00:08:04 * sbp updates http://infomesh.net/2001/07/swad/ 00:08:19 Now contains the links that I found... some quite interesting ones, too 00:11:07 googler has quit 00:11:07 deltab has quit 00:11:07 sbp has quit 00:11:23 googler has joined #rdfig 00:11:23 deltab has joined #rdfig 00:11:23 sbp has joined #rdfig 00:12:49 GabeW has quit 00:20:28 sbp has quit 00:22:35 christian has joined #rdfig 00:25:48 GabeW has joined #rdfig 00:36:47 ChanServ has changed the topic to: 00:36:47 This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome 00:36:51 sbp has joined #rdfig 00:37:04 christian has left channel 01:17:10 * AaronSw smacks chanserv 01:17:27 lol 01:20:54 AaronSw has changed the topic to: RDF and Semantic Web Chat: Two great tastes that taste great together. (http://www.w3.org/RDF/Interest#irc) 01:22:24 in your face links, still: http://infomesh.net/2001/07/swad/ 01:22:52 yes, for printing 01:23:01 Who'd print a list of links? 01:23:11 What good would that be? 01:26:05 sbp has quit 01:32:17 TimBL on WSDL: [[[The description of services is addressed by WSDL, also submitted to W3C. Whilst this defined the functionality well and was clearly requested to become a work item, WSDL in practice defines abstract objects such as services, ports and bindings between them which in the W3C metadata architecture would normally be modelled in RDF.]]] - http://www.w3.org/2001/04/30-tbl 01:35:25 sbp has joined #rdfig 01:55:20 GabeW has quit 02:29:38 sbp has quit 02:37:47 sbp has joined #rdfig 02:51:55 wmf has joined #rdfig 02:52:03 * AaronSw waves to wmf 02:52:08 howdy 02:52:16 How's the new job? 02:52:19 * wmf has some bad Advogato karma 02:52:24 the job is good 02:53:46 What happened on Advogato? How'd you get decertified? 02:53:54 no idea 02:54:03 probably, somebody upstream from me lost a cert 02:54:43 I'm a bit disappointed 02:56:00 I'd certify you, but I already have. 02:56:06 yeah 02:57:48 Hi! How are you? 02:57:55 I send you this file in order to have your advice 02:58:01 See you later. Thanks 02:58:02 What is it this time? 02:58:17 You gotta love that virus, really 02:58:20 the files are all called "DATE" 02:58:27 I got hot stock tips yesterday. ;-) 02:59:13 only 8 this week though 03:00:10 amazing to think that for every time you recieve one of those files, someone got duped by it 03:01:31 heh 03:07:41 tav! you have not certified me on advogato! 03:11:14 not that my certification will change much, but there ya go ;p 03:13:21 wmf / AaronSw - both been certified ;p 03:13:37 thanks tav! 03:14:55 btw, my fellow espians and i have been working more on that nice python bot tonight 03:16:01 cool 03:19:04 hmz 03:19:14 question: what kind of features would you like to see in it? 03:20:05 I'd like it to keep a state stack for me. 03:20:19 hiroko, push: eat dinner 03:20:23 hiroko, pop 03:20:28 eat dinner 03:23:06 hmz 03:23:27 relationship between 'push' and 'pop' ? 03:23:41 err, they're standard stack operations, no? 03:23:55 I suppose in Python it's append and pop... 03:24:21 >>> foo = [] 03:24:21 >>> foo.append(1) 03:24:21 >>> foo.pop() 03:24:21 1 03:24:27 oh 03:25:52 hmz, that can be arranged 03:26:12 It can't be too hard. 03:26:27 To make it a little more difficult, make it so that it can keep a personal one and a group one. 03:28:05 well, i have .reminders in-dev 03:28:21 i guess, i can add .note to that 03:29:24 AaronSw: why do you want stacks? 03:29:33 there are too may stacks in my life 03:29:42 Because I always pop my stack and forget what I was supposed to be doing. 03:33:28 sbp has quit 03:35:25 tav` is now known as tav 03:54:24 tav, where can I get this bot? 03:54:55 hmz 03:55:23 is that, 'where can i get the source code' or 'where can i get to see it working?' 03:55:32 source code, preferably 03:55:40 if not, then playing with it is fine 03:56:20 source code, soon 03:59:56 i'll have a play version in here in about 5 minutes 04:04:26 hmz 04:22:55 killarny has joined #rdfig 04:23:43 hmz 04:23:51 disconnected on me 04:24:23 msg it again 04:24:40 Hiroko has joined #rdfig 04:24:51 Hiroko is now known as hiroko 04:25:04 woo 04:25:21 AaronSw: here she is ;p 04:25:29 .time 04:25:29 2001/08/03 04:25:30.1996 Universal 04:25:39 .push release source code 04:25:46 .help 04:25:46 usage: help [] 04:25:47 use: commands for a full command listing 04:25:48 wmf has left channel 04:25:48 - if command is specified, specific usage help is displayed for that command 04:25:49 - [] characters denote optional parameters, <> indicates a variable 04:25:53 .help note 04:26:08 lemme whip up note 04:26:22 look at the other commands whilst i do that 04:26:34 help commands 04:26:36 .help commands 04:26:36 usage: commands 04:26:37 - lists valid commands, both local and from the Zope IRC Commands Interface (ZIRCI) 04:26:50 .commands 04:26:50 valid commands: 04:26:51 - babel, change, commands, deadline, dns, email, excuse, exec 04:26:52 - flight, foo, fortune, google, help, imdb, intelliquote, join 04:26:52 - modules, news, news source, nick, part, ping, plexname, quit 04:26:53 - quote, raw, register, rehash, search, seen, symbol for, time 04:26:54 - translate, tv, weather, when did, whois 04:28:01 heh 04:28:04 .babel 04:28:04 usage: babel  04:28:05 - translates the given phrase from one language to another, as specified by the translation method 04:28:06 - languages: german (de), english (en), spanish (es), french (fr), italian (it), portuguese (pt) 04:28:07 - translation methods: de-en, de-fr, en-de, en-es, en-fr, en-it, en-pt, es-en, fr-de, fr-en, it-en, pt-en 04:28:13 aah, too many choices 04:28:26 hehe, wait till you see .translate ;p 04:28:59 .translate eng-fre hi! my name is Aaron Swartz 04:29:00 salut!! je m'appelle [Aaron] [Swartz] 04:29:11 ;-) 04:29:20 god this cli irc sucks 04:29:33 bitchx killarny? 04:29:37 yeah 04:29:40 it's lagged 04:29:47 * AaronSw plays with it on private msg not to distrub everyone else 04:30:16 well, i'm interested in how ppl use it 04:30:24 wanna take it to #hiroko? 04:30:30 sure 04:30:36 .join #hiroko 04:34:27 killarny has quit 04:37:20 hiroko has quit 07:59:35 tav has quit 08:07:29 danja has joined #rdfig 09:00:45 http://images.google.com/images?q=rdf 09:01:18 A: http://images.google.com/images?q=rdf from danja 09:01:50 A: Google image search (beta) 09:01:50 commented item A 09:07:57 danja has quit 10:27:26 ambient has joined #rdfig 11:17:49 ambient has quit 11:21:41 Phil has joined #rdfig 11:22:18 Phil has quit 11:27:50 rrs has joined #rdfig 12:02:05 ambient has joined #rdfig 13:22:47 tdxdave has joined #rdfig 14:14:56 ambient has quit 14:52:16 sbp has joined #rdfig 14:59:04 * DanC_ waves from 1st telcon of the day 14:59:30 Hi Dan 15:01:02 sbp has quit 15:11:06 sbp has joined #rdfig 15:24:53 ambient has joined #rdfig 15:30:05 tdxdave has quit 16:03:32 sbp has quit 16:12:24 sbp has joined #rdfig 16:15:43 GabeW has joined #rdfig 16:36:26 jhendler has joined #rdfig 16:36:40 Hi Jim 16:36:54 morning 16:36:59 afternoon 16:37:14 day 16:37:28 (evening soon) 16:37:41 There goes the Sun King... 16:38:59 Sandro, what are your feelings on the pts scheme? 16:39:50 * AaronSw has decided it's best to ask difficult questions about things when people are asleep. 16:40:22 I agree 16:41:08 jhendler is now known as jimH 16:42:58 pts? 16:43:18 see http://logicerror.com/ptsURN 16:43:53 and http://infomesh.net/2001/08/pts/ 16:45:48 * sandro was indeed taking a nap just now. :-) 16:45:55 :-) 16:47:06 AaronSw (and sbp), I was a little surprised you didn't ask me first before going public with it. And I didn't understand all of it. Can you summarize how it's different from tags. 16:47:19 Er, actually I did ask you. you didn't reply 16:47:38 Did you ask here? or in e-mail? 16:47:47 I asked here, in private message 16:48:13 different from tags: 1) automatic persistence. tags have a wider scope. 2) people can't use email addresses, 3) people have to hold the authority component for an entire month 16:49:15 plus, we're actually planning to register this thing. I'm openly frustrated at the speed of the TAG registration. And because it's a new URI scheme, it's going to take longer to get consensus, and go through the IETF 16:49:25 Ah, sorry -- I subscribe to the W3 practice of believing IRC is inherently an unreliable medium (because it's so hard to make it reliable). Sorry I missed your message. 16:49:48 Sorry, we probably should have emailed you first. 