00:11:06 googler has quit 00:11:06 sbp has quit 00:11:21 googler has joined #swhack 00:11:21 sbp has joined #swhack 00:20:27 sbp has quit 00:36:53 sbp has joined #swhack 00:48:11 lol! http://www.metafilter.com/comments.mefi/6048 00:51:24 Are you happy Aaron? :-) 00:51:24 Oh man, some of the people on that thread are real down-to-earth jerks. I mean, just total bastards 00:52:45 But it has to win comedy of the week award. It'd be one of those things I'd print out and hang on my wall 00:52:58 "hey, look what this idiot advised me to do" 00:53:33 I'm undecided about which is the funniest comment 00:54:32 Oh, it has to be the first capn_stuby comment. That just kicks-ass 00:55:00 Then again, I also like "How many weeks of allowance does it take to buy 5+ domain names?" 00:56:23 "If certain posters on this thread are exemplars of proper social development, I hope the kid sticks to his XML." 00:56:49 Socializing with assholes, or hacking XML... I chose T.V. 00:57:15 Oh this bit's brilliant too: "Don't tell me that a group of 25 year old XML monkeys or a group of 14 year old supergeniuses is a good representative model of society that will give the child a sense of how to act in the world." 00:57:28 Love it! 00:58:14 "Ooh, my world is full of scary monsters and shit, and if you don't interact with a tonne of piss-heads, and drug dealers, you'll never learn how to get through life, blah blah blah" 00:59:57 "All of them were Mensa meterial and super smart, but sociopaths to the 19! degree." 01:00:12 See, the problem is, he's talking about clever people there, not intelligent people 01:00:34 "Kids are messy, destructive, snotty, candy-snarfin', soda-swillin' leisure monsters." - so are adults 01:00:46 s/candy/coffee 01:00:58 s/soda/beer 01:01:09 s/leisure/sex 01:01:46 "one (being smart) doesn't automatically preclude the other (fucking around)." - damn straight! 01:02:24 "Actually the kid's very socially adjusted." - pfffffffffffff 01:03:35 anyway, I love it 01:05:12 And I'm very proud to know the "kid" who has an entire Metafilter thread devoted to him... it just makes me dran proud :-) 01:05:22 s/dran/darn 01:05:34 Hmm... I seem to have problems typing the word "darn" 01:05:57 I could ask the OED to change it to dran. Maybe if I slipped them Ł10 or something 01:17:38 Hey, good news! 01:17:44 What? What? 01:17:48 I was just at a bookstore and guess what I saw? 01:17:56 My book? 01:17:57 s/?/./ 01:18:01 You guessed it! 01:18:07 Really?! Yay! 01:18:10 I was going to buy it, but it was too expensive. 01:18:24 Heh! Yeah, it is a bit costly. Did you at least read it? 01:18:30 It was sort of surprising -- they didn't have any Python books, but oh, sure, they had your book. 01:18:37 We were sort of in a hurry so I skimmed it. 01:18:44 Did you write the chapter on the W3C also? 01:18:59 Nah... just the Accessibility thingy 01:19:11 It seemed pretty good... a lot of good advice. 01:19:42 thanks 01:20:13 Don't forget to buy one in future, and one for all your friends, etc. 01:20:28 Right, right. 01:20:40 Anyway, in the meantime, I've remembered to let you know about the SWAG UTIL thing 01:20:48 http://purl.org/swag/util# 01:21:21 what about it? 01:21:43 Is it any good? If so, could you release it 01:21:50 (back into the wild) 01:22:06 isn't it already in the wild? 01:22:17 not yet 01:22:52 how so? 01:26:04 sbp has quit 01:35:25 sbp has joined #swhack 01:35:46 it hasn't yet been released onto the plains of the SW mailing lists 01:36:23 hmm 01:37:03 And another thing... you mihgt be really impressed with this, or you might not be 01:37:22 XNote 01:38:49 XNote? xWebL in a nutshell? 01:39:14 I dunno. It started out as something completely different, or rather, a sideline to xWebL 01:39:16 It looks pretty good at first glance. 