation timed out)) [07:11:26] vlado|lunch is now known as vlado [07:12:52] ** debugger has joined us [07:13:12] hi [07:17:50] hello [11:35:22] ** vlado has left IRC ("Leaving") [11:55:51] ** bear has joined us [17:19:54] ** arg has joined us [17:20:02] hola [17:45:23] etrepum : got any feature requests for a new blogging app? [17:45:58] arg: the list you came up with is pretty much what I want [17:49:21] yeah [17:49:27] writing the web frontend is a PITA [17:49:36] marsedit works fine tho :) [17:49:52] im working on colorizing python code and comments now [17:50:08] I don't really care about a web frontend.. I use MarsEdit too [17:50:13] its my first nevow project tho, so i may be doing some things naively [17:50:40] ill setup svn soon, maybe you can check it out [17:53:40] cool [18:26:40] what about TypeKey for auth [18:28:49] ahhh n/m... its not what i thought it was [20:53:51] ** debugger has left IRC () [22:49:55] * arg considers ditching sqlobject for PEAK [22:50:14] SQLObject kinda scares me [22:50:32] zope.interface can't declare a module or object as providing an interface [22:50:38] it's nice for prototyping [22:50:52] but i may get rid of it eventually [22:51:05] im more concerned with getting something working first [22:51:18] zope.interface sucks compared to pyprotocols, yeah. [22:51:22] i figure by tomorrow evening ill can check in something useful [22:51:45] yeah, peak.model is very nice too [22:52:00] i'd like to help pje out too by releasing something that uses peak [22:53:02] he wanted the source to my shitty bot framework that used dispatch but its too shitty [22:53:15] I've used it for proprietary stuff, nothing OSS [22:53:29] yeah i have too [22:53:44] for work i've used it in 3 or 4 things we've sold [22:54:09] i cant even talk about them tho, nonetheless release them [22:54:38] i asked if i could give a little endorsement quote for twisted 2.0 and got denied [22:54:38] yea [22:54:45] that sucks [22:55:04] lawyers [22:55:42] SQLObject classes look like peak.model clases, but they do all the SQL stuff without a DM [22:56:01] you dont get as much control tho, or the richness of peak,.model [22:56:23] yeah [22:56:26] maybe i'll switch over when peak.schema materializes [22:56:34] pje does a really good control about giving you control [22:56:43] but also making it easy when you don't need it [22:56:52] man, peak was like an enlightenment experience [22:57:00] about framework architecture/design [22:57:05] yeah [22:57:46] it's definitely influenced how I think about writing frameworks, but a lot of the crap I do depends on poorly designed things like distutils so I can't do everything in that style :) [22:57:47] when you use the whole thing... peak.naming and a domain model with storaqge and .model.. [22:57:58] peak.events is blindingly elegant too [22:58:09] transparent twisted integration [22:58:31] yeah, peak rules :) [22:58:47] you're going to pycon right? [22:58:59] I hope so, I've got two talks :) [22:59:06] oh yeah thats right hehe [22:59:11] are you going for the sprints? [22:59:17] yeah I'll be there [22:59:42] will probably sprint on PyPy unless something more interesting comes up [22:59:46] i'm trying to get akamai to pay for me to extend my trip [23:00:24] i'd like to go for the sprints [23:00:29] ill be there for the conference tho [23:00:46] usually the talks are the least useful part of PyCon, it's talking to people and shit, which you get a lot more of during the sprints [23:00:50] yeah [23:01:02] ill prolly just pay for it myself if they wont foot the bill [23:01:21] cause if you just want the talks, you can get the notes and slides afterwards and not even go [23:01:32] yeah i know [23:01:38] i mostly want to meet people [23:01:48] also maybe try to hire people [23:01:59] last year there was a group of people doing SubEthaEdit notes during all of the talks, it was pretty cool [23:02:18] i've never actually used it in collaboritive mode [23:02:41] if you pop it open during PyCon you'll be able to see notes in real time for talks you're not even at, it's pretty slick [23:02:53] wow, that rules [23:02:59] im going to try to do a talk next year [23:03:05] about the stuff we've done with python aaat akamai [23:03:14] "at", even [23:03:30] speaking of that I should probably integrate that into a PyObjC demo, a Rendezvous browser or something and poke at some of the data coming through SEE [23:03:34] and subverting and overturning the Perl culture [23:04:03] pretty good success story, if the lawyers will let me tell it [23:04:07] I used to use Perl, years ago.. it's brain damage, like using BASIC [23:04:11] yeah, i want to write a cocoa/PB client for Minnow [23:04:20] with pyobjc [23:04:37] i went through