16:49:51 That's O.K. Sorry I didn't take more time to track you down :-) 16:50:07 Yeah, I've been meaning to check on where tags are. 16:50:10 Yeah, it really was more of a spur of the moment thing... but enthusiasm for the idea too 16:50:13 em_ has joined #rdfig 16:50:22 * AaronSw waves to em_ 16:50:25 Hi em 16:50:28 * em_ nick em 16:50:32 karld has joined #rdfig 16:50:35 err /nick em 16:50:42 * em_ waves back to AaronSw 16:50:45 http://www.ietf.org/internet-drafts/draft-kindberg-tag-uri-00.txt 16:50:45 B: http://www.ietf.org/internet-drafts/draft-kindberg-tag-uri-00.txt from sandro 16:50:48 * AaronSw nominates http://www.w3.org/Talks/2001/07/30-swws/slide4-0.html for picture of the year! 16:50:53 * em_ waves back to sbp 16:51:13 Yeah, that photo is great! 16:51:15 :) 16:51:36 thanks! 16:52:04 Hmm... the ID expires on Jan 2002. How long do you have to wait before you can do anything with it? 16:52:14 B: so it's an internet draft form -- I don't think there's any rule stopping it from becoming an informational RFC, but I don't remember the proceedure. 16:52:14 commented item B 16:52:20 B:|The 'tag' URI scheme 16:52:20 titled item B 16:53:05 * em_ tries to play catchup on 19 phone messages from last week 16:53:13 sbp was also somewhat annoyed that TAGs aren't necessarily persistent 16:54:20 * em_ reads rdfig logs for the past week 16:54:29 So (1) I'm unclear on the semantics of persistence. What does it mean for an identifier in the semantic web to be constant or persistent? I think if you're taking a monotonic view, everything is naturally persistent, and if you're not, things are naturally messy anyway. So I don't think calling an identifier for things more complex than, say, bits, "persistent" means anything. 16:54:34 That was *the* reason for the PTS scheme 16:55:39 anyone using mirc by chance having problems connecting to rdfig? 16:55:59 The semantics of persistence is the right to go to people's houses with a baseball bat if they break persistance... 16:56:01 Yes... I'm well aware of the philosophy behind URI persistence, but I think that there are levels. With TAG, you're saying that denotations can be temporary - and I interpet that as meaning that resources can change to the extent that Joe Public would say "oh, you've changed what that means" 16:56:06 [I kind of like the hack about turning PTS into HTTP, 'though I'm not sure what do about it.] 16:56:23 mIRC - yes, I had to try 8 different windows to connect to OPN 16:56:30 [It works automatically for many date uri servers, like w3.org and infomesh.net] 16:56:42 I even wrote a Python CGI script for the PTS => HTTP thing 16:57:04 WIPO could resolve pts URN 'owernship' disputes, but not tag URI disputes 16:57:17 [Of course, it misses things like the "www." which is strongly perferred on w3.org.] 16:57:22 (at the bottom of http://infomesh.net/2001/07/urn/urn-pts.txt) 16:57:47 [sandro, urn:pts:www.w3.org,2001-01:WSWS] 16:57:57 yeah, subdomains are allowed 16:58:20 Can you give me a tag ownership dispute scenario? That relies on now being clear about who was the regestrant of some domain at some time, right? 16:58:47 * sbp was wondering about that too 16:59:07 I realize (assumed) that subdomains were allowed, but you don't want to have to specify them for a tag/pts kind of thing. They're just silly for that application. 16:59:35 pardon? It's quite easy to get free subdomains on the internet... that was the thinking 16:59:36 why? 16:59:55 so we don't have to use email addresses :-) 16:59:57 No, I mean tacking the www on w3.org is just silly. 17:00:17 does it matter whether you tack it on or not? 17:00:20 So you complain if we leave it off, but you say putting it on is silly? 17:00:48 It all works out in the end. 17:01:25 I'm just saying that the conversion between pts and http URIs brings up silly issues like the "www." prefix, which is preferred for http and annoying for pts. 17:01:43 I'm not saying it's a showstopper.... 17:01:54 Well, it's only a resolution mechanism. 17:01:56 w3.org redirects you 17:02:17 it makes no claims about what's on the http end of the urn 17:02:49 Yeah, we put in the clause that PTS resolution using this method is obviuosly not relaible. I mean, of course not, otherwise you'd just use the URLs as names! 17:03:30 jimH is now known as jimH2031268 17:03:34 It's just that if you use this resolution mechanism, you may end up with some information related to that PTS... maybe 17:04:05 jimH2031268 is now known as jimH 17:04:55 PTS does not eliminate the need for tag URIs. tag URIs are still fine for when you need temporary denotations, want to use an email address in the authority componenet, or can't wait an entire month for a domain 17:05:32 Is there any reason not to put the same kind of resolution into tags? 17:05:45 (Maybe I should say "resolution"...) 17:05:52 pardon? 17:07:08 sbp has quit 17:07:10 If a tag is a domain-based one, you can turn it into http the same way, right? There's nothing in the design of pts vs. tag relevant to this issue. (also, abbreviated dates are not generally supported on web sites, but that's an authority's issue) 17:07:40 Sure, feel free to borrow the idea. 17:08:42 Okay. Do you think it's important to be in the RFC? Or just something people kind of know about....? 17:10:52 sandro_ has joined #rdfig 17:11:21 I don't really care... 17:11:27 I'm still against the ,1 notation, though 17:11:47 I mean, if folks are willing to encode the whole date into HTTP URLs, I'd be surprised if they couldn't do it with a URN 17:13:07 You know the current draft allows either date form? (so you can have two distinct tags with the same date & data, by using the two date forms -- an unfortunate side effect) 17:13:21 No, I hadn't seen that. 17:13:40 Are thags with the same date equivalent? 17:13:58 i.e. tag:sandro@hawke.org,1-6-5 vs. tag:sandro@hawke.org,2001-06-05 17:14:28 No, they are distinct. We recommend you not use both. 17:14:52 I personally, would recommend you only use the abbreviated form for year and maybe month forms. 17:15:00 Ah, it's not on taguri.org. I should update that. Does it satisfy you? (most folks -- 99.999% of the web? -- are demonstrably not willing to encode the date in their HTTP URIs) 17:15:26 sbp has joined #rdfig 17:15:40 Most folks (99.99999999999% of the Web) are not going to use a tag: URI. ;-) 17:16:35 * sbp reminds himself to add acknowledgements to all of the previous tag:/TANN stuff, including debates on uri@w3.org etc. 17:16:40 most folks don't seem particularly interested in persistence 17:16:51 right. 17:17:06 sandro, are you trying to convince us to drop the proposal? 17:17:07 though some news sites have impressed me with well-designed URLs 17:17:26 Yeah, some are. W3C are, and people who use PURLs are 17:17:40 But I think that the main use for PTS' etc. are for namespaces 17:17:50 PTS' and tags can be used on the SW as well 17:18:40 Seth has joined #rdfig 17:19:10 I'm not sure if it's better tactics to have both proposals go forward or not. 17:20:03 The big thing seems to be persistence of the binding between an identifier and some identified object. Can you guys give me a use case where calling it a URN matters? 17:20:34 I can always assume that a URN was intended to be permanent 17:21:05 (If there is a good reason, then Tim Kindberg and I had been thinking we'd also do urn:tag:, with exactly the same qualities otherwise. But we couldn't come up with a case where that urn: tag was useful.) 17:21:16 That assumption does not hold true with a tag, since the owner is free to change the meaning whenever. 17:21:36 sbp has quit 17:21:43 I don't want my URNs to change out from under me, especially not on purpose. 17:21:59 a pts should be a commitment to permanence on the part of the minter. 17:23:09 I think persistence should be a quality of an assertion, not an identifier. It doesn't make sense to talk aboutwhether "sandro" is changing, but it does to talk about whether (the truth of) "sandro is alive" is changing. 17:23:50 Huh? It seems it's much more important that Sandro doesn't change, since that breaks all assertions about him. 17:24:28 I can deal with a few assertions going bad, but I can't deal with the resources falling out from under me. 17:24:35 What consititutes a change in Sandro? Every breath I take, I change a little, right? 