00:00:00 --- log: started haskell/02.02.02 01:06:58 --- quit: lesner (Read error: 104 (Connection reset by peer)) 03:24:17 --- join: shapr (~user@p-c2fbab34.easy.inet.fi) joined #haskell 04:15:44 --- join: atun (~cogito@217-125-226-125.uc.nombres.ttd.es) joined #haskell 04:17:23 --- quit: atun (Read error: 104 (Connection reset by peer)) 05:26:30 well crap 05:26:36 that bites ass. 05:26:45 what now? 05:26:48 keep missing people? 05:26:51 as far as I can tell, there are no remaining copies of the HaskellWikiInHaskell *anywhere* 05:26:54 heh, no :) 05:26:55 sheesh, you complain a lot. ;) 05:26:59 * shapr grins 05:27:03 yah I probably do 05:27:42 what's up in your life yusri? 05:27:55 little 05:28:01 Going to waste some money today 05:28:10 I'm going to shool for a few weeks here 05:28:20 devoid of any IT applicances 05:28:25 * yusri shivers in fear 05:28:48 so I'm going to buy a PS2 today to waste my time while away, keep my mind of all the porgramming I'm missing 05:29:09 what kind of school is devoid of IT appliances? 05:29:44 Managerial/supervisory training 05:29:53 My laptop died :( 05:30:12 suckage 05:30:22 (the laptop death, not the school :) 05:30:30 It's one of those fairly general, conceptual, think-outside-the-box kind of classes 05:30:36 so the school too 05:30:47 Though it's necessary for my pay raise 05:31:35 what do you do for a living? 05:32:48 general linguistic translation 05:33:09 usually treaties and goverment/coorporation contracts 05:33:15 wow, that's hardwork 05:33:19 *very* hard work 05:33:29 very tedious, very exacting 05:33:44 I hope it pays very well 05:33:49 obviously wording and nuance are important :) 05:33:55 yah, really 05:34:13 It pays enough, I guess. Though I wouldn't complain if they doubled it ;) 05:34:23 heh 05:34:40 my significant other does translations of scientific documents sometimes 05:34:44 (among other things) 05:34:56 and I get to see just how difficult it really is 05:35:02 Ah- what lang? 05:35:41 she does german, french, english, and some dutch translations 05:35:54 and we're both learning finnish (though *very* slowly) 05:36:15 Ah- I know very little norwegian, which I suppose is fairly similar (?) 05:36:30 I don't think so 05:36:42 with the disclaimer that I am not a linguist :) 05:36:53 * yusri is; will ahev to look ;) 05:36:56 I think that most of the scanned in avian languages are closely related except for finnish 05:37:13 oh, shapr, I got ghc with hopengl to compile. No idea if it works, though :-) 05:37:13 norwegian, swedish, and danish sound pretty similar, 05:37:19 Heffalump: awesome! 05:37:28 only problem, it's nearly 200MB worth of debs 05:38:07 Ah, i recall- finnish is an oddity 05:38:14 yusri: those last three sound a lot like a strange dutch and german mixture 05:38:22 * yusri nods 05:38:42 one third dutch, one fifth german, one fifth english, a few french words, and some other random stuff.. 05:39:04 * yusri nods 05:39:19 languages like that make me yearn for my [relatively] pure arabic. 05:40:11 finnish is very pure in that respect 05:40:24 estonian is the closest thing to finnish, and it's not very close 05:40:33 * yusri looks at the time and realizes he needs to sleep 05:40:39 g'night yusri 05:40:48 sorry to cut you off- these ARE my favourite- but I need to crash 05:41:05 no worries :) 05:41:26 talk to ya tomorrow 05:42:26 prob just a nap- too much to do 05:42:50 heh 05:45:12 * xbill hacks the kernel. 05:45:25 * shapr hacks his locale 05:54:31 --- join: Yurik (~yrashk@gw.telcos.net.ua) joined #haskell 05:54:38 hi 05:54:44 shapr are you here? 05:55:39 guys, can anybody say where can I read any good tutorial on ASN.1 and BER/PER? 05:56:51 hi Yurik 05:56:55 are you serious? 05:57:00 why do you want to know about ASN.1? 05:57:09 it's most evil technology I've ever had to learn! 