#rdf-uri log

This is a log for the #rdf-uri channel, starting at Tue Mar 27 22:22:16 2001 UTC.

bot $Revision: 1.1 $ of $Date: 2000/11/08 19:26:46 $ by Dan Mar 2000

22:22 sbp: hey, scribey!
22:22 AaronSw: so what fancy things does scribey do?
22:22 DanC: so... the command line is:
22:23 DanC: python scribe-bot.py scribey '#rdf-uri' ~/public_html/2001/ http://208.190.202.227/~connolly/2001 irc.openprojects.net
22:23 AaronSw: dajobe, your html should be more like scribey's
22:23 sbp: Right, I'll give it a whirl - better use a different room...
22:23 sbp: Hmm, the HTML is odd:
22:24 sbp: it keeps bunting the > to the next line
22:24 dajobe: I refer you to my 1st XSLT script note I made previously
22:24 AaronSw: yeah, i know
22:24 dajobe: looking...
22:25 dajobe: so is there an XSLT for something related to URI schemes then?
22:26 AaronSw: yes, to make the schems-gen page
22:26 dajobe: what's the route? n3 -> RDF/XML -> XHTML?
22:27 dajobe: hmm, cwm.py I have barfs on schema.n3 - raise internalError # Strings should not match if not same object
22:27 AaronSw: yeah, think so -- look at the Makefile
22:27 AaronSw: danc did some messing with cwm
22:28 dajobe: great!
22:28 DanC: I've got a fix for that, dajobe
22:29 dajobe: got new version from http://www.w3.org/2000/10/swap/cwm.py
22:29 dajobe: and got results
22:30 sbp: yeah, I had the same problem
22:30 sbp: the old CWM had a bug
22:31 dajobe: new version is CVS/RS $Id: 1.30
22:32 sbp: Yep, that's the latest: I have some logs at http://infomesh.net/2001/03/cwmarchives/
22:33 sbp: Correction: http://infomesh.net/2001/03/cwmarchive/
22:37 sbp: Does anyone have CWM v1.28 to complete my collection? :-)
22:40 sbp: Let's see... stuff for schemes.n3 - well, we need consistency on author names: Tim is listed as TimBL and T. Berners-Lee and so on
22:41 sbp: Why not use an arbitrary URI (e.g. a homepage, a mail address) and then hang stuff off of that?
22:41 dajobe: yeah! sablotron moaned about stylesheet so hacked it and now got it running server side from RDF+XSLT->HTML
22:43 DanC: scribey groks links ala:
22:43 sbp: Yes...
22:43 DanC: -> http://www.w3.org/Addressing/schemes WWW Addressing schemes
22:43 DanC: scribey also tracks agenda requests:
22:44 DanC: scribey, agenda request: splitting titles out of schemes.n3 and merging back in from rfc-index.n3
22:44 DanC: oops
22:44 DanC: scribe, agenda request: splitting titles out of schemes.n3 and merging back in from rfc-index.n3
22:44 dajobe: hmm
22:44 dajobe: logger_1, hello
22:45 dajobe: excellent
22:45 DanC: scribey marks up topic changes in the log:

URIs in the Semantic Web

Tue Mar 27 22:45:22 2001 UTC

22:45 AaronSw: neato
22:45 AaronSw: where do agenda requests go?
22:45 sbp: Did you see the SWAG post about URI stuff to look into?
22:45 DanC: they go into scribey's head.
22:46 DanC: scribe, agenda summary?
22:46 AaronSw: i see -- they're not in the html?
22:46 DanC: er... no, I don't think so.
22:46 sbp: Hey great, the logs get a new heading for a new topic...
22:46 DanC: if you're clever and look at the source of the logs, you can link to the headings:
22:47 DanC: -> #t985733122 the earlier topic change
22:47 sbp: name="t985733122"
22:47 sbp: ugh! Use id=""