17:25:05 You change, but the conceptual mapping to your person doesn't. 17:25:18 You understand the difference between resources and entities, no? 17:25:34 Heh heh. Nope. Explain it to me. 17:25:42 Oh, come on. 17:25:51 Here, I explained this to sbp just yesterday... 17:26:08 (It's why we created pts...) 17:26:49 * sandro_ will return in 3 minutes to read what's here. 17:26:55 A resource is a conceptual mapping, like, for example you... 17:27:07 An entity is the physical thing, your body, a bag of bits, etc. 17:27:40 You as a person will stil have lived, even when you body is limp, lifeless, and underneath a pile of earth. 17:28:10 Every time I retrieve the resource Sandro, I get back a different entity, one a little older. 17:28:34 Just as when I retrieve the resource identified by http://www.w3.org/ I get a different bag of bits back. 17:29:04 (Understand that "the resource Sandro" is shorthand for a more complicated explanation.) 17:33:48 * DanC_ works on travel expense report... 17:34:54 So can you identify the "physical thing, your body, a bag of bits, etc" with a URI? 17:35:11 Yes. 17:35:20 And that's a URN? 17:35:47 It could be. 17:36:01 A URN just means that the conceptual mapping doesn't change. 17:36:19 The mapping could be to you as a concept, with a slightly different body every momemnt 17:36:35 or it could be you at any instant, the physical person that exists the second i hit the enter key 17:38:52 the conceptual mapping doesn't change for all URIs, not just URNs. A URN is just a URI that starts with "urn:". 17:39:16 I think the right way to model (formally talk about) all this is with assertions, not with weird semi-formal conventions about the logical symbols themselves. 17:39:36 The conceptual mapping for http://knownow.com/ changed when KnowNow bought it from Instant Knowledge Technologies, inc., no? 17:39:59 yes, that was a protocl error. shit happens. doesn't change the design/specs. 17:40:34 So I think in the SemWeb *all* logical symbols must have a constant mapping to object in the domain of discourse. Considering "http://www.w3.org/" a logical symbol is basically a red herring. 17:41:04 Well, DanC, the tag spec says that the owner is free to change the meaning of the URI at any time. 17:41:08 I don't want that for my URIs. 17:41:30 s/spec/documentation/ 17:41:40 huh? http://www.w3.org/ is a perfectly good logical symbol. you didn't mean to suggest otherwise, did you, sandro? 17:41:43 If the tag spec says that, it's broken. 17:41:56 Right, which is why I don't want to use it. 17:41:58 But when you use ANY URI in a monotonic part of the SemWeb, you better not -- regardless of what URI specs you might be using. 17:43:07 yes, there's a strong social convention that if you publish some assertion about some URI, it'll continue to be true. I'm not sure about forever, though. I've been noodling on reasoning in context. 17:43:21 GabeW has quit 17:43:31 If you take "http://www.w3.org/" as denoting a some kind of service from which you can perhaps gets some bits, then yes, it's a perfectly reasonable symbol for humans, and maybe even for computers. 17:44:30 Dan -- you came in after my stating my basic position: persistence should be a quality of an assertion, not an identifier. It doesn't make sense to talk about whether "sandro" is changing, but it does to talk about whether (the truth of) "sandro is alive" is changing. 17:45:04 So sometimes we'll want to put asserts somewhere, and later change them, sure.... Or let them become false. 17:46:11 DanCon has quit 17:46:26 DanC_ has quit 17:47:18 Obviously DanC_ is not persistent. ;-) 17:47:40 No, DanC_ (the person) is VERY persistent. :-) The text "DanC_" as an identifier here,... well, no. 17:48:05 ;-) 17:51:04 http://lists.xml.org/archives/xml-dev/200108/msg00115.html 17:51:04 C: http://lists.xml.org/archives/xml-dev/200108/msg00115.html from AaronSw 17:51:19 C:|Daniel Veillard on the Semantic Web 17:51:19 titled item C 17:51:34 C:"""After 3 years of rpmfind experience (large catalog/database of mostly 17:51:34 Linux free software [1]), well I don't believe much in a Semantic Web as 17:51:34 commented item C 17:51:34 defined by most people. I do provide an RDF [2] catalog for all the database 17:51:34 content. I think I can try to draw a few lessons from this experiment up to 17:51:35 now:""" 17:52:21 C:Linux free software [1]), well I don't believe much in a Semantic Web as defined by most people. I do provide an RDF [2] catalog for all the database content. I think I can try to draw a few lessons from this experiment up to now:""" 17:52:21 commented item C 17:52:34 Without knowing if you're convinced or not, let me try another angle. You're afraid someone will change the meaning of a tag out from under you. But to use someone else's tag in a way that involves the meaning, you must have studied it extensively -- that would have exposed you to a variety of possible declarations from a variety of sources as to the variability over time and minds of its denotation as a symbol. 17:53:36 I'm not sure about that... 17:53:44 if you say it's a namespace, i want to use it as a namespace. 17:54:57 DanCon has joined #rdfig 17:56:46 What is a XML namespace name except a string of characters? It can no more change than the word "hello" can change. 17:57:26 Again, the application makes the mapping from strings to objects constant by definition. 17:58:36 Your POV is interesting, but it does seem like the tag spec is broken 17:59:18 sbp has joined #rdfig 18:00:19 * sbp catches up on the logs 18:01:00 DanCon has quit 18:01:42 The tag spec is written with the same POV -- that it's the business of protocols specs to talk about the persistance of identifier-to-object bindings, and that tags are useful across many different protocols. I'd very much appreciate suggestions on the wording that might make the spec more clear in this regard. 18:02:45 I think there is a difference between the original denotation of a mapping from identifier to a resource and the interpretation of the semantics of that denotation. That denotation is open to change, of course, just as I can start using other people's URIs for different thing (using the XHTML namespace in place of the XLink namespace perhaps), but there should be some method for people to say what the original denotation is. That's what the authority component ios 18:04:08 I agree with Dan that all URIs should have this property, but the practise is that either they don't, or people don't see them that way. That's the problem that we have in using HTTP for names, in that domain names change hands, and people use them for different things. People either intentionally or unintentionally break the original denotation, and that isn't good. URNs, in being names, mean that there is less of a chance of that denotation being broken 18:04:18 it's all a consequence of the address vs. names debate 18:05:16 it would be nice if domain names could not change hands by force. :-/ 18:05:34 The reason we pushed this through in the first place is that Aaron and I couldnt' agree on a new namespace for a SWAG project. I wanted to use a "#" on the end, as TimBL advises and for similar reasons, but Aaron pointed out the argument to the contrary, which is conceptually valid 18:05:42 persistent domains: if only! 18:06:43 TimBL was here talking about that the other day... I suggeted that DNS is the bane of Web persistence, and he noted that if someone whipped away w3.org from the W3C, there'd probably be a fuss involving people changing their nameserver roots and so on 18:06:49 I wonder.... Are domains ever taken away for reasons other than non-payment and (alleged) trademark infringement? 18:07:07 DanC has his taken away! Ask him! 18:07:14 :-) 18:07:18 I'll get the URI... 18:07:28 Make that alleged non-payment, also 18:07:57 yeah 18:07:57 Uh oh, w3.org is registered with netsol 18:08:03 it's vulnerable too! 18:08:22 here: http://www.w3.org/People/Connolly/9903webtechs-snafu.html 18:08:29 got it: http://www.w3.org/People/Connolly/9903webtechs-snafu 18:08:32 oh, damn you! 18:08:39 Ah. Hm. On the trademark issue I have an angle: use numeric names, like 100.com, since numbers cannot be trademarked. Alas, probably with the .com part they can be..... 18:09:14 You think people haven't trademarked 800.com? 18:10:28 Well, I know Intel could not get trademark protection for are 80486. 