05:57:16 but, yes I can point you to lots of info on it 05:57:27 well, just read some introduction on it (in Erlang/OTP docs and some webpages) 05:57:32 I also know that erlang and Haskell have excellent ASN.1 tools 05:57:35 and found that it is quite interesting in E/AS 05:57:41 really? how so? 05:58:26 Using ASN.1 usage seems to be very nice in E/AS to support non-native method implementors to exchange data.. 05:58:40 I think *anything* would be better than ASN.1 05:58:52 ? 05:59:08 I've been working with ASN.1 for the last six months or so 05:59:24 I'm writing an X.509 PKI timestamp client in Java 05:59:28 rfc3161 05:59:46 and, in my humble opinion, ASN.1 is *evil* 05:59:58 I'd recommend XML far above it. 06:00:42 er, let me give some detailed examples: 06:00:56 there are about eight basic datatypes in ASN.1 06:01:12 but one type is called ANY 06:01:22 and it can be a any other ASN.1 datatype 06:01:48 but in 99% of the cases I've seen, it's actually an OCTET STRING (just a bunch of bytes) 06:02:15 still sounds ok, until you realize that most people embed another ASN.1 structure in there 06:02:27 so... 06:02:47 hmm. isn't ASN.1 widely used? 06:02:53 yes it is 06:03:04 and there's no central repository for ASN.1 definitions 06:03:14 nor for OID definitions 06:03:27 so even in framework of one project you propose to use XML? 06:03:31 and there are at least three different syntaxes (1988, 1991, 1994, and I think maybe a 1997) 06:03:42 I'd say "anything other than ASN.1" 06:03:48 heh 06:04:08 I could go into extreme detail about why I think ASN.1 is the most evil technology I've had to deal with... 06:04:11 do you want more detail? 06:04:32 yep, I want. I want to make a right decision 06:04:37 ok 06:04:49 just a moment, let me get some urls together 06:04:58 first, the layman's guide to a subset of ASN.1 06:05:24 http://www.rsasecurity.com/rsalabs/pkcs/ 06:05:39 near the bottom of that page, you'll see the layman's guide. 06:06:07 I see 06:06:34 I have an asn.1 response that's been sent back from a timestamp server, want me to send you the file? 06:06:35 * Yurik is downloading layman.ps.gz 06:06:47 http://www.cs.auckland.ac.nz/~pgut001/ 06:06:51 that's peter gutmann's homepage 06:07:00 he has a fascinating view of the X.509 standards 06:07:31 if you need to work with ASN.1 you *will* need his dumpasn1.c code 06:07:34 as well as his list of Object Identifiers 06:08:08 there are two books about ASN.1 that are available from ... oss.com I think... 06:08:11 and they're free 06:08:17 and I've read both of the books 06:08:22 and I've read the layman's guide 06:08:31 and I *still* have daily problem with encoding and decoding ASN.1 06:08:33 not only that... 06:08:47 but there around 30 different encodings for ASN.1 that I'm aware of 06:08:53 and at least three major encodings 06:08:59 those being BER, PER, and DER 06:09:15 --- quit: dennisb (carter.openprojects.net irc.openprojects.net) 06:09:25 there's even XER, the XML encoding, but that's used so that you can encode something, and the actually look at it in a text editor 06:09:40 "then actually" 06:09:45 well, seems that I understand why ASN.1 can be an evil technology.. 06:10:02 yes, especially because the standards don't cover everything themselves 06:10:10 some asn.1 parsers allow empty elements 06:10:16 --- join: dennisb (~dennis@as2-4-3.an.g.bonet.se) joined #haskell 06:10:16 some require a NULL in empty elements 06:10:27 some disallow both of those 06:11:09 you will need a piece of software to automatically compile your ASN.1 specification file into a chunk of code that can turn it into an ASN.1 representation 06:11:22 then, you'll need at least one encoder to encode that representation 06:11:40 I require two encoders/decoders at work, for DER and for BER 06:11:50 well, as I see, erlang has ASN compiler.. 