01:39:40 and grew into something that can do something a lot like xWebL, but with a slightly different scope 01:40:15 swag/util looks good (aside from the #) 01:41:00 I want to keep the "#", because there's no harm at all in using it - i.e. what have you got to lose? 01:41:23 well, when the new regime comes, it won't work 01:41:32 pardon me? 01:41:33 all the properties aren't resources 01:41:42 it's not even a URI 01:41:46 er, what? 01:42:01 The usual # problems 01:42:25 Oh, right. But this is RDF, so the MIME type... do I have to go on? 01:43:04 Er... you're writing the MIME spec for RDF, aren't you? Don't forget to say what the fragment bits mean :-) 01:43:07 well you know my argument... it's just not a resource 01:43:20 It is according to RDF M&S 01:43:26 pffft 01:43:40 well then, don't use HTTP. That's the only option 01:43:50 Umm, no! 01:44:26 if you start using HTTP resources as RDF terms, you lose a way to address the HTTP resource as a network retrievable entity 01:44:39 case in point: your logicerror.com stuff 01:45:07 @prefix : . :AaronSw :writtenBy :AaronSw . 01:45:11 use ? 01:45:18 Pardon? 01:45:47 Well that needs to be sorted out, but # is not the solution. 01:46:04 the network retrievable entity never had the URI 01:46:15 Precisely, it needs to be sorted out. HTTP URIs can't identify two different resources. So, either "#" or "urn" 01:46:15 HTTP is very clear on this: a URI represents a Resource, which can be anything 01:46:25 the server just sends back a bag of bits which is somehow a resource. 01:46:30 err related to the resource 01:46:51 The easy solution is for the bag of bits to say what the resource is 01:47:19 So how do I identify the bag of bits, given that that is what people usually associate with a URI? 01:47:28 No it's not. 01:47:31 Practically speaking, it won't work 01:47:43 That's not true, hardly anyone associates a bag of bits with a URI. 01:47:52 They associate a URI with a concept, which expresses itself as a bag of bits. 01:47:54 bag of bits => I mean, the page itself 01:48:04 like http://aaronsw.com/ is the concept of "my homepage" 01:48:19 the bag of bits changes over time, and in the future may be an amusement park or something 01:48:23 but it is still my homepage 01:48:38 log:includes "Aaron Swartz" . 01:48:52 There, now that's referring to the bits... even CWM does it 01:49:33 [:date "2001-08-02" ; :resource ; :protocol :HTTP ] log:includes "Aaron Swartz" . 01:50:10 that's just not practical 01:50:22 well, log:includes is a shorthand for that 01:50:32 with the parameters set to the default 01:51:17 as TimBL said, you can't ask for Dan over HTTP 01:51:51 Of course not! 01:52:14 what's your point 01:52:30 you can't identify Dan in HTTP space 01:52:41 because he's not there 01:52:45 That's not true. 01:53:03 HTTP identifies resources, of any sort. 01:53:12 But it can't return resources, no protocol can. 01:53:31 Amazon is an implementation of the isbn: scheme 01:53:48 but it doesn't return the resource behind isbn:, it returns an entity -- a book 01:53:51 URIs can map many to one, I don't care about that 01:54:10 That's not my point. 01:54:14 I never said that. 01:54:16 Interesting: http://home.snafu.de/castor/projects/pyx/ 01:54:22 I care that you're trying to represent Dan in HTTP space, when that's impossible 01:54:29 why is it? 01:54:35 what in the spec implies that? 01:54:43 Try readint the HTTP specification 01:54:53 Yeah, I have. 01:55:50 PyX: that's clever. They take the worst feature of Python, and add it with a metalanguage that no one in the world really understands (or very few people) 01:56:07 Worst feature of Python? Not so. 01:56:17 And how many people understand HTML? 01:56:18 significant whitespace? Eek 01:56:29 very few... perhaps none 01:56:33 exactly 01:56:41 Hmm, I sort of like StructuredText better. 