00:31:37 Morbus has quit 00:57:33 MikeM has joined #rdfig 01:03:02 sbp has joined #rdfig 01:06:27 xena has quit 01:06:56 xena has joined #rdfig 01:23:30 Morbus has joined #rdfig 01:24:21 Morbus has left #rdfig 01:27:56 MikeM has quit 01:31:38 xena has quit 01:33:37 Morbus has joined #rdfig 01:34:05 Morbus has left #rdfig 01:35:10 xena has joined #rdfig 01:41:12 <[filsa]> [filsa] has joined #rdfig 02:11:59 * SethR waves to sean 02:12:20 * sbp waves to Seth 02:12:54 and great changes in rdfwiki? 02:14:18 none that have been released :-) I've been hacking URI schemes/URN namespaces 02:16:34 what you think of http://www.ietf.org/internet-drafts/draft-mealling-uri-ig-00.txt ? 02:18:21 woah! Where did you get that from? 02:18:32 it was chumped earlier today 02:18:41 I thought it was O.K., but it doesn't answer many questions 02:19:03 it just says "most URI materials are badly out of date". Pffff, tell us something we don't know 02:20:34 how come tdb is not on the list at http://uri.net/urn-nid-status.html ? 02:20:38 lol! "Request For A Formal URN Namespace: "geo" by the Great Sean B. Palmer" 02:20:52 Isn't tdb a URI scheme? That's just for URNs 02:21:27 F:|GepGratis 02:21:27 titled item F 02:21:30 F:|GeoGratis 02:21:31 titled item F 02:21:47 F: 02:21:47 http://geogratis.cgdi.gc.ca/frames.html 02:21:48 GeoGratis 02:21:50 (sandro) Canada's "site that distributes 02:21:51 (sandro) geospatial data of Canada." 02:22:10 neat 02:22:33 Hmm... tdb is a URN 02:22:53 J:|URN NID Assignment Status 02:22:53 titled item J 02:22:58 sean, i gues i just assumed if it not a url ist a urn .. but i guess not 02:23:28 :-) 02:23:40 It's a URN if it starts with "urn:" 02:24:05 yep, keep forgettijng that 02:24:45 see, why have long URIs like http://ca.maps.yahoo.com/py/pmap.py?Pyt=Tmap&addr=615+Booth+St.&city=Ottawa&state=ON&slt=45.399400&sln=-75.704200&zip=&country=ca&mag=9&cs=9&name=&desc= when you could have a "geo" URN? :-) 02:26:57 * gerald is 2.1km from there :) 02:27:30 Hmm... there really are a lot of potential applications, aren't there? 02:27:49 cool! 02:27:49 xena has quit 02:28:04 What's at that address? 02:31:00 the center that creates the maps 02:31:23 sean, is urn:urn-n primarially a formal method of coding gen id's ? 02:32:31 it's an informal namespace space 02:32:50 oh, sorry... you're referring to the urn-anon informal proposal thing, I guess 02:33:07 er, yes, it's for representing anonymous nodes 02:33:22 "existentially quantified variables" 02:33:34 Sean, what's at 615 Booth St.? 02:33:47 I dunno! 02:33:53 Booth St., where? 02:34:12 aaron, that map uri is linked from http://geogratis.cgdi.gc.ca/frames.html 02:34:18 [ :at [ :name "615 Booth St." ] ] . 02:34:26 i see 02:35:21 sean, i though gen id's were used to identify annon nodes 02:35:58 yeah, they are. But they're not URIs. People seem to be using any old pile of crap at the moment, and it may confuse people into thinking that they're URIs, when really they're not 02:36:49 SiRPAC used genid:9725238 02:37:34 so your proposal just formalizes how to put identifiers on annon nodes ? 02:37:51 formalizes=standardizes ... that is if it catches on 02:39:06 yep 02:39:33 but, I'm not going to register it, because they're so much confusion at the moment when it comes to anonymous nodes 02:41:58 yeah 02:44:19 Have you seen the new esl URI scheme? 02:44:31 try: http://purl.org/net/esl 02:47:17 signed uri .. how kewl ! 02:47:32 thanks :-) 02:47:45 Hmm, should the labels themselves be URIs? 02:48:18 no, because that would be very confusing 02:49:14 if I digitally sign http://www.mycompany.com/, I own that signed URI, but myCompany (whimsical example) aren't going to be happy 02:54:29 sbp, I'm not sure I understand the role of the label, or at least not why it's mandatory. I'd like to have some kind of URI scheme to refer to a particular pile of bits, e.g. an MP3 file; why does it need a label in addition to an sha1 hash? 02:56:04 oh, I think I'm confused about what this URI scheme is for. 02:56:29 these uri are very persistant .. they can never go away ... and can only get reallocated by the person who signed them ... unlike a http://host type uri 02:56:45 is that right? 