06:12:12 oss.com sells their ASN.1 compiler for USD $25,000 06:12:19 it comes in C++ and Java flavors 06:13:02 Using ASN.1 usage seems to be very nice in E/AS to support non-native method implementors to exchange data.. 06:13:30 ASN.1 is not a good way to exchange data with another platform/language/host 06:13:35 that's my solid opinion 06:13:41 well, probably you're right :-) 06:14:01 so, what will be better? XML (I'm very pessimistic on wide XML usage) 06:14:03 I've spent three weeks straight trying to debug a very simple ASN.1 encoding problem 06:14:03 ? 06:14:22 one of the great strengths of XML is that I can read its contents even with Notepad 06:14:35 trying to debug ASN.1 is a complete and total nightmare 06:15:05 my overall opinion of cross-anything encoding is leaning towards lisp-style SEXPs lately 06:15:14 XML is second on my list 06:15:47 SEXP? what is this? 06:15:48 want me to explain why? 06:15:58 Lisp source code is most often written in Cambridge Polish notation: from a set of atomic symbols, you build Symbolic EXPressions (S-exp or SEXP -- once was SEX), a SEXP being either atomic, or a list of smaller SEXP, read and written as the parenthesized space-separated sequence of the representations of its components 06:16:10 that's a short quote from this page: http://tunes.org/Review/Languages.html 06:16:23 if you jump down to the LISP section, they have some interesting stuff 06:16:56 * shapr takes a moment to compose his thoughts on the subject of data interchange 06:18:52 written as the parenthesized space-separated 06:18:52 sequence of the representations of its components 06:18:52 that's a short quote from this page: 06:19:03 sorry, cut-n-paste 06:19:04 :-) 06:21:15 no worries 06:21:18 ok, I've been thinking about it 06:21:28 I read a funny quote somewher 06:21:47 it said "some has a problem, they need to match some text, they decide to use regular expressions... now they have two problems" 06:22:07 what it means to me is that regular expressions are powerful, but can easily bite you if you don't know what you're doing with them 06:22:32 that's why I'm the only person I work with who uses regular expressions, no one else understands them well enough to use them 06:23:07 with ASN.1, it can be useful, but it's so much more complex than regular expressions that it has fifty-nine ways to bite you for ever one way it could be useful 06:23:14 er, "every one" 06:25:17 heh :-) 06:25:38 so... 06:25:44 I'd say that ASN.1 isn't useless 06:26:09 I understand your point of view concerning ASN.1 :-) 06:26:11 but it's so complicated that you should write your own ASN.1 parser before you use it, because otherwise it will suck away your sanity. 06:27:09 if you still decide to use it, I'll help you howevery I can, but I would strongly recommend against using it. 06:28:31 probably I will not use ASN.1. I do not like such kinda techs... so complicated and can really make a kinda harm ( :)) )... 06:32:07 --- quit: shapr (Read error: 104 (Connection reset by peer)) 07:13:53 --- join: shapr (~user@p-c2fbab34.easy.inet.fi) joined #haskell 07:13:54 --- quit: Yurik (Read error: 104 (Connection reset by peer)) 07:54:17 --- join: tmoertel (~chatzilla@pa-mtlebanon2a-268.pit.adelphia.net) joined #haskell 07:54:47 hi tmoertel 07:54:50 what's up? 07:55:03 your Haskell, Perl, and XSLT article is nifty 07:55:08 howdy. just stoppin' by to see what's going on 07:55:26 thanks. hope it wasn't seen as overly harsh. 07:55:38 I liked it 07:55:47 I've given the url to a bunch of people to read 07:55:47 (I *like* XSLT, but its weaknesses drive me crazy.) 07:55:51 yah, I understand 07:56:34 I just figured out how to generate a ssh2 public key 07:56:38 now I gotta see if it works 07:57:09 tmoertel: written any cool code lately? 07:57:33 * tmoertel got a phone call . . . 08:04:32 * tmoertel is back 08:04:47 still there? 08:05:06 yah sure 08:05:07 what's up? 08:06:01 Let's see, Haskell-wise, I've been doing a few of John Hughes' courses online. 08:06:55 The "Designing and Using Combinators: The Essence of Functional Programming" course is wonderful. 