01:56:46 err ZopeStructuredText 01:56:50 anyhoo, back to Dan over HTTP 01:57:24 right, can i have a spec quote? 01:57:59 I have to cite this bit... just have to:- 01:58:00 [[[ 01:58:03 It is important, on the Semantic Web, to be clear about what is identified. An http: URI (without fragment identifier) necessarily identifies a generic document. This is because the HTTP server response about a URI can deleiver a rendition of (or location of, or apologies for) a document which is identified by the URI requested. A client which understands the http: protocol can immediately conclude that the fragementid-less URI is a generic document. This is true e 01:58:08 ]]] - http://www.w3.org/DesignIssues/Fragment 01:58:12 now onto the specifications... 01:58:35 Fragment is blantantly false, IMHO. 01:59:19 * sbp remembers the latest RFC for HTTP; does Aaron? 01:59:21 :-) 01:59:23 RFC 2616 01:59:29 good job. ;-) 02:00:02 Point me to the acceptable return code for a query on Dan 02:00:11 http://www.w3.org/People/Connolly/Dan 02:00:34 200 02:00:49 200 OK 02:01:26 10.2.1 200 OK 02:01:27 The request has succeeded. The information returned with the response 02:01:27 is dependent on the method used in the request, for example: 02:01:27 GET an entity corresponding to the requested resource is sent in 02:01:27 the response; 02:01:33 what entity should be returned? 02:01:49 is has to be a representation of Dan. That's impossible 02:02:05 A page that states, the URI you have requested represents Dan Connolly and gives a description of him, etc. 02:02:08 a representation? 02:02:12 it doesn't say that! 02:02:21 it says: "an entity corresponding to the requested resource" 02:02:28 that's what's being returned, my friend 02:02:58 Well, I'm very unhappy about it. I'd say that it's information *about* Dan, not Dan himself 02:03:34 Well it corresponds to the resource, no? 02:03:39 I mean, I don't want to give you Dan. 02:03:47 Perhaps I should send you a Not Authorized? 02:03:58 Don't take away my Danny-boy! 02:05:22 I think you're twisting what HTTP should be able to do... it's a hypertext transfer protocol: transferring data suitable for HyperText systems. That's just data, MIME an' all 02:05:35 I never contradicted that. 02:05:42 I don't disagree. 02:06:00 er... so you agree? 02:06:11 Yes, I agree that's what HTTP Is supposed to do. 02:06:22 * sbp wonders why you didn't save a few characters 02:06:26 But there's also a social contract, of sorts, involved. 02:06:36 It's a very weak social contract 02:06:39 If I request a resource, I want something related back. 02:06:52 Otherwise URIs wouldn't be very useful. 02:07:03 404 02:07:05 I'd give you one but it'd be something different every time you visited it. 02:07:17 sometimes they're not... 200 O.K. is a wonderful 02:07:30 er.. 02:07:33 hang on 02:08:34 Ah, bollocks. Conceptually, I should concede. But practically and intuitively, it seems very wrong 02:09:07 Why? 02:09:18 :AaronSw :wrote :AaronSw . 02:09:19 Perhaps you should read whats-his-name's thesis. 02:09:25 See, that's wrong. 02:09:29 We need to be careful. 02:09:34 It's the "Mona Lisa Problem". 02:09:57 I know what's wrong, I know how to correct it 02:10:05 I'm saying that most people don't, and won't 02:10:22 And using # will correct that? 02:10:58 It'll stop them from getting confused. A URN scheme would clearly be better 02:11:35 OK, so let's go with URNs then... or TAG/TANN/RAKJKS. 02:11:56 But don't try and screw with implemented Web Architecture! 02:12:44 Fine, let's go with URNs 02:12:51 You can help me with the PTS scheme 02:13:31 http://infomesh.net/2001/07/urn/urn-pts.txt 02:14:10 RAKJKS? 