02:58:28 ok, I think I get it. It doesn't seem directly related to e.g. an md5: uri scheme that could identify a specific file. 03:00:20 correct 03:00:20 So what good are these signed URIs? 03:00:33 they're like tags, except that they're "secure" 03:01:15 if you see one, you know proof positive that identity was created by a specific person 03:01:16 you can sign a document describing the tags using the same key... 03:01:20 secure meaning what? 03:01:22 exactly 03:01:27 what Seth said 03:01:33 No you don't! 03:01:46 well by a person that knows the key 03:01:49 created by a specific person, or someone with access to their computer and key :) 03:01:56 No... not even that. 03:02:01 access to the key 03:02:04 Nope. 03:02:11 aaron, why not 03:02:12 If you sign something and give it one label. 03:02:20 I can take the signature and give it another label 03:02:26 and create a new URI 03:02:31 without knowing your key 03:02:52 i though the label was signed along with the authority 03:03:17 Oh? 03:03:26 sean is that right? 03:03:47 you can't sign using someone else's key 03:04:01 We're not talking about that. 03:04:11 what are you taking about, then? 03:04:24 You sign an email, and send it to me. 03:04:29 I take the signature and make a URI out of it. 03:04:40 yes... 03:04:54 So that URI is not secure. 03:05:18 you cant sign your mail with my sig cause you dont know my key 03:05:51 how can it not be "secure"??? I still own the URI. You can verify that I created it 03:05:58 that's what digital signatures are for 03:06:01 You didn't create it though! 03:06:05 I did! 03:06:31 Huh? If I digitally sign a document, then that is my digitally signed document, no matter how other people use it 03:06:40 Umm, yes... but the URI... 03:06:54 the URI is a digest of a digitally signed string 03:07:12 Like if you make urn:esl:kdfkdjfkdjfklsdf:iloverdf I can turn around and make urn:esl:kdfkdjfkdjfklsdf:rdfreally sucks 03:07:40 err, minus the space 03:07:49 and your uri will not pass sig test 03:07:57 exactly 03:08:42 woops, sean, dont you need the public key in the uri 03:08:54 otherwise you cant check it 03:09:15 you include the label in the URI? 03:09:15 err, the signature 03:09:24 Ahh, never mind. 03:09:29 that's true, but I expect that people will use these in such a context that the signing public key will be known 03:09:37 I misread the draft. 03:09:48 :keyLocation . 03:10:45 but if you dont know the public keyLocation, the uri is just a string of giggerish 03:10:57 cept of course for the label in the clear 03:11:24 I guess so... but how do people use digital signatures? they don't tend to send over their public key every time that they sign a document 03:11:39 i see 03:12:16 Hmm, for long documents these URIs will get pretty long. 03:12:24 long documents??? 03:12:37 you may need to provide a use case where this kind of thingy is needed 03:12:54 "the "esl" URI scheme is really just a method of putting a digitally signed message unambiguously and fairly tersely into URI space." 03:13:17 If my digitally signed message is long, this is going to be a really long URI. 03:13:23 For a start, a digital signature has a set length based upon a hash. secondly, this is for identifiers, not documents... yuo could theoretically use a document 03:13:43 aaron, if that all you want just sign the message 03:13:48 well, a message in this case is just a short string. I'm not going to define short, just like the URI specification doesn't 03:14:09 URIs need not be short... 03:14:38 esl doesn't constrain the length 03:14:53 er, well, the ABNF defines the shortest syntactic length 03:16:07 So, what's the usecase for this? 03:16:33 Web of trust, that sort of stuff 03:16:44 Can you explain? Like a real usecase? 03:17:00 plexnames people plexnames 03:17:07 plexnames are the future! 