08:07:12 I suspect it would still be a bit over my head... but maybe not... 08:07:16 His paper on Global Variables in Haskell is also a great read. 08:08:25 do you do much programming outside of Haskell? 08:08:36 I can guess you do since you deal with XSLT and Perl at least 08:08:59 Yes. Mostly Perl, C, C++, bash, SGML/XML-related. 08:09:23 Perl and Haskell are my favorites. 08:10:01 How 'bout yourself? 08:10:10 Python and Haskell are currently my favorites 08:10:22 but I'm having lots of fun with Joy lately 08:10:34 I do Java at one of my jobs also 08:10:38 I'm not familiar w/ Joy. What's it like? 08:10:53 it's a concatenative stack based functional lang 08:11:06 kind of like PostScript/Forth meets Haskell, and they have kids 08:11:06 Postfix? 08:11:09 yah, very 08:11:27 Is it used for embedded stuff, like Forth? 08:11:29 I didn't know postfix/stack languages existed before I found Joy, it's fun! 08:11:47 One of my favorites is the Unix 'dc' program! 08:11:50 heh 08:12:04 yah, the concatenative group talks about that occasionally 08:12:25 Can you give me a line or two of Joy code to give me a feel? 08:12:34 sure, just a moment 08:13:04 jewel and I are working on a connected graph path finder in Joy 08:13:14 DEFINE pairs == [["ward" "shae"] ["shae" "irene"] ["shae" "al"] ["al" "jewel"] ["irene" "lynda"] ["lynda" "jewel"]] . 08:13:19 DEFINE buildFilter == [] cons [first] swap concat [=] concat . 08:13:26 DEFINE collectEdges == pairs swap buildFilter filter . 08:13:49 Postfix-riffic! 08:13:54 totally 08:14:00 Stack-mazing! 08:14:10 absolutely! 08:14:21 it's great fun 08:14:30 Are you on the pragprog list? 08:14:33 sure! 08:14:52 Did you work on the circuit exercise from one of Prof. Hughes' lectures? 08:14:59 nope, not yet 08:15:31 I just did it last night. Very fun. 08:16:57 I keep getting distracted into doing fun stuff 08:17:04 well, other fun stuff :) 08:17:05 It happens. ;-) 08:17:09 I really enjoy using Haskell 08:17:22 I'd like to be able to recommend it to my bosses 08:17:35 Every time I do work in mainstream languages, I miss foldr. 08:17:44 Woe to me without foldr! 08:17:59 happily for me, Python has map, filter, and reduce (foldr) 08:18:07 that's what got me into FP 08:18:43 I realized how amazingly useful those commands are 08:19:07 Foldr doesn't really shine w/o lazy eval. 08:19:15 Then, it's a thing of beauty. 08:19:17 I have a feeling that the referential transparency requirement of FPLs ends up creating far more orthogonality in the languages 08:19:32 well, more than procedural languages 08:19:40 True, true. 08:19:56 yesterday I hypothesized that higher order functions in FP work like metaclasses in OOP 08:20:00 what do you think about that? 08:20:27 Interesting hypothesis. 08:20:59 btw, there's also the #loty channel here 08:21:12 I checked it last night, but it was dead. Active now? 08:21:23 not so much the last week or two 08:21:41 --- join: Yurik (~yrashk@gw.telcos.net.ua) joined #haskell 08:21:45 hi Yurik! 08:21:52 did you decide anything about ASN.1? 08:21:58 shapr hi 08:22:05 seems that decided not to use it :-) 08:22:09 * shapr cheers 08:22:22 hehe 08:22:29 tmoertel: hey, have you heard of the sf.net/projects/haskell-libs/ project? 08:22:46 I just checked it out last night. The idea is to collect snippets, right? 08:23:03 that's part of it 08:23:18 another part of it is to write up some of the standard libraries that are missing 08:23:26 such as smtplib, imaplib, poplib 08:23:53 I've got some lazy base64 and quoted-printable stuff I could contribute. 08:23:58 coool! 08:24:06 that would rock :) 08:24:18 * Yurik has downloaded USSR Anthem and got merry modd :o) 08:24:27 s/modd/mood/ 08:25:50 tmoertel: can you think of any haskell libs you'd use if they existed? 08:25:56 sounds from childhood.. yeah :-) 08:26:36 Yes: we need more robust XML support, easy HTTP (ala Perl's LWP::Simple), ... 08:27:19 modules for common text-processing shells... 08:27:22 hm, you think HaXML needs improving? 08:27:29 what kind of modules? 