a few of the examples, its amazing how well the two object systems integrate [23:04:41] yeah ReST integration would be hot [23:04:52] well, ytou could start with ReSTEdit [23:05:00] yeah fortunately Objective-C and Python have very similar semantics [23:05:09] slap some xmlrpc mumbojumbo onto that [23:05:49] ReSTedit definitely needs a couple more features, like block indent/dedent [23:06:10] I might do that in the next couple days so I can use it to write my talks [23:06:10] yeah, i use it a lot at work for important emails... it impresses people [23:06:48] right now I just use Vim, and keep ReSTedit open for live preview (it uses kqueue notifications) [23:06:58] as soon as I save, it rerenders [23:06:58] oh, cool [23:07:05] yeah it's pretty slick [23:07:09] im a recent mac convert [23:07:16] well, 2 years or so [23:07:23] pje needs to Switch [23:07:31] I converted when 10.1 came out [23:08:41] theres no going back [23:09:29] yeah, for a developer, I don't see why you would want to use anything else [23:10:13] well, some of the .NET dev tools are pretty cool, but not enough to switch for [23:10:26] mono works [23:11:08] someone needs to make an easy to install way to get Mono on OS X [23:11:15] there is one [23:11:23] I guess I haven't seen it [23:11:24] its pretty new [23:11:31] and hidden at the bottom of the page [23:11:43] getting gtk# and monodevelop to compile is another story [23:11:57] I can imagine [23:11:57] but theres an installer for the libraries and compiler [23:12:43] brb, cats harassing me for food [23:19:15] i think i'll use the blogger style yyyy/mm/dd/the-post-title.html naming [23:19:33] well, actually, i have INamingStyle with nameForPost and postForName [23:19:33] sounds good to me [23:20:04] wordpress does the same thing s/\.html/\// [23:20:22] im conflicted about how to deal with static rendering [23:20:41] or whether to put that off for later [23:20:47] well the way that some things do it, is that they take the things that need to be dynamic, like comments, and shove that into javascript [23:21:32] right now my blog is slow cause it uses wordpress.. I made it use cached reST fragments though [23:21:57] basically I take the md5 digest of the input text, and make a temp file with that name.. if it exists, it's used as the html, otherwise it rerenders [23:22:09] it's ghetto, but so is php [23:22:17] this is pretty fast with sqlite [23:23:00] the good thing about static rendering is that you don't need to care about setting proper cache headers and all that crap [23:23:10] it's also easier to host something statically rendered [23:23:18] i might just release a 0.1 as is once i get comments and tags and the rest of the usual stuff [23:23:23] so others can look at it [23:23:29] yeah it definitely is [23:23:34] plus if you get slashdotted [23:23:39] yeah [23:24:02] theoretically, it shouldn;t be hard, but it has subtle implications on naming and stuff [23:24:25] and on updates.. like if you change your template, do you expire the whole site? [23:24:54] i was considering a two-stage xslt transform [23:25:03] but thats too much [23:25:13] i guess you can cache parts of those [23:25:15] yeah, xslt in general is probably a bit much [23:27:02] wingware needs to stop doing their own web pages.. they're hideous and make no sense [23:27:23] hahaha i know [23:27:50] they look like circa-1995 hobbyist web pages [23:27:55] the photos are hilarious, i bet an intern took them [23:28:07] those guys are right by me, in Brookline [23:28:17] i've been meaning to go to one of their python meetups [23:28:35] I'm in NYC right now, but I'm probably relocating to SF this year [23:28:59] wow, i've always wanted to go to SF... almost did once [23:29:08] for a different job? [23:29:24] well right now I work for myself [23:29:32] part owner of a company and I do consulting on the side [23:29:51] but I'm starting another company and my business partner and the investors are in SF [23:31:52] I almost moved out there last year anyway, and I have some friends in SF, so I might as well [23:32:14] cool [23:32:29] my company has an office in san mateo [23:32:41] so im constantly flirting with the idea of going out there [23:32:51] but most of the stuff im into is happening here in cambridgre [23:32:53] I'll be able to file radars in person :) [23:32:58] haha [23:33:40] im going to stick it out with these guys for a while... im sure in a few years when people are fully vested there will be a lot of startup opportunities [23:34:06] Where are you at? [23:34:10] Akamai [23:34:23] cool [23:34:29] have you heard of them? [23:34:39] yeah [23:34:50] who hasn't? [23:34:56] haha, i dunno, you'd be surprised [23:35:18] if you pay attention to the