02:14:22 I was thinking of PTS, couldn't remember the name, tho 02:14:26 :-) 02:14:33 What do you need help with? 02:14:48 I basically haven't got a clue what I'm doing 02:15:11 How so? 02:15:53 in that the RFC track scares the shit out of me. I'm not sure what format it has to be in, or who to send it to, because the IETF documents appear to be in conflict 02:16:22 Well, put my name on it and we'll do it together. ;-) 02:16:24 the registration documents for URNs, and the IETF registration details... all so complex and stupid. And the damn text-formatting 02:16:28 O.K., fine 02:18:13 refresh it 02:18:47 OK, i'm with you. 02:19:41 err, and at the bottom of the page 02:19:57 Can't you capitalize email? 02:20:22 Where? 02:20:30 email address: 02:20:44 and why doe the * go before the term in your BNF? 02:20:56 O.K., done it 02:21:07 * before: er... perhaps that's an ABNF thing 02:21:21 It seems really strange. 02:21:28 are you sure? 02:21:48 * sbp checks 02:22:07 path_segments = segment *( "/" segment ) 02:22:11 from the URI RFC 02:22:13 oddd 02:22:14 and why no email addresses? 02:22:25 no email addresses where? 02:22:27 in pts 02:22:30 only domain names 02:22:40 ah yes: couldn't be bothered 02:23:16 domain names are just as easy to get as email addresses, anyway 02:23:29 not true 02:23:33 email addresses are free 02:23:35 domain names are not 02:23:46 yes true. it includes subdomains, and I know places that give away subdomains 02:24:03 good point... 02:24:22 hmm, not sure if that is a good idea or not... 02:24:29 who owns foo.blogspot.com 02:24:32 or foo.weblogs.com 02:24:50 who owns blargh@yahoo.com? 02:25:05 ok, i take the point 02:25:28 that's part of the pain-in-the-arse about URIs and security 02:26:25 the only odd bit in the specification is:- 02:26:26 For example, if Fred Bloggs owns the domain example.org for the 02:26:26 entire month of May, in the year 2002, then he acquires the right 02:26:26 to create names under the authority component ():- 02:26:26 example.org,2002-05 02:26:26 This is a non-transferable right. Persons or entities MUST NOT 02:26:28 create URNs using authority components () that they do 02:26:31 not own based upon the rules above. 02:26:36 I couldn't think of an easy way around it 02:26:45 What's odd about that? 02:29:36 sbp has quit 02:37:46 sbp has joined #swhack 02:38:05 I'm just cleaning up the PTS spec now... 02:38:10 Cheers 02:38:32 note the spelling mistak 02:38:37 lol 02:38:42 s/mistak/mistake 02:38:44 multiple ones. 02:38:45 in the draft:- 02:38:51 s/curent/current 02:38:56 yep, it's just a draft 02:40:37 in the script, could you replace:- 02:40:37 auth = m[0] 02:40:37 auth = string.replace(auth, ",", "/") 02:40:37 path = m[1] 02:40:41 with:- 02:40:55 I'm taking out the script, since it's rather self-evident 02:40:58 Do you mind? 02:41:06 auth = string.replace(m[0],",","/") 02:41:27 path = string.replace(m[1],":","/") 02:41:30 no, fine... 02:42:32 Hmm... perhaps that second line should be an error. ":" shouldn't convey any sense of structure by default 02:42:47 But then again, neither should "/" :-) 02:42:53 ;-) 02:43:29 are your cleanup results WWW available yet? 02:43:33 s/WWW/HTTP 02:43:35 almost 02:45:16 oh crap, change:- 02:45:18 month ::= "0" | digitx0 | "10" | "11" | "12" 02:45:20 into:- 02:45:25 month ::= digitx0 | "10" | "11" | "12" 02:45:28 What do you think of: 02:45:30 Process for identifier resolution: 02:45:30 It is possible (but not required) that the identifiers may be 02:45:30 resolved by converting the authority into a request using the 02:45:30 HTTP protocol. In such a method, all commas (,) and hyphens (-) 02:45:30 are replaced with slashes (/). The colon (:) separating the 02:45:31 authority and the name is also replaced with a slash. 