Scribey Tips and Tricks

Tue Mar 27 22:47:26 2001 UTC

22:47 AaronSw: my favorite part is the easter egg
22:47 AaronSw: spam
22:47 DanC: I fought a war with the HTML WG about name vs. id; I lost; that doesn't mean they're right, though. ARGH!
22:48 DanC: whoah! an ester eg in scribey!
22:48 sbp: Yeah, don;t get me starte don the HTML WG!!!
22:48 sbp: Still, what's so great about name?
22:48 AaronSw: danc, you sounds surprised ;-)
22:49 DanC: the easter egg is from code I didn't write.
22:49 sbp: Ah, just got it:-
22:49 sbp: def spam(self, sender, recip, forum, addl, raw):
22:49 sbp: forum.tell("Spam, spam, spam, spam, Spammity Spam! Wonderful SPAM!")
22:49 AaronSw: ahh, i see
22:49 DanC: what's great about <a name>? how about 10 years of existing practice and deployed software?
22:50 sbp: But isn't ID more widely recognized as being "the FragID bit"?
22:50 DanC: what if the DRUMS WG decided to change, say, Message-Id to Msg-Identifier for no particular reason?
22:50 DanC: no, I've got a lot more software that groks <a name> than <a id>.
22:50 AaronSw: so should we change xpointer to use name?
22:50 sbp: If the DRUMS WG were the HTML WG, they'd get their way whatever screwball idea they came up with
22:51 DanC: yes, we should have changed xpointer to use name. ARGH!
22:51 AaronSw: and xml too?
22:51 sbp: And XHTML Basic?
22:51 DanC: no change to XML is required.
22:51 AaronSw: isn't there a thing about IDs in there -- oh wait, it's in the dtd, isn't it
22:51 DanC: all the HTML WG had to do was to say "the name attribute on the a element has type ID". Then xpointer would work just fine.
22:51 AaronSw: ahh, i see. yes, you're right then
22:52 sbp: Hmm... I hadn't actually thought about it before
22:52 sbp: What was the WGs excuse?
22:53 DanC: The WG's excuse is actually better than I'm letting on... there's a complex interaction between CSS selectors and XHTML element ID's.
22:53 sbp: a#id
22:54 DanC: basically, TimBL admitted that he should have picked 'id' in the first place, so there wasn't a really good architectural argument for keeping <a name>.
22:54 sbp: So it's just a mater of waiting for implementations to change over then...
22:54 sbp: and most *do* grok ID
22:55 sbp: still, it doesn't matter anymore - text/html is dead as far as extensibility is concerned, always has been. Validation is a mess, and even XML didn't fix it, because of the inflexibility en system
22:56 DanC: er... clearly, both of us could go on at length about XHTML and such; but let's not, at least not in this channel.
22:56 sbp: good point of order...
22:56 sbp: But one thing we could discuss is the namespace: http://www.w3.org/1999/xhtml
22:56 sbp: *** dajobe has left #rdf-uri (night all)
22:56 sbp: no hash on the end y'see
22:56 DanC: ah.
22:57 DanC: stop, stop, stop, sbp; no fair opening up *all* the old wounds at once.
22:57 sbp: heh!