18:10:36 try: http://302452351/ 18:10:50 I mean really, go to it in your browsers 18:10:54 sbp, my client doesn't do stupid uris like that 18:10:58 But that's a hard-coded IP address, and you can certain lose those. 18:11:24 Yep 18:11:46 anyway, what were we talking about? :-) 18:12:03 Shall we go ahead with PTS or not? 18:13:02 sbp has quit 18:14:43 * AaronSw wonders why http://valsvc.webtechs.com/ no longer works... 18:16:01 Does PTS do anything for you than urn:tag:.... does not? (except for be a bit simpler)... 18:16:48 Simplicity is a benefit. 18:17:06 (that's why I said "except" :-) 18:17:24 Don't think so. 18:17:34 Here's a tip: registering urn:tag gives you a lot of leaverage? 18:17:40 err s/?/. 18:17:59 ;-) 18:18:02 What leverage? 18:18:38 If you guys followed through and registered one or more of you ideas, we probably wouldn't have proposed pts. 18:20:08 Uh huh. :-) (it is in the pipe, as you can see.) 18:20:27 Yah, but moving slowly... ;-) 18:21:57 It seems unlikely any incompatible changes will be made. Strictly speaking, I don't think there ever has been one. You're probably safe using tags whenever you want. 18:22:26 It's like using DAML or RDF-Schema -- not a standard yet, but usable. 18:22:34 That's not a fair comparison. 18:22:40 sbp has joined #rdfig 18:22:45 RDF isn't a standard either. 18:22:57 BTW, I could ask a similar question: does tag: give you anything that urn:nid-: informal registrations do not? 18:23:01 DAML and RDFS have made commitments that backwards-incomaptible changes will cause the namespace to change. 18:23:05 In the W3C universe it is. 18:23:17 No, it's a Recommendation. 18:23:31 :W3CUniverse :minisculeSubClassOf :TheUniverse . 18:23:39 ;-) 18:23:57 So I feel safe using them. 18:23:58 Some people think a W3C Recommendation is more of a standard than an ISO Standard, at least in some cases. 18:24:03 tag isn't even registered! 18:24:13 Registered where? 18:24:34 with the Internet Community. 18:24:56 It's not on any of the URI scheme lists... 18:25:02 Who owns that trademark? 18:25:30 I don't think it's a trademark, but I was thinking of the IETF. 18:25:32 sbp has quit 18:25:51 ...and IANA, I suppose 18:25:54 And URI's are not an IETF standard, either. 18:26:13 they are in the IETF Universe ;-) 18:26:23 I mean, the URI specification is an IETF specification! 18:26:43 But IETF clearly says it is not yet a "standard". It's only some kind of draft. 18:27:10 Ummm.... 18:27:43 * AaronSw checks 18:27:43 Anyway -- I get your point -- I should be pushing on IETF, IANA, DanC, ..., and that includes registering as a URN at the same time, since that might make it official faster. 18:27:43 yes 18:28:11 2396: (Status: DRAFT STANDARD) 18:29:12 What about HTTP, while you're there.... 18:29:52 2616: (Status: DRAFT STANDARD) 18:30:05 the IETF standard qualification is very hard to achieve 18:30:24 Indeed. :-) 18:31:42 From the W3C Press Release for HTTP: 18:31:45 [[[Draft Standard is the second of the three step IETF standardization process; it recognizes that that HTTP/1.1 is stable, and has multiple interoperable implementations, and that all known technical issues have been resolved in the specification. A Draft Standard is considered to be very close to a final specification, and changes are likely to be made only to solve specific problems encountered. The RFC, based on draft 6 of the specification r 18:32:25 My point is simply that "standardness" involves a lot more that what various organizations publish. That every user has to do some evaluation themselves, involving the real likelyhood that adopting the specificaiton now will cause compatiblity problems in the future. I think tags are pretty stable, by their nature it's very hard for them to be otherwise, so there's little danger in adopting them now. 18:32:48 I'm against unregistered URI schemes. 18:33:10 It's like saying, oh, you can visit my site at lkdl;fsk;dlf.com although I haven't registered the domain name yet. 18:33:20 I'm pretty sure it won't be taken, though! 18:34:29 I understand that political stance. 18:34:45 It's not political. 18:34:49 It's purely pragmatic. 18:35:17 Give me a pragmatic use case. No software is supposed to look at a tag anyway. 18:35:34 Err, what do you mean? 18:35:47 What happens if the IETF registers pts under the tag: name for some reason? 18:36:38 While I think that's highly unlikely, but even than, your rules are close enough to ours that collisions would still not be possible, right? 18:37:13 But would my tags be valid? It's hard to say... 18:38:22 They might be invalid by some spec, but they would still work just fine, since the only thing you can count on doing do with a semantic web identifier (or XML namespace name) is compare it with another on a character-by-character basis. 18:39:23 So the only danger is two parties using the same tag unintentionally. Or maybe the spec might say something about intention and rights to use other's. That's a whole other fuzzy territory....... 18:39:30 * sandro_ has to go cook some food 18:59:22 tdxdave has joined #rdfig 19:04:01 DanC_ has joined #rdfig 19:07:55 danja has joined #rdfig 19:36:31 sbp has joined #rdfig 19:36:44 danja has quit 19:38:13 ambient has quit 19:52:44 danja has joined #rdfig 19:53:59 Ahh, ChanServ 19:55:28 Go ChanServ! 19:55:36 tdxdave has quit 19:57:02 * AaronSw cheers 19:59:43 oh, ChanServ... what went wrong? 20:00:04 * sbp hopes this is captured in the logs accurately 20:00:14 I was just wondering about that. 20:00:42 Nope, it isn't. 20:01:03 Oh well, we'll let people wonder 20:12:18 ambient has joined #rdfig 20:32:34 has bwm been sighted recently? 20:33:10 I suppose he's still on the plane. 20:33:26 ah - of course, thanks. 20:34:15 * danja wondering why Resource in Jena doesn't have a setURI() method 20:38:01 Seth has quit 20:38:04 sbp has quit 20:38:16 jimH has left channel 20:38:25 So, while folks are dropping like flies... 20:38:35 em, can we get a PNG or GIF of the RDF Logo? 20:39:25 or a t-shirt ;-) 20:40:03 Has "the" RDF Logo been chosen? 20:40:20 * AaronSw guesses Sandro hasn't been following the chump ;-) 20:40:34 See: http://blogspace.com/pictures/photo-view?photo_id=4277 20:40:47 * AaronSw his wearing his T-Shirt now 20:41:04 I hope that's not "the" RDF logo. 20:41:18 Well, it's "today's RDF logo". 20:41:27 Why do you hope that, DanC? 20:41:31 I think that was my favorite, so I don't mind. :-) 20:41:49 I liked the empty one better. 20:42:48 sbp has joined #rdfig 20:43:12 why not: no labels on the arcs and no directionality to the arcs. 20:43:49 He he he. 20:43:53 Sounds like a graph, not a logo. :-) 20:43:57 I told that to Eric! 20:43:58 aaron, remind me how I get the RDF stuff about this photo of me? http://blogspace.com/pictures/photo-view?photo_id=4290 20:44:08 Click the "View the RDF" button. 20:44:13 ;-) 20:44:23 And he said "In that case, DanC can just go buzz off" or something to that effect 20:45:19 Hmm... 20:45:41 hmm... photo-rdf says lots of relevant stuff about that photo, but not enough to tell that I'm in it, nor enough to tell where it was taken. 20:45:46 * sandro has updated taguri.org, although it's still not perfectly clear, I suppose. 20:46:18 and :photographer ... no namespace decl. 20:47:01 * AaronSw adds namespace decls.. I seem to have gotten distracted 20:47:06 when i first did them 20:47:29 * sbp is a bit baffled at:- 20:47:40 [[[ In general, it probably makes more sense to talk about the persistence or constancy of statements ("There are about 250,000,000 people living in the US") than about identifiers ("the US"). ]]] - http://www.taguri.org/ 20:48:04 surely the statements 250,000,000 people living in the U.S. depends on using "the U.S." as an identifier? 20:48:13 s/statements/statement 20:48:29 danja has quit 20:48:38 Yeah. I need to work on that some more. 20:48:44 DanC, as I recall location and content was not a requirement... the agreement was merely RDF. 20:50:32 What the the prize/forfit? 20:50:38 true, I wasn't explicit about those. and I can record who/where myself. 20:51:02 odd that the HTML page says "exposure date" but the RDF just says dc:date 20:51:20 * AaronSw isn't sure what exposure date means for a digital camera 20:51:24 so I didn't translate it.. 20:51:26 * DanC_ wonders if the "Add a comment" accepts RDF... 20:52:18 * sbp is still baffled by the "differences from URNs" part in the tag: ID 20:54:04 * AaronSw mutters something about off-the-shelf software... 