08:27:42 stuff like sort,uniq,tr,comm ? 08:27:54 I don't know perl, what does LWP::Simple do? 08:27:58 wget? 08:28:16 Yes. HaXML's support for % entities in DTD -> Haskell type leads to undesireble type duplication. 08:28:53 *Example* perl -MLWP::Simple -e'getprint("http://www.moertel.com/~thor/icfp2001/")' 08:29:10 I don't know anything about that, do you have a small snippet of code that shows the duplication of % entities? 08:29:46 sounds like the wget binary, urllib in python 08:29:58 or httplib 08:30:37 Many DTDs use % entities to define common classes of content: 08:30:47 08:31:07 And then those entities are referenced throughout the DTD: 08:31:17 08:31:49 08:32:16 But when HaXML constructs a type for Para and Intro, it will expand %inline into the Br Emph Strong Link Img like so: 08:32:39 Para_Br Para_Emph Para_Strong Para_Link Para_Img *and* 08:33:00 Intro_Br Intro_Emph Intro_Strong Intro_Link Intro_Img, and so on 08:33:02 hm 08:33:53 what would be better? 08:33:58 So you end up writing wacky handlers to get all the Br's, Emph's, ... 08:34:22 It would be better i f you could tell HaXML not to expand certain entities. 08:35:12 So you could create a "meta element" Inline that contained Inline_Br, Inline_emph, etc.and Para and Intro could each include Inline 08:35:30 rather than the expansion of Inline into all its components. 08:35:49 that makes sense to me 08:36:03 So, maybe we could do some work on HaXML, too. 08:37:16 sounds like it :) 08:37:42 since you said 'we' would you like for me to add you to the list of developers on haskell-libs ? 08:38:18 Sure. 08:38:25 ok 08:39:25 same name? tmoertel? 08:40:12 Yup. 08:45:17 I see that the current tasks are smtplib and imaplib. 08:47:46 welcome to the haskell-libs project! 08:48:04 oh, do you want to be an admin? 08:48:04 Yup. I can see that I'm on the list hoiw. Thanks. 08:48:38 No, that's okay. 08:48:45 yah, imaplib and smtplib were two projects that seemed needed 08:49:01 I actually think that poplib and smtplib would be best 08:49:29 or at least, most likely to be used by a lot of people 08:49:48 but I don't know the needs of the entire community by any means 08:50:06 ah, sandwiches are good things.... sometimes obsessive means I forget to eat 08:50:27 ahem, "obsessive coding" 08:50:29 whoops 08:50:34 So is espresso. Nothing cuts through "code fog" like it. 08:50:51 * shapr slurps on a cup of coffee 08:51:08 btw, what part of the world do you live in? 08:51:17 * tmoertel laments his empty espresso cup. 08:51:31 Glorious Pittsburgh. 08:51:33 You? 08:51:47 well, I'd email you some of my coffee, but I'm afraid what it might do to the computer 08:52:02 I'm in snowy Tornio, Finland 08:52:09 where sweden, finland, and the gulf of bothnia touch 08:52:29 I hear folks drink a lot of coffee in that part of the world. True? 08:52:46 er, I'd say more alcohol than coffee actually 08:53:03 tmoertel programmers drink a lot of coffee. in every part of the world :-) 08:53:04 Hmm.. Probably isn't so good for the code. ;-) 08:53:16 understandably so, the winters near the arctic circle get depressing 08:53:26 True. 08:53:32 I'm originally from Birmingham, Alabama, only been here ~2.5 years 08:53:50 * Yurik tried to avoid drinking alcohol a lot 08:53:51 Now, that's what I call a move! 08:53:54 * Yurik tries to avoid drinking alcohol a lot 08:54:45 yah, me too 08:54:48 I try to avoid it 08:55:41 * Yurik got a bad vodka-n-beer trip few days ago :-( 08:57:42 hm, any latin experts here? 08:58:11 If so, I'm not oen of them. 08:58:15 s/oen/one/ 08:58:18 heh :) 08:58:36 tmoertel: any suggestion for the haskell-libs group? 08:58:56 juhp suggested that I announce its existence on the haskell list and solicit lib suggestions and contributors there 08:59:20 That would be a good idea. 08:59:33 any other suggestions? 08:59:46 You might also want to ask about orphaned research projects. 