03:17:14 uh oh 03:17:17 Not until you put in some add infrastructure! :-) 03:17:29 plexnames only map to URIs, anyway 03:17:40 .e plexname 03:17:43 well, it protects against people contesting your decentalized URI. Even TAG (basing itself on DNS) can't guarantee that 03:18:02 xena has joined #rdfig 03:18:06 * sbp wonders what channel he's in 03:18:08 .e plexname 03:18:20 .e plexnames 03:18:26 Hmmph. 03:18:32 Plexnames - http://plexnames.com/ 03:18:40 whats a plexname ? 03:19:22 Tav's DNSy system 03:19:36 well, the current implementation of plexnames is very simple 03:20:06 whats a plexname ? 20 words or less 03:20:55 * sbp wonders how long before he finds people using plexname:esp 03:21:09 plexnames remind me of dtrt, http://impressive.net/services/dtrt/ 03:21:19 dtrt is cool 03:22:06 a system of decentralised indexes of persistent resource identifiers 03:22:47 you'd be able to say, use my personal index, my families index, my companies index, the rdfig's index, and the global esp index 03:22:48 hmm... decentralized, but all requests go via plexnames.com? 03:23:02 tav's a dreamer, what can i say. ;-) 03:23:11 Not all plans are finished. 03:23:55 gerald: no 03:24:06 they go through the indexes you subscribe to 03:24:45 what can we do right now to help make the dream become a reality? 03:25:00 code? 03:25:05 anything 03:25:31 hmm, that means plexnames are context-sensitive 03:25:40 download, install, rant, propose, code, whatever 03:25:49 sbp: yup 03:26:18 How are plexnames both persistent and context-sensitive? 03:26:29 AaronSw? 03:26:38 That's my name. 03:26:47 elaborate 03:26:56 I wouldn't want to use identifiers that aren't available to most people, or that give different results to different people 03:27:17 yeah, that's it 03:27:37 plexnames resolve differently for different people -- they aren't globally consistent. 03:27:39 universality... 03:27:45 What makes them persistent? 03:27:45 gerald, well you might want to use one of the indexes that allocate a plexname to you on a first come firt serve basis.... 03:27:51 * gerald is also reminded of protozilla, http://protozilla.mozdev.org/ 03:28:04 ah, Protozilla 03:28:21 you won't get a pleasant plexname, but it'd be yours... 03:28:49 i need a bloody password to make a plexname 03:28:56 hmz, lemme change that 03:28:58 O.K., I want plexname:sbp to refer to my homepage, right now. How? 03:30:00 * sbp found the form on plexname.com... 03:30:34 * sbp filled it in 03:30:46 wait while tav fixes things 03:30:49 * sbp found that the site asked for authentication 03:30:55 * sbp tried it again 03:31:07 .me found that "sbp" had already been taken 03:31:13 * sbp oops 03:31:56 * gerald notes that dozens of companies have tried similar things in the last decade or so, but most of them didn't have decentralization as a goal 03:32:46 tav, the add function is really screwed up 03:33:03 if I add a name but don't define it, it still goes in the database 03:33:41 that's the problem that I had 03:34:47 here's one from 1998: (wow, the uri still works! cnet++) http://news.cnet.com/news/0,10000,0-1005-200-330700,00.html 03:36:11 Yeah, but those aren't decentralized/contextualized... what's the point? they're just new uris 03:37:02 New URIs are not good... 03:37:16 er... new URI schemes 03:37:55 But new URNs are just fine, right? 03:38:07 Sometimes 03:38:25 informal ones are fine, register as many of them as you want 03:38:32 formal ones are a different matter 03:39:35 I wonder whether MikeM/TimBL would agree with your comments on informal ones. 03:40:32 I'm sure that they would 03:41:05 that's what informal URN NIDs are for. OF course, they still need to be unique/persistent, etc., and they are reviewed by URN NID 03:41:07 are pelxinames like netword (s) see http://www.netword.com/ ? 