20:54:06 are universally quantified variables an example of somthing which is "contextual" in that particular part of the draft? 20:55:39 er... existentially, even.. and no 20:56:27 * sandro is too. :-) The test case that generates all that is this: a tag printed on the side of a can of beans. When you scan it in the store, it tells you about its price. When you scan it in your kitchen, it tells you about how to cook it. That means same URI -> different text, depending on the situation. That's absolutely forbidden by the URN spec. But in my mind, the denotation map is still constant, you're just asking >> 20:56:41 context sensative questions about the thing identified by that tag. 20:57:06 Hmm... that's interesting 20:57:55 but it's fair to say :price "2 dollars"; :instructions "open can, empty it, cook contents" . 20:58:07 Yes. 20:58:26 what you're doing is querying information about a persistent node... not changing the details about the node itself 20:59:01 so yeah, I guess it is still a constant denotation... 20:59:09 Indeed. So address that to this usecase. The user types in a URL and gets back different information. So it's not a URN. 20:59:58 that's a deferencing problem, not a denotation problem 21:00:11 That's less clear cut, because, as Aaron points out, you get back a representation of the resource, not the resource itself 21:00:30 When you type in a certain URI, you expect to get back some service of that resource, otherwise you wouldn't have tyoped in the URI 21:00:54 But yes, it can be changed drastically enough to make you interject "damnit, they moved the page again!" 21:02:34 DanC_, you can now add comments in RDF/N3 21:02:42 So that's where the practise appears to diverge from the theory. In that people aren't all that concerned about persistence because it's difficult to keep an address persistent... really, no addresses are persistent, although stuff like the W3C site comes as close as you could get it 21:02:53 danbri has joined #rdfig 21:02:57 comments in N3: woohoo! lemme try that out... 21:03:01 Oh, and the PURL stuff 21:03:08 Sandro, that is not forbidden by the URN spec. 21:03:16 anyone seen libby today? 21:03:23 nope 21:03:35 no worries 21:03:36 bye 21:03:39 danbri has quit 21:03:52 Then what is forbidden by the URN spec, Aaron? What URI is not also a URN? 21:04:03 Changing what a URN denotes. 21:04:11 In both cases of your example, it denotes a can. 21:04:17 but you resolve different information about the can 21:04:45 One day, I use http://infomesh.net/someanimal to denote a rabbit. Next day, I use it to denote a bird. That's not a URN 21:05:00 I use the word "denote" rather loosely there... 21:06:02 Then again, perhaps the denotation there is simply "this page represents some animal" 21:06:33 O.K., so what if the next day it talks about world peace, and then the next day I move the page somewhere else and don't put up a redirect? 21:06:36 I can't seem to log in AaronSw. I have cookies turned on, I think. 21:06:52 Hmm... 21:07:29 so perhaps the denotation then is "the page /someanimal on the infomesh.net server"... 21:07:35 I think it must be cookies... I just logged in fine. What browser are you using? 21:07:41 addresses make tacky names 21:08:12 I like William's: "59th Street Warehouse... where the name is the address!" 21:08:36 But you're just talking about denotation in some internal-to-your-head sense. Use a web example. Tell me a working HTTP URI that is not a URN. 21:09:10 I just gave you one. I can make that example happen 21:09:40 http://localhost 21:09:40 D: http://localhost from ambient 21:09:42 How about http://info.cern.ch/WWW/hypertext/ or whatever it was 21:10:10 it no longer redirects properly: CERN broke it somehow 21:10:25 it changed. Lots of URLs change. They shouldn't do, but they do do 21:11:40 So it doesn't matter which one we choose? Let's pick http://www.cnn.com. That always denotes the web browsing experience one gets from pointing a web browser at that address. That cannot change. So http://www.cnn.com is a URN. Yes? 21:12:33 No... the first owners original denotation was as a news site. Perhaps in ten years, the next owner will be a pine furniture shop or something... the denotation would have changed, in my opinion 21:12:56 and in most people's opinion. Try taking a straw poll 21:13:15 That's just denotation in the fuzzy human mind sense. To a computer, what you get from one day to the next on cnn.com is as different as a furniture store is from the current cnn.com. 21:13:22 AaronSw, I'm using Netscape 4 via wwwoffle. any clues from your logs? 21:13:38 you don't get the denotation, though, sandro. 21:13:39 * AaronSw checks 21:13:47 you get a representation of the state of the denotation. 21:13:57 yeah 21:14:29 * DanC_ despairs of ever getting the community to grok this. 21:16:29 another way of thinking about it is to say that http://www.cnn.com is an object with a method getWebPage(Date date), such that it only makes sense to say what page you get if you also specify a date. The actual object always remains the same 21:16:59 Well, you're just making a new resource out of the data that is returned, the original location, and the date of the dereferening 21:17:23 DanC, I know the bits you get are some kind of representation of knowledge. But I don't know what that has to do with the denotation of the URI as a symbol. As a symbol to humans and computers, the URI seems to denote a browser experience, which change with time and perspective in soft and indescribable ways. 21:17:26 [ :from "http://www.cnn.com"; :date "2001-08-03" ] . 21:18:31 The concrete question here is this: what can software assume about a URN that it cannot assume about other URIs? 21:18:47 how can a URI be used as a denotation of a symbol? The URI is the symbol, the resource is the thing that that symbol denotes 21:19:08 unless you use a name to denote a name... 21:19:17 DanC, I see your machine in the logs, but I don't see it ever POSTing the registration form 21:19:28 Agreed the URI is the symbol. I think you misparsed my sentence or I mistyped it. 21:19:49 assume from a URN: that if it is bound to some particular resource, that resource will not change in such a manner as to break the common contextual use 21:20:03 I probably misparsed it... let me read it again 21:20:40 i.e. assertions involiving only urns always hold? 21:20:47 s/resource/denotation 21:20:56 Yes! 21:21:31 You mean if you type a URN into the location box of a browser, you'll always get the exact same display, regardless of who you are, where you are, or what time it is? 21:21:50 why would you type a URN into the location box of a browser? 21:21:51 no, you always get a representation of the same thing. 21:21:56 no, that's a derefencing issue not a denotation issue 21:22:13 why put a URN in location box: cuz it's a URI. 21:22:34 ugh... I still like to separate names and addreses instinctively, sorry 21:22:43 and the representation you get may very well depend on who you are and/or when you ask. 21:22:52 but don't call it a location box. Call it a resource box :-) 21:22:59 names/addresses: I suggest you get over it. ;-) 21:23:09 address box is my preference. you don't put resources there. 21:23:13 what security settings you have etc. 21:23:20 you ask for a representation of a resource there 21:23:28 So: if you type a URN into the location box of a browser, you'll always get a respresentation of the same thing. The respresentation may vary because of, say, your screen resolution or language preferences. But it will not change because of who you are, where you are, or when you are asking. 21:23:56 it may change because of any sort of information in the request, including who/when. 21:24:32 How is that different from any other URI, then, DanC? (example, please) 21:24:32 the *representation*, yes 21:24:39 not the resource, though 21:24:43 it's not different from any other URI. 21:24:59 (he said, for the 1000000000th time.) 21:25:42 So DanC and sandro are in agreement here. But everyone else seems to think only some URIs are URNs. Yes? 21:25:51 evidently. :-{ 21:26:06 I still think that the denotation changes on some URLs 21:26:11 i don't think http://localhost is a URN 21:26:16 or rather: I agree that only some URIs are URNs. a URN is just a URI that starts with "urn:". 