08:59:53 * Heffalump tries to remember what happened to the Haskell libraries project that was discussed there 09:00:11 Heffalump: it turned into the haskell-libraries mailing list, which has very little traffic 09:00:20 at least, that's what I've seen 09:00:22 There's a lot of good stuff that started well but never was finished. 09:00:30 hmmm :/ 09:00:32 yah, I'd like to collect stuff like that 09:00:38 maybe the new FFI stuff would help 09:00:46 s/would/will/ 09:00:47 for example, I found a Hsh.hs, a minimal demo Haskell shell 09:01:11 and someone just finished writing an LDAP client in Haskell that might show up in the CVS tree if they approve 09:01:14 Someone here was doing that for his project. I wonder how he got on 09:01:25 Igloo: was doing what? 09:01:26 igloo: doing what? 09:01:29 Haskell shell 09:01:30 collecting stuff? 09:01:32 ah 09:01:53 "here" meaning your physical location? 09:01:55 or #haskell ? 09:02:03 and who, if the former? 09:02:08 * shapr pries proper out of Igloo 09:02:13 ahem, "proper nouns" 09:03:07 the problem with a Haskell shell is that dynamically loading Haskell scripts isn't actually possible 09:03:07 * shapr googles for haskell shell 09:03:13 http://www.informatik.uni-bonn.de/~ralf/software/examples/Hsh.html 09:03:16 Physically here, and Gareth 09:03:23 Gareth who? 09:03:31 nullity...ermmm... 09:03:35 ah, right. 09:03:42 It was hugs based I think 09:03:46 Heffalump: yah, dennisb and xbill helped me clean up Hsh.hs so that it works with ghc5 09:03:53 Gareth Smith 09:04:09 Oh, OK, I was never going to get that 09:04:25 shapr, ah, cool 09:04:55 Heffalump: yah, I'll be putting it in as the first snippet for haskell-libs 09:05:17 I'd really like to see a haskell shell that could load haskell scripts 09:05:28 even if has to recompile and restart itself :) 09:05:53 I guess if it was based around ghci, it'd be feasible 09:05:56 but that's rather big 09:06:05 better than nothing 09:06:32 --- join: ski (~md9slj@fraggel39.mdstud.chalmers.se) joined #haskell 09:06:35 hmm. 09:06:38 hi ski 09:06:43 hi ski! 09:06:45 hi Heffalump 09:06:47 hi shapr 09:06:58 w00, we got 17 people on at the same time 09:07:02 I think that's a record 09:07:28 how many are (at least a little) active of them ? (approx.) 09:07:35 aha 09:07:41 :) 09:08:18 hmm. What dynamic loading needs is essentially some kind of reflection module 09:08:55 Heffalump: that's something you wished for yesterday, right? 09:08:59 what would that involve? 09:09:51 * tmoertel is on the phone (again) . . . 09:10:07 tmoertel: tell me you have a nice comfy headset ;) 09:10:29 well, firstly a runtime library that could load up a haskell module (including parsing and type-checking) 09:10:48 that would be fun. 09:11:08 then, represent the types of the things that module exported in a Haskell data structure, and return those to the function that called it 09:11:41 and finally, provide a means for the functions in the newly loaded module to be actually called 09:12:08 that strikes me as the nastiest bit, conceptually 09:12:20 because if that module defines data types that are needed to call it 09:12:43 then you have to dynamically create them? 09:12:59 the function doing the calling needs to dynamically construct instances of that data type. So there'd need to be some kind of marshalling 09:13:06 well, instances of them, yes 09:13:20 Python has an interesting solution... there's a "new" module 09:13:36 which can turn an instance into any other instance, even one that doesn't exist 09:14:06 well, Haskell is strongly typed, so... 09:14:14 yah... 09:14:26 Python is strongly typed also 09:14:30 should one be able to load *any* module into the system with this ? 09:14:31 but dynamically strongly typed 09:14:59 errm, Python doesn't give types to higher-order functions, does it? 09:15:02 ski: I'd be happy with just enough to let me define new commands for a haskell shell myself 09:15:10 (i.e. even modules that one don't know the interface of yet ..) 09:15:12