Getting the URI scheme index into the Semantic Web

Tue Mar 27 22:57:31 2001 UTC

22:57 sbp: There was some talk about changing the RDF concatenation system to add a hash to any URI not ending with a / or #: *big* mistake IMO
22:58 sbp: why break existing tools and schemas to cover up others mistakes?
22:58 AaronSw: ok, first parse: -> http://blogspace.com/rdf/iana-uri-schemes a quick parse of the uri scheme list
22:59 DanC: er... the -> gizmo is only recognized at the beginning of the line, I think.
22:59 AaronSw: oh, ok
22:59 sbp: Aaron: N3?
22:59 AaronSw: it's coming
23:00 DanC: -> http://blogspace.com/rdf/iana-uri-schemes a quick parse of the uri scheme list, by AaronSw
23:01 sbp: Nah, I already got 1000 for schemes.n3
23:01 sbp: Plus Aaron will deserve wuite a few if he creates a PUT Wiki
23:01 sbp: s/wuite/quite
23:02 DanC: AaronSw, is that thing generated by machine from the IANA registry? if so, please make a dc:source link back to that registry (and, if at all possible, a link to the source code of the transformation)
23:02 DanC: I wonder if I can get cwm to exploit the fact that inet:rfc a daml:UnambiguousProperty.
23:03 sbp: Um, is this link in the scribey source allowed? ./2001-03-27T222216Z-%23rdf-uri.html#t985733122
23:03 sbp: oh, hang on, what am I saying...
23:04 sbp: IE6 doesn't keep the URI escaped when you hover it!
23:04 sbp: grr...
23:04 DanC: -> http://dev.w3.org/cvsweb/2000/scribe-bot/ scribe-bot sources
23:05 DanC: (another cool scribey feature: it splits the log every 12 hours, making links between segments)
23:06 DanC: the logger grep feature is way cool, though.
23:06 sbp: demo?
23:06 DanC: logger_1, grep daml
23:06 logger_1: I'm logging. I found 1 answer
23:06 logger_1: 0) 2001-03-27 23:02:58 <DanC> I wonder if I can get cwm to exploit the fact that inet:rfc a daml:UnambiguousProperty.
23:07 sbp: Oh, I thought you had one on scribey
23:07 DanC: nope.
23:08 sbp: Does scribey have an [off] equivalent?
23:08 DanC: no.
23:08 sbp: aw...
23:08 DanC: SMOP, though, no?
23:08 sbp: Well, here is a demo of what I'm doing recently... not a Cool URI: http://ch8as13-77-168-136.cw-visp.com/infomesh/2001/sw/sbp/default.html
23:09 sbp: fairly small
23:09 sbp: http://info.astrian.net/jargon/terms/s/SMOP.html
23:09 AaronSw: you all can make fun of my n3 now
23:10 DanC: scribye RFE: google integration
23:11 DanC: ala:
23:11 DanC: google, SMOP?
23:11 sbp: Aaron: where are your "."'s?
23:11 sbp: google is a search engine
23:11 sbp: SMOP is an acronym
23:11 sbp: http://blogspace.com/rdf/iana-uri-schemes
23:11 AaronSw: sbp, fixed
23:11 DanC: oh. duh.
23:12 sbp: and you forgot the ;
23:12 DanC: AaronSw, you've got the right raw materials, but it's a little jumbled.
23:13 sbp: after the dc:title
23:13 DanC: e.g. [ :schemeName "ftp" ]
23:13 DanC: dc:title "File Transfer Protocol"
23:13 DanC: :specifiedBy [ inet:rfc "1738" ] .
23:13 DanC: should be: [ :schemeName "ftp" ] :specifiedBy [ inet:rfc "1738"; dc:title "File Transfer Protocol" ].
23:13 sbp: scribe, agenda request: log "/me" commands
23:13 AaronSw: danc, but it's not the name of the rfc
23:14 DanC: huh?
23:14 AaronSw: it's the description of the scheme
23:14 sbp: anon rfc x; title FTP
23:14 AaronSw: danc, iana calls it a description
23:15 DanC: in English: "Something with schemeName "ftp" is specified by something with dublin core title 'File Transfer Protocol' and internet RFC number 1738".
23:15 DanC: oops. I see.
23:15 sbp: Exactly, that "something"
23:15 sbp: oh
23:15 AaronSw: rfc-index can link in the rfc names
23:15 AaronSw: i won