21:26:29 I think that if you asked the majority of people in the streets that same question, they would agree that the denotation has changed 21:26:37 quite. "http://localhost" doesn't start with "urn:". 21:27:10 so why did the IETF come up with URNs? 21:27:12 But that "denotation has changed" notion is too fuzzy. It's like defining beauty. 21:27:28 Why does HTML 2.0 have a urn atribute on the element as well as an href attribute? 21:27:32 that the denotation of some URIs changes is, unfortunately, true, but that's a bug. bugs happen. shit happens, even with "urn:" identifiers. 21:28:21 HTML 2.0 urn: I don't recall any interesting motivation for it. 21:28:37 I don't think it's a *bug*, I think it's a fact of life, a practical fact 21:28:59 Well, it's in there, and the text for the "urn" attribute is quite baffling 21:29:18 Denotation has changed == browser experience has changed more than expected, yes? 21:29:27 well, then, it's a practical fact for "urn:" URIs too. I don't see any guarantee against it. a technical guarantee is impossible, and I don't see any social guarantee (i.e. zillions of dollars in escrow). 21:29:44 more than expected: yeah, something like that. 21:29:57 kind of.. throwing the authors feelings into the mix too 21:30:39 [[[ 21:30:40 URN 21:30:40 specifies a preferred, more persistent identifier for 21:30:40 the head anchor of the hyperlink. 21:30:52 ]]] - urn:ietf:rfc:1866 21:30:56 'more'.. fuzzy 21:31:06 I agree 21:31:18 how do you measure persistence? 21:31:36 Since there is no "browser experience" with a tag-style URI, the only thing you can talk about (outside someone's head) when talking about the "denotation changing" is what assertions people make using that URI. Right? 21:31:44 so linktext is short for a link to "abc" with a { :availableAt } annotation. 21:32:00 or perhaps a link to with that annotation. 21:32:06 neat... 21:32:26 no browser experience: there will be, if they become valuable. I can proxy any URI thru HTTP. 21:33:04 denotation of URI changes -> all previously made asserions involving URI no longer hold (unless the new denotation had occurred sometime before!) 21:33:27 context is everything 21:33:39 no longer hold: thereby creating chaos and devaluating the web. Hence the architectural axiom stating that this shall not happen. 21:33:52 but it does, and it does regularly with URLs 21:33:56 and I think that is the distinction 21:34:18 I like ambient's definition of dynamic denotations 21:34:20 yes, it happens. shit happens. <6% of the time. just like credit card fraud. that doesn't mean the rules aren't what they are. 21:34:36 Pfffff, <6% - reference please 21:34:47 That's still enough to warrant the clause 21:35:04 you doubt me? what will be my reward if I can substantiate the claim? looking it up is not without cost. 21:35:08 if 6% ish of cases are broken, then I'd say that the rule is stupid, or needs rejigging 21:35:23 reward: 250 brownie points back 21:35:58 [I don't think I have anything that you'd value] 21:36:31 "6% of the links on the Web are broken according to a recent survey by Terry Sullivan's All 21:36:32 Things Web. " -- http://www.useit.com/alertbox/980614.html 21:36:48 * sbp hands over the 250... 21:37:05 * DanC_ collects 21:37:48 that 6% is nothing to do with changing denotations, it is to do with the complete lack of denotations 21:37:52 Nielsen evidently subscribes to the principle of unambiguity: "Even worse, linkrot contributes to dissolving the very fabric of the Web: there is a looming 21:37:52 danger that the Web will stop being an interconnected universal hypertext and turn into 21:37:52 a set of isolated info-islands. Anything that reduces the prevalence and usefulness of cross-site 21:37:52 linking is a direct attack on the founding principle of the Web. " 21:38:18 true, ambient, this is not exactly the failure mode we were discussing. but it's related. 21:39:12 relevant axiom: 2a: sameness. http://www.w3.org/DesignIssues/Axioms.html#same 21:39:49 sbp has quit 21:40:16 sbp has joined #rdfig 21:40:44 sorry 'bout that... 21:42:10 O.K., well I'd be happy if the tag: Internet Draft got rid of the stuff about temporary and persistent denotations rubbish, and just left that up to these "axioms of Web architecture", and 2) if it would actually get to RFC status 21:42:34 Dan, is the TAG meant to be clearing up stuff such as this? 21:42:44 re 6%: there's a big difference between 6% of links are broken vs. 6% of access requests fail. most attempts won't be near the broken links. so the reliability of the web is likely much higher than 94%. 21:43:03 yes, the TAG is chartered to clarify this stuff. I give it about 45% odds of success. 21:43:30 O.K., but remember how many reqests there are per day... 0.01% even of that amount is a big number still 21:43:42 45%: I'd give it more, depending on who the members are going to be 21:44:29 And I take it that the charter is active, so I can propose this to www-tag@w3.org as an agenda item... 21:46:36 * sbp wonders what liason powers W3C (and hence TAG) has with the IETF 21:48:08 ah, it's a TAG role: "establishing liaisons with groups outside of W3C involved in the development of Web architecture" - http://www.w3.org/2001/07/19-tag 21:50:43 more on the axiom of sameness: "Web Pages Must Live Forever" http://www.useit.com/alertbox/981129.html 21:51:43 the best page really for the sameness axiom is "Cool URIs Don't Change"... but because of that, I still don't get the point of the { :availableAt } annotation in HTML 2.0 21:54:35 I'm not sure I get the point either, but... 21:54:49 imagine you have a loose federation of mirror sites for some photo archive. 21:55:31 say http://photo.example/2001/phot1221 is mirrored at a zillion places, but you don't have the horsepower to do the DNS stuff to make it invisible 21:56:20 so you might make links to http://mycopy.example/photo-archive-mirror/2001/phot1221 with an annotation that this is really a copy of the thing at the above address... 21:56:53 and a cache might be smart enough to know that if { :x :availableAt :y, :z } and it has a fresh copy of :y, that can be used to service a request for :z. 21:57:10 Is that telling a machine what the URL denotes, or is it just processing information? 21:57:36 certainly not telling (everything about) what the URL denotes... 21:57:52 just giving some aliasing hints. 21:58:27 Hmm... it's just added context... much like linktext means { xhtml:title "def" } 21:58:31 RDF for cache info, mirroring, aliasing, etc. is long overdue. 21:58:43 added context: quite. 21:58:53 if RDF had have been around in the early days of the Web... 21:59:02 * DanC_ hopes to get wwwoffle to grok RDF 21:59:03 Can someone give me an example (for taguri.org) of a URL that clearly, publically, changed resources? 21:59:23 http://knownow.com/ ; see above about ownership change, sandro. 21:59:23 E: http://knownow.com/ from DanC_ 21:59:32 E:|oops 21:59:32 titled item E 22:00:24 * DanC_ gets back to expense report... 22:01:18 Sandro - could you possibly give me a guesstimate of RFC status for tag:? 22:02:09 s/of RFC/of time until RFC 22:02:23 I'd like to think a month, but I have no clue, really. 22:02:43 How long has it been an internet draft for? 22:02:45 And of course that's purely informational RFC, whatever that means. 22:03:01 Since July 7, I think. 22:03:03 If it's more than two weeks, with no feedback, you can submit it for an informational RFC right now 22:03:13 That's long enough... hang on a sec 22:03:19 Oh -- I've been meaning to look up the proceedure. 22:04:01 4.2.2 Informational in RFC 2026 22:04:07 - http://www.ietf.org/rfc/rfc2026.txt 22:04:34 er... 4.2.3 is more relavent :-) 22:04:50 [[[ 22:04:50 documents 22:04:51 intended to be published with Experimental or Informational status 22:04:51 should be submitted directly to the RFC Editor. The RFC Editor will 22:04:51 publish any such documents as Internet-Drafts which have not already 22:04:51 been so published. 22:04:53 ]]] 22:05:14 [...] 22:05:15 The RFC Editor will wait two weeks after this 22:05:15 publication for comments before proceeding further. 