03:41:12 Why is an informal URN OK when a new URI isn't? 03:41:12 s/URN NID/urn-nid 03:41:40 because informal URNs are informal 03:41:50 but they are centrally registered 03:42:13 really, the problem is not with new schemes, but new schemes that are not registered, or tacky ones that are 03:42:35 I'm not so sure... 03:43:43 well, usually, when someone comes up with a new URI scheme, you can guarantee that it will be a crud idea. Because people don't understand the relative merits of URIs 03:44:38 e.g. the host of URI schemes that I've come up with... 03:44:43 wooo! 03:44:49 :-) 03:44:50 ok, fixed 03:45:04 * sbp gets his plexname 03:45:13 tav, are pelxinames like netword (s) see http://www.netword.com/ ? 03:45:15 It's been taken! 03:45:27 tav. mine's been taken =( 03:45:30 what do I do? 03:45:48 you need to buy it off the current owner ;) 03:46:00 http://plexnames.com/index/edit?plexname=AaronSw 03:46:01 N: http://plexnames.com/index/edit?plexname=AaronSw from tav 03:46:04 eeek 03:46:07 damned blogger 03:46:28 heh 03:46:33 * sbp senses plexname teething problems... 03:46:37 N:|ignore! 03:46:37 titled item N 03:47:04 .e plexnames aaronsw 03:47:04 http://plexnames.com/uri?plexname=aaronsw 03:47:05 O: http://plexnames.com/uri?plexname=aaronsw from xena 03:47:08 at least it works :-) 03:47:11 heh 03:47:17 O:|Ignore this too! 03:47:17 titled item O 03:47:46 aaronsw plexname doesnt exist 03:47:48 Plexnames and the chump don't get along too well 03:48:02 ok now i got a plexiname=SethsConjecture at plexiname.com ... so what is it supposed to mean ? 03:48:27 Hey! I made that plexname... it better exist 03:48:45 Net, mine works 03:49:07 s/Net/Neat 03:49:17 .e ta 03:49:18 .e tav 03:49:18 error: Site Error occurred: AttributeError 03:49:19 I wonder why mine doesn't 03:49:20 sethr, now you either convince the whole world to use the same plexname resolvers you do, and start publicizing the SethsConjecture: URI scheme, or you don't, and don't ;) 03:49:52 .e tav 03:49:52 tav: http://espians.com/tav 03:49:56 there 03:50:00 chump friendly 03:50:07 thank you, tav! 03:50:08 .e edit AaronSw 03:50:25 .e sbp 03:50:26 sbp: http://purl.org/net/sbp/ 03:50:43 AaronSw: capitalisation... 03:50:56 They're case-sensitive? ick 03:50:59 .e AaronSw 03:50:59 error: Site Error occurred: UnboundLocalError 03:51:01 I don't think that the plexnames are meant to be URI schemes... 03:51:14 sbp: this is just a simple implementatino 03:51:18 UnboundLocalError -- I have all the luck... 03:51:18 "plexname" itself would be the URI scheme :-) 03:51:28 How would it work on a larger scale? 03:51:37 what would a uri in the alleged SethsConjecrure URI scheme look like ? 03:51:43 I assume you'd have an ordered list of plexname resolvers... 03:51:47 and it'd query them in order to find the answer 03:51:51 *nod* 03:51:57 in order of preference 03:52:00 plexname:SethsConjecture (ugh) 03:52:07 * AaronSw came up with this idea a few years back ;-) 03:52:16 I could set it up in Protozilla right now. In fact, I will 03:52:20 nah, sbp: http://plexnames.com/uri?plexname=SethsConjecture 03:52:26 is seth's new uri 03:52:33 they're just PURLs... 03:52:37 sbp: just type 'sbp' 03:52:50 AaronSw: well, they could be pure ips... 03:53:05 that's how I think mid: resolvers oughta work (ordered list, in order of preference) 03:53:05 what could be? 03:53:09 your plexname, then acts as a primary index 03:53:26 how does my plexname act as an index? 03:53:29 within which you could have entire namespaces 03:53:41 gerald, does that mean the W3C will start providing mid: resolvers for its mail messages? 03:53:51 sure, RSN ;) 03:54:02 what's the difference between this an a purl ? 03:54:24 .e hmz 03:54:28 * sbp tries out plexname:sbp in Netscape6... 03:54:37 It works! 03:54:39 sbp: got protozilla setup ? 03:54:46 of course 03:54:58 nice 03:55:22 .e google foo 03:55:22 google foo: http://www.google.com/search?q=foo 03:55:28 .e google foo bar 03:55:28 google foo bar: http://www.google.com/search?q=foo bar 03:55:33 I simply created a file called plexname.url in the PRotozilla directory, with the string http://plexnames.com/uri?plexname= in it 03:55:39 really easy 03:55:53 sbp: yup, i do that on every machine i use ;p 03:55:54 hmm, functional args could stand to be encoded 03:55:57 that or the IE.... 03:55:58 :-) 03:56:11 how does one do it in IE? 03:56:18 the reg file on the frontpage 03:56:28 I installed it, what URI do I use? 03:56:34 you then type 'e sbp' in the address bar 03:56:42 why e? 03:56:53 i dunno 03:57:01 neat, it works 03:57:12 e for esp ;p 03:57:13 heh 03:57:23 can't you get it to use a proper URI scheme? 