status bar of your browser you have :) [23:35:41] apple does a bunch of business with them [23:35:45] yep [23:36:00] we do the storage backend and xml serving for ITMS [23:36:12] among other things [23:36:15] yeah [23:36:23] we usually stream their keynotes [23:36:30] and i think we do the software updates [23:36:46] and all the movie trailers [23:36:50] yep [23:37:07] pretty crazy place [23:37:14] robbed me of my youth :) [23:37:27] but they got me an ipod for my 5 year anniversary [23:37:27] I can imagine [23:37:40] I haven't worked anywhere that long [23:37:47] so you're not at masters of branding anymore [23:37:52] i hadnt before this [23:38:15] I still consult for them, but I left there over a year ago [23:38:22] i see [23:38:49] doing RFID work with giant megacorps is a good way to hate life [23:38:54] haha [23:39:10] I made great money but I just didn't want to go to work in the morning [23:39:15] yeah, lucky for me i got out of the stressful stuff, working on the production network [23:39:25] where any fuckup is amplified x 15000 or so machines [23:39:35] im in the custom eng department now [23:39:37] haha [23:39:40] doing one-offs and prototypes [23:39:40] that sounds good [23:39:46] that's perfect [23:39:49] yep [23:40:03] carte blance wrt. langauge, within reason [23:40:09] i.e. not ocaml or anything [23:40:10] the fun part is definitely not production [23:40:12] but python is a-ok [23:40:16] event obj-c [23:40:17] yeah [23:40:28] err [23:40:32] "even" obj-c [23:40:50] some mac fanatic did all the work to get the libaries to build in our build system [23:40:59] haha [23:41:13] all sorts of hacks to sneak the compiler in as a data resource [23:41:18] Obj-C is a very sane language [23:41:32] it really is [23:42:32] it doesn't have the performance characteristics of C++, but you mostly make up for that with quick development/debugging cycle, and it doesn't take a week to compile [23:43:04] yeah, the looser typing and nice degradation to Object makes things much faster [23:43:08] i cant write c++ [23:43:13] i have to write it in python first [23:43:18] and translate it [23:43:35] any language without print vars() or equivalent is broken :) [23:43:55] BOOST makes some c++ palatable but only if you have a shitload of RAM and a very fast computer [23:47:30] yeah [23:47:40] the guy who sits next to me is a really good c++ hacker [23:47:48] who uses boost extensively [23:47:55] I think you pretty much have to [23:47:57] boost.spirit is really cool [23:48:22] i showed him the *infix* operator hack and challenged him to do it in c++ [23:48:23] macstl also really intrigues me, but I don't do enough C++ to have needed it yet [23:50:02] david abrahams actually inspired donovan and I to come up with nevow's stan [23:50:12] at pycon 2003 [23:50:14] http://rafb.net/paste/results/S3voRe63.html [23:50:19] theres the infix hack [23:50:22] stan is awesome [23:50:38] yeah, even more awesome in 2.4 with genexps [23:51:01] yeah [23:51:20] i just took out all the genexps in Minnow because I should prolly be 2.3 compatible [23:51:27] my only regret about stan is choosing to do class_ instead of _class [23:51:37] for kwargs syntax [23:51:53] I hadn't done the research to know that _class is more common [23:52:03] since I came up with it in like half an hour in a hotel room [23:52:09] without interweb [23:52:16] and it stuck [23:52:18] hehe, thats a pretty minor wart tho [23:52:22] yeah it is [23:52:49] using object adaptation to build a web page was the best idea ever [23:52:57] and the way nevow can deal with deferred execution is great [23:53:17] compiling the parts of the templates that it knows, so they're optimized, and do the dynamic parts at runtime [23:53:26] yeah, im only 2 days into learning nevow, so im sure you can clean some things up once I post this code [23:53:36] i have to figure out how to nest templates [23:53:45] like, having a template for just the post [23:53:53] and an overall layout template [23:53:59] and the sideboxes, etc [23:54:08] you can probably just put the object in there and it will render [23:54:15] yeah [23:54:25] a lot of stuff in twisted/nevow works that way [23:54:48] yeah [23:55:12] im actually using xmlfile so far figuring people would be more comfortable extending/modifying that [23:55:22] but im wondering if thats worth it [23:56:02] it's probably worth it for the "outside" template [23:56:11] css is probably enough to customize the "inside" templates [23:56:20] true [23:56:59] stan is great for prototyping though, so if it's more work to do xmlfile, you can always put that off [23:57:28] you can always get an xmlfile with a little hackery from a stan template, then give folks the option [23:57:46] true