02:45:33 02:45:35 For example:- 02:45:37 02:45:39 urn:pts:example.org,2002-05:foo/bar 02:45:41 02:45:43 would become:- 02:45:45 02:45:47 http://exam 02:45:49 ple.org/2002/05/foo/bar 02:45:51 02:45:54 Obviously such a mechanism of resolution would not be persistent. 02:46:06 ...Heh, month 0 02:46:15 fine, apart from the fact that foo/bar is not a valid instance of that component 02:46:21 try foo:bar 02:46:21 really? 02:46:33 name ::= *( char [ ":" ] ) 02:46:33 char ::= 1*( unreserved | escaped ) ; [RFC 2396] 02:46:55 oops 02:47:02 perhaps we should add slash? 02:47:25 shouldn't it be: 02:47:29 char | ":" 02:47:30 ? 02:47:42 That was one of the first things that I thought of, but I think people would be too hung up on directory structures 02:47:47 er, yes, I think it should 02:48:06 no, actually, it isn't 02:48:26 because otherwise it would allow urn:pts:example.org,2002-05:::::::blargh::::: 02:48:52 umm, ok 02:49:34 take my word for it 02:49:46 OK, so: 02:49:47 For example:- 02:49:47 02:49:47 urn:pts:example.org,2002-05:foo:bar 02:49:47 02:49:47 would become:- 02:49:50 02:49:51 http://example.org/2002/05/foo/bar 02:50:05 yeah, that's absolutely fine 02:50:38 and how about this: 02:50:38 Rules for Lexical Equivalence: 02:50:38 Two PTS URNs are equivalent if the strings are 02:50:38 character-for-character equivalent. 02:50:52 fine 02:51:23 OK: http://logicerror.com/ptsURN 02:52:26 Yep, that's neat 02:52:40 And spell checked. 02:52:47 Should I submit it? 02:53:07 Hmm... take out the "::" in the BNF... that's ABNF 02:53:23 What should I replace it with? 02:53:32 delete it 02:54:26 OK, now? 02:54:52 The , and tokens are imported 02:55:03 check that out! Must have copied and pasted it from a former draft 02:55:04 so? 02:55:15 there are no tokens of those names! 02:55:27 err, yes there are 02:55:32 er... of hostname there aren't 02:55:37 domain = hostname ; [RFC 2396] 02:55:40 oops, sorry 02:55:41 and it's imported 02:56:00 yep, my mistake 02:56:21 O.K., it's fine. But not formatted correctly yet 02:56:34 That's OK, that's what RFC Editors are for. 02:56:35 ;-) 02:56:55 Ahh, RFC editors 02:58:15 OK, so you want to send it or should I? 02:58:25 you can... but now? really? 02:58:31 Yes! Why not? 02:58:42 If not now, never. 02:58:59 Well, we haven't asked Sandro, or read the URN materials (well, I have) 02:59:20 Well that's what the 2-week discussion period is for. 02:59:30 correct:- 02:59:30 Registration date: 2001-07-21 02:59:36 and thne send it :-) 02:59:53 Should you send it or should I? 03:00:19 you can, if you really really want to 03:00:33 well i'm not the declared registrant 03:00:40 send it from seanaaron@blogspace.com :-) 03:01:01 No, seriously. 03:01:25 Oh fuck, just change the registration details then! 03:01:31 Thank you. 03:02:14 That part of the registration form is actually a bit bizarre 03:02:22 Why? 03:02:26 Registration date: 2001-07-21 03:02:29 oops 03:02:56 URNs are supposed not to change, and this particular scheme delegates away the authority according to a strict algorithm 03:03:18 how can one "own" an algorithm, and why the feck would one want to? 03:03:52 anyway... have you sent it yet? Who did you send it to? Can you CC it to www-archive if you haven't sent it already? 03:04:10 urn-nid, uri@w3.org, sandro, tim kindberg, you 03:04:15 do we really need www-archive too? 03:04:19 Wow... nope 03:04:39 It's 2001-08-03 03:04:52 :-) 03:04:58 Not where I am. ;-) 03:05:06 He he he 03:05:08 And I'm registering it. 03:05:15 sent 03:05:19 Cool! 03:06:11 It's been two weeks since http://lists.research.netsol.com/pipermail/urn-nid/2001-July/000253.html 03:06:15 What should we do now? 03:06:31 http://lists.w3.org/Archives/Public/uri/2001Aug/0000 03:07:01 we could send it to the proper registration email address, but there's no point if we move PTS through 03:07:18