03:57:32 Hmm, you should change the name to plexnamesp ;-) 03:57:39 tav, so this is like http://www.netword.com/ ... huh ???? 03:57:43 plexnamESP ;-) 03:57:47 SethR: to an extent 03:57:50 I've just realised that that can be combined with bookmarklets 03:57:53 a better example would be realnames.com 03:57:59 no, sethr, nimroy doesn't plug plexnames 03:58:03 :-) 03:58:12 deltab: who chose the initial 'e' ? 03:58:20 you did 03:58:26 did i say why? 03:58:30 and we didn't know why then either 03:58:35 ah good 03:58:42 bookmarklet with similar idea: javascript:void(str=prompt('Do%20the%20Right%20Thing:',''));if(str){location.href='http://impressive.net/services/dtrt/dtrt?text='+escape(str).split('%20').join('+');} 03:58:44 (nimroy doesn't plug plexnames) means ?x 03:58:58 It was just a joke... nevermind 03:59:00 SethR: those are all centralised for one 03:59:13 how will plexnames be decentralized? 03:59:18 netword is just a free version of realnames 03:59:24 meaning anyone can make a plexiname server ? 03:59:28 sbp: anyone can run an index 03:59:46 any code available yet? How would one address those servers: using IP? DNS? 03:59:59 Freenet? 04:00:10 Hmm, tav, you should be able to make chains of plexnames... and then my preferred plexname chain could be a plexname itself? 04:00:12 well, i was trying to come up with a system that provided it with anonymity 04:00:21 sbp, using plexnames, duh! 04:00:26 reminds me of http://www.cfp2000.org/workshop/materials/projects-dns.html 04:00:34 oh blimey... 04:00:56 i.e. http://plexnames.com/uri?backup=http://plexnames.aaronsw.com/uri?backup=... 04:01:13 lol! that's just a very very long HTTP URI 04:01:27 it depends upon DNS, so why not just use DNS? 04:01:30 the idea is, an object is at location A... now, i know where locatin A is because i know the route to location A 04:01:40 O.K. 04:01:54 now, if i were to take a different route to location A i still know where location A is... 04:02:06 now, if location A were to have multiple addresses... 04:02:11 oh, so you're giving the *routes* a name? 04:02:15 e.g. PO BOX 3285902 04:02:23 e.g. 9 mrin gardens 04:02:27 and.... 04:02:37 if my route to them were preset 04:02:58 albeit altered on a fairly regular basis 04:03:11 by me throwing in signed packages into the "mesh" 04:03:21 me = site owner 04:03:21 mesh? 04:03:40 mesh is a place which picks up all this random data you throw about ;p 04:03:46 have you written all this down somewhere? 04:04:15 do irc logs count? ;p 04:04:28 On paper this doesn't make any sense at all, but none of the coolest plans ever do 04:04:38 * AaronSw is getting very excited. ;-) 04:04:40 no, IRC logs don't... well, this one might 04:05:02 well, lemme continue, and hope it makes sense 04:05:09 tav, if i add that download on the front page to my registry, am i deactivating thigs like netword ? 04:05:16 no 04:05:18 SethR: no 04:05:31 you just have an extra key in your registry fo 'e' 04:05:33 for 04:05:42 REGEDIT4 04:05:42 [HKEY_CURRENT_USER\Software\Microsoft\Internet Explorer\SearchUrl\e] 04:05:42 @="http://plexnames.com/uri?plexname=%s" 04:05:42 " "="+" 04:05:48 etc. 04:06:02 so, the mesh picks up the packets 04:06:03 (a few character encodings) 04:06:12 equivalent to you updating the ns records 04:06:20 and the global network picking up on it 04:06:56 now, servers, harvest the data going into the mesh 04:07:19 i presume everyone here is familiar with a dictionary? ;p 04:07:26 a what? 04:07:36 sbp: a less detailed ontology 04:07:37 * sbp looks up that word in his dictionary 04:07:52 er, less rigorous 04:07:59 plexiword:SethsConjecture still doenst go nowhere 04:08:19 if you want to find a word, you can do so in less than 9 turns, by halving the namespace boundaries every time... 04:08:26