22:06:25 Well, it's not in the queue: http://www.rfc-editor.org/queue.html 22:07:04 it then goes on to sec. 6.1.1 22:07:07 It's been two weeks so you need to email rfc-editor@rfc-editor.org 22:07:18 According to http://www.rfc-editor.org/howtopub.html 22:11:22 So send it! Send it now! :-) 22:16:53 Thanks for looking that up. 22:17:04 Send it now? I thought you were arguing for making a few changes! :-) 22:17:21 Don't listen to Sean... 22:19:21 * sandro_ has revised the URN bit on taguri.org and is now much happier with it. Sean? 22:21:23 Yep, it seems fine, although I'm not at all concerned about the site, just the specification 22:23:04 I really think that this should be an RFC already 22:23:09 * sandro_ is loathe to try to make this kind of change to the spec. I think the spec is neutral on this, but I guess I should review it carefully. Maybe Tim Kindberg will be happy to do it. :-) 22:24:26 So you vote is: CHANGE-FIRST or SUBMIT-NOW. 22:24:29 I don't think it is neutral about it at all! It says that tags cannot possibly be URNs because tags can have temporary denotations... and since we've decided that there's no difference between URNs and URIs other than the string "urn:" on the front 22:24:48 CHANGE-FIRST 22:25:04 vote... definitely CHANGE-FIRST, with the added comment [SUBMIT-DAMN-SOON] 22:25:07 :-) 22:26:09 Wanna send me a diff on the i-d? :-) 22:27:17 just whip out the bit starting "URNs [RFC2141] are intended " would probably do it... unless there's more stuff hiding in there 22:27:52 er... where did "t1" come from? 22:28:26 Oh yeah......! I forget about that. 22:28:48 :-) 22:28:57 So what's that all about? 22:29:00 We changed the abbreviated forms when we introduced the long form, so there would be no conflict in 2000 years. 22:29:42 good point... 22:29:54 the URIChar production doesn't exist 22:29:56 it's uric 22:30:15 s/specific ::= *(URIchars) ; [RFC 2396] 22:30:39 - /specific ::= *(uric) ; [RFC 2396] 22:33:07 and I can't have month long authority components? 22:33:46 example? 22:34:11 Do we need to wait another two weeks after making the change? 22:34:13 tag:infomesh.net,2001-07:mytag 22:34:45 wait another two weeks: there are a few places in the process where you can make changes... 22:36:00 It's probably best to wait two weeks... 22:36:06 No, I think not. What would it mean? The same as -01 ? That's another extra bit.... 22:36:35 Hmm... 9) in http://www.rfc-editor.org/rfcfaq.html but it only mentions typos 22:36:58 This isn't really a typo/misspelling. 22:37:10 what would it mean: either that I've held that domain for the whole month, or, like the abbreviated thing, that it means the first day of the month 22:37:39 Yeah, it's a change... aaaaargh! So another two damn weeks 22:37:39 And you want it for http-like-ness? 22:38:10 http-like-ness? no, just brevity 22:38:10 I can store files in /2001/08/03/ for example 22:38:26 Well, dropping a non-nomative paragraph about URNs, and urichar -> uric; those dont seem substantian. 22:39:29 If you want previty, use t1-8, if you want clarity use 2001-08-01. Okay? 22:41:36 karld has quit 22:41:48 sbp has quit 22:42:37 sbp has joined #rdfig 22:43:47 "If you want previty, use t1-8, if you want clarity use 2001-08-01. Okay?" - fine 22:44:05 what does "t" stand for? 22:44:17 cool. have you reviewed the rest of it? 22:45:38 t for tag. we're open for other ideas. We were toying with "+" but it looks too much like a math expression. 22:46:00 so you did make a backwards compatible change! aha! 22:46:24 tut tut :-) 22:46:42 Well, not really, Aaron, since we say how anything outside this syntax must be treated. 22:47:01 You do? that seems a bit odd 22:47:14 so how must tag:infomesh.net,1-5-28:blackbird be treated? 22:47:20 namely you just compare bytes, like always! 22:47:54 tag:infomesh.net,1-5:blackbird isn't a valid tag anymore thouh, according to the BNF... 22:48:30 Oh, and perhaps you should add support for IRIs? 22:48:45 Its outside the grammar, but all software which sees tag: must still do the right thing. 22:49:07 what is the "right thing"? 22:49:45 * sandro_ ids ignorant of IRIs. 22:49:46 right thing == compare characters for equality. 22:49:52 [[[ 22:49:53 Each disallowed character is converted to UTF-8 (as defined in RFC 2279) as one or more bytes. 22:49:53 Any bytes corresponding to a disallowed character are escaped with the URI escaping mechanism (that is, converted to %HH, where HH is the hexadecimal notation of the byte value). 22:49:53 The original character is replaced by the resulting character sequence. 22:49:54 ]]] 22:50:47 simple really, and it makes tags i18n-ized 22:52:58 So why aren't tags URNs? 22:53:18 because URIs are URNs 22:53:28 er... well... 22:53:30 No they're not. 22:53:42 * AaronSw thinks pts should be renames stag. 22:53:45 except they don't have the "urn:" bit on the beginning :-) 22:53:45 ;-) 22:53:55 ranames? stag? 22:53:59 renames, even 22:54:06 renamed, rather 22:54:13 stag = simple tag 22:54:19 I think that's wrong. Tags are URIs and URIs dont have, eg " " in them. You can put a %20 if you want, but that's different. 22:54:22 it's a funny name, and it's easier to pronounce. 22:54:49 you think what's wrong? 22:55:02 your IRI bit 22:55:22 Tell Duerst that 22:55:46 But first, try using a Japanese charater in a URI! 22:56:14 So you can have a space in a URI? 22:56:31 %20 22:56:50 Sandro, the tag spec gives the example tag:myIDs.com,t1:TimKindberg/doc.101 22:56:55 but that's not legal according to the BNF 22:57:01 (i.e. / is not allowed) 22:57:10 yes, "/" is allowed 22:57:16 uric* 22:57:23 well, since the URIChars isn't specified it's hard to tell ;-) 22:57:35 that's a typo :-) 22:57:41 II has joined #rdfig 22:57:47 Oh, it's uric -- ok 22:57:50 also, why is it using ::= a la ABNF rather than = as in BNF? 22:58:12 Aaron, "So why aren't tags URNs?" Assuming you mean in the DanC sense, because (1) that's four wasted and confusing characters, and (2) some people will think it means URN in the sense of same-content-web-pages. 22:58:28 I'd rather it followed the example of not having reserved characters in the component, apart from perhaps ":" as a hierarchy delimiter 22:58:33 I'm not sure what that sense is. 22:58:43 read the logs 22:58:49 I read the logs. 22:59:00 The DanC sense == a URN is a URI that starts with "urn:" 22:59:07 cool, but have you read them? 22:59:14 err, I meant the "sense of same-content-web-pages" 22:59:24 I think that's a rare mistake. 22:59:42 Since urn:pin:1 -- one of the first URNs -- refers to Michael Mealing 22:59:52 Tags are not same-content-web-pages URIs because of the can-of-beans example. 23:00:17 But I think you're the only person who thinks that's what URNs mean. ;-) 23:00:18 neither are URNs; the can of beans example can be a URN 23:00:45 As I said, there are two meanings of URN! 23:01:03 Err, the meaning of URN is well-defined. 23:01:39 if there were two meanings of URNs, they wouldn't be holding up to their task as URNs or URIs :-) 23:01:51 Certainly not between Dan Connolly and Mike Mealling. If they disagree deeply, how can you say it's well-defined? 23:02:31 I don't see what you mean. 23:02:43 I think that URNs represent a better quality of service, and than DanC and MM disagree on where the line is 23:02:55 i.e. DanC says there is no line :-) 23:02:58 They both agree that they are URIs that start with the characters 'urn:' to be used as "s persistent, location-independent, resource identifiers" 23:03:20 What do you mean? DanC drew the line! cf http://www.w3.org/Addressing/ 23:03:30 I think DanC thinks the standards are the same for all URIs, URN or not. 23:03:40 yeah 23:03:46 But that's not the issue, sandro 23:03:55 the issue is what is a URN, not what are the standards for URIs. 23:04:03 and DanC and MM agree on what a URN is. 23:04:10 URN: "An URI that has an institutional commitment to persistence, availability, etc." 23:04:10 Label URN not found. 23:04:16 grpmh 23:04:42 The issue is what you learn from those four characters. DanC says "nothing", and I'm inclined to agree until someone gives me an example. 23:05:02 You make people happy, Sandro. What is worth more than that? 23:05:03 you learn that that URI has an institution commitment... 23:05:19 what I can learn ,or what a machine can learn? 23:05:20 s/institution/institutional 23:05:31 what a machine can learn. thanks. 23:05:57 ambient, you learn that the person who created "has an institutional commitment to persistence, availability, etc." 23:06:02 I don't understand your "happy" comment, Aaron. 23:06:05 a machine can think "oh, I don't need to look for a persistence policy for this URI"... how about that? 23:06:32 The can-of-beans example will fail that test. 23:06:38 what??? 23:06:58 a can of beans is a can of beans: I could maintain that quite easily 23:07:04 For four characters you can make me and the URN folks happy, Sandro. That seems like a good trade to me. 23:07:26 I could write a certificate to say that I declare this URN to identify this can of beans throughout time... 23:07:33 But it'll make TimBL and DanC and about 7 billion people over the next 100 years unhappy. :-) 23:07:38 the can of beans must be represented by a URI, since I can meltdown the can - the can no longer exists and the persistence test has failed 23:07:55 sandro, why will it make them unhappy? 23:07:58 no, the concept of that can still exists 23:08:05 read the URN RFC 23:08:42 Because they wont be bothered by four useless characters all the time. 23:09:18 1) they're not useless, 2) TimBL hasn't died from typing http:// all these years... 23:09:43 But no one's happy about typing http:// all these years! 23:09:59 * sbp wonders where he read that bit about resources not being available anymore 23:10:03 But they're not that unhappy. 23:10:13 That's right sbp, we're sold out of them. 23:10:33 What do they really tell me? They are a HINT about persistence policy, which cannot be defined across applications. In an application like RDF you can define a far, far more precise and clear system of persistence policies. 23:11:00 RFC 2611:- 23:11:01 [[[ 23:11:02 Although non-reassignment of URN identifiers ensures that a URN 23:11:02 will persist in identifying a particular resource even after the 23:11:02 "lifetime of the resource", some consideration should be given to 23:11:02 the persistence of the usability of the URN. This is particularly 23:11:04 important in the case of URN namespaces providing global 23:11:06 resolution. 23:11:08 ]]] 23:11:12 I don't disagree, but those four character should inspire a commitment, a sense of duty and loyaly, a sense of importance, a question of -- am I really going to make this URI persistent? 23:11:33 RDF should inspire that even more...... 23:11:35 all URIs should do that, though 23:11:40 Yes. 23:11:47 But they don't. 23:11:58 RDF is what drives us to tags in the first place, and the stupidity of HTTP for names 23:12:03 RDF isn't part of the URI -- it's out-of-band. 23:12:15 * sandro_ is being dragged away, kicking and laughing from this argument to play Myst III with his family. 23:12:23 RDF is a good application for URIs... 23:12:31 Myst III is fun, so I'm not too upset. ;-) 23:12:33 Myst III: cool. Have fun! 23:13:22 And it SHOULD be out of band. If you get the URI without any context, you should not be assuming just by the leading "urn:" that everything you learned about it yesterday is everything you can learn about it today. 23:13:37 I never said that! 23:13:49 Put it this way: I'm not upset about PTS using a "urn:" scheme. I'm not even too bothered that "tag:" doesn't use one. I am bothered that people can't see that the persistence w.r.t. URIs depends often upon the protocol and scheme being used 23:13:58 ideals are just ideals, not practice 23:14:18 Sean, talk to Aaron. I'll be back (eventually). :-) 23:14:24 sandro_ has quit 23:14:26 O.K. :-) 23:14:42 Is the reason for RFC2611 stating that URNs cannot be reassigned that URIs can be reassigned? 23:14:49 I'll be back Sunday, so see you all then. 23:14:58 s/URIs/non-URNs 23:14:59 I'm not sure: that's a good question 23:15:00 (Or late Saturday night...) 23:15:05 Have a nice Sunday! 23:15:10 er... Saturday 23:15:12 :-) 23:15:16 er... both! 23:15:44 Perhaps we should be asking these questions on uri@w3.org as well... 23:16:09 i think i'm way out of my depth as it is ;-) 23:16:16 I think that the practice is that, yes, URIs other than URNs are often reassigned 23:16:37 not at all, you've made a whole host of comments more pertinent and insightful than my own 23:16:41 then what sense can you make of RDF assertions that use non-URNs? 23:17:00 I suppose that depends upon context... 23:17:05 Here's an example 23:17:06 exactly 23:17:37 at the moment, I want to create a new namespace for a set of terms that I want to use in my documentation. Let's say I want a URI for the predicate "countryOfOrigin" 23:18:28 Now, I own the domain infomesh.net - so you would have thought that it would make sense to put that predicate there. Ignoring what HTTP can retrieve, let's put it at http://infomesh.net/2001/xountryOfOrigin 23:18:32 s/x/c 23:19:38 Now, I'm pretty poor. There's a good chance I won't be able to afford infomesh.net when it comes up for renewal, or I may just choose to spend the money on something else. So, now the new owner decides to make a competing term at that same URI, and issues a load of software to back it up 23:20:12 clearly, software that I create that uses "my" term as a built in, and software that uses the "new" term as a built in cannot interact: I think that's a given 23:20:41 So, HTTP, due to the instability of the DNS system, cannot guarantee persistence *within an SW context*, or at least, not very often 23:20:52 i think this a problems with URLs rather than URIs in general. 23:21:09 I like Freenet URIs, they're based on content hashes 23:21:09 O.K., but then you're saying that URLs have a problem that URNs don't 23:21:16 Yeah, they're cool 23:21:40 URLs are based on DNS, therefore they have a problem 23:21:47 yes 23:21:53 URNs aren't based on DNS, they have no problem 23:22:25 well, its more of a grade of a problem, but yeah, sure 23:23:26 I don't think it's a boolean "URLs bad, URNs good", but I think the situation is much like linking to Webpages 23:23:35 linkrot is a real problem on the WWW. It's something that we can aviod on the SW, because we're not all that concerned about locations! We're concerned about names usually 23:24:07 so we can aviod linkrot by switiching to schemes which are more stable, persistence-through-context wise than URLs, and other schemes that use the DNS system 23:24:14 but remember what happened when the RSS schema disappeared off netscape.com 23:24:18 s/linkrot/termrot 23:24:19 tdxdave has joined #rdfig 23:24:25 what happened? 23:24:47 some RSS software broke because they couldn't retrieve the RSS definition 23:25:18 Well, remember the urn in HTML 2.0 thing? :availableAt , . 23:25:25 then use the URN 23:25:34 rather:- 23:25:42 s/availableAt/schemaAvailableAt 23:25:53 or even, rdfs:isDefinedBy, or whatever 23:26:01 No, that wasn't the issue. 23:26:32 It was a DTD, and many DTD-aware tools couldn't parse the document without retrieving the DTD. 23:26:42 So a URN wouldn 23:26:46 So a URN wouldn't really help. 23:26:55 well, DTDs already have the PUBLIC identifier, people should have used that 23:27:08 I know, it's a case of bad tools, IMO. 23:27:38 II has left channel 23:28:40 it's odd though: TimBL says that you should have a name *and* an address for something 23:31:06 in practise, I think URNs should make you think "persistence" in much the same way as an http://www.w3.org/xxxx/ prefix should 23:32:12 the problem is that it's not boolean... if it were, then we could just say "URLs are not persistent, URNs are". All we can really say is "in theory, all URIs should be fully persistent, but in practice, the quality of service for URNs is most certainly higher than that of URLs" 23:34:23 danja has joined #rdfig 23:35:39 anyone had comm probs with uk tonight? 23:36:15 it's still here as far as i can see :-) 23:36:29 ouch - just my isp 23:36:32 I've had big comm problems, but I found a way around it 23:36:49 what was that? 23:36:52 Simpyl open up five different mIRC windows (or whatever), and connect to five different servers 23:37:08 sbp^5 23:37:09 I use irc., carter., benford., forward., and sagan. usually 23:37:28 no, because then you simply use the first one to get onto the servers 23:37:42 you can never have two clients with the same nick on one server 23:37:53 I've not had to think that far 23:38:03 prob wasn't IRC 23:38:08 Oh... 23:38:17 coupla uk sites 23:38:25 and mail (via uk) 23:39:13 btw sbp - is there a faq on mirc server probs? 23:40:11 never run one meself, so it looks silly when it does its silly stuff 23:42:08 * danja no offense to opn intended 23:43:30 sbp has quit 23:49:51 ambient has quit 23:53:28 * danja remembers he's using the same ISP as sbp... 23:55:17 karld has joined #rdfig 23:55:36 hi karld - what's new? 23:57:28 not much 23:57:36 how bout with you? 23:58:38 i am trying to follow this discussion, but I am new to RDF