23:15 AaronSw: 'tbother to do the extra parsing
23:16 DanC: then dc:title is the wrong property; use dc:description or rdfs:comment
23:16 AaronSw: i switched it to dc:description
23:16 sbp: he is
23:16 sbp: http://blogspace.com/rdf/iana-uri-schemes
23:17 sbp: I noted that when my old broser cached something, all I had to do was add a query string,
23:17 DanC: woohoo!
23:17 AaronSw: it also hase "reserved url scheme names" should i include those?
23:17 sbp: a la http://blogspace.com/rdf/iana-uri-schemes?a
23:17 sbp: and it would refresh it for me... odd hack, but it worked
23:17 AaronSw: why the woohoo?
23:17 DanC: pls replace :specifiedBy with is :specifies of; i.e. please hold off on using daml:inverseOf, OK?
23:18 AaronSw: oh, i can do that? awesome
23:18 DanC: woohoo: cuz the IANA registry of URI schemes is now in the semantic web!
23:18 AaronSw: heh woohoo!
23:18 sbp: yep, N3 can go backwards...
23:18 sbp: but sometimes the grammar doesn't makes sense
23:18 sbp: it would be great if:-
23:18 sbp: a = an .
23:18 sbp: and so on
23:19 sbp: then :mySub a earl:thing . could be :mySub an earl:thing .
23:19 DanC: the conceptual graph folks went nutso with a = an and the like. TimBL's thinking about it. I think rdf:type is not so much longer than a, some days.
23:19 AaronSw: iana also hase "reserved url scheme names" should i include those?
23:19 DanC: reserved: yes, please.
23:20 DanC: i.e. :ReservedURISchemeName rdfs:subClassOf rdfs:Literal.
23:20 sbp: Hmm, so all it takes are three bored SW folk to get one important page up on the SW in a few minutes? That's not too bad I suppose
23:21 sbp: Hmph, it's half midnight here
23:22 sbp: You've still got the rest of the day to hack about!
23:22 DanC: you could get rid of the awkwardness of _is :specifies of_ by just writing the sentence the other way around: [ inet:rfc "xxx" ] :specifies [ :schemeName "foo"; dc:description "a hoopy frood"].
23:22 sbp: oh yeah, duh
23:22 DanC: I'll probably be back after putting the kids to bed; but 6pCT to 10pCT is family time.
23:23 AaronSw: dan, do the reserved ones look alright now?
23:23 AaronSw: wait a sec...
23:23 DanC: tn3270 is goofy
23:24 sbp: Aaron, needs """
23:24 sbp: because it spreads over multi-lines
23:25 sbp: oh, it just messed up
23:25 AaronSw: i
23:25 AaronSw: 'm working on it
23:25 DanC: I was adding a pointer to Aaron's thingy in schemes.n3, and I noticed:
23:25 DanC: <ftp://ftp.isi.edu/in-notes/iana/assignments/url-schemes>
23:25 DanC: dc:title "Internet Standard list of url schemes";
23:25 DanC: :reserves
23:25 DanC: [ :schemeName "afs" ]
23:26 DanC: that's an alternative to using a class called ReservedSchemeName.
23:26 sbp: so its inherent to the property then, a domain
23:27 sbp: s/its/it's
23:28 DanC: er... there's a question of whether we want to say that there's a scheme named "afs" that's reserved -- which is sorta silly; there isn't really any such scheme -- or just say that the *name* "afs" is reserved...
23:28 sbp: heh:-
23:28 sbp: [ :schemeName "Andrew" ]
23:28 sbp: dc:description "File" ;
23:28 sbp: a :ReservedURISchemeName .
23:28 DanC: ... and there's the question of using a one-place predicate Reserved("afs") or a two-place predicate reserves(aDoc, aSchemeName)
23:28 AaronSw: danc, you call them schemes in schemes.n3 and they're far from reserved!
23:29 sbp: the name and the scheme are one and the same I suppose
23:29 DanC: no, no, no, the name and the scheme are not one and the same. I'm not my name.
23:29 AaronSw: danc, i know
23:29 sbp: but it doesn't change by adding implementations does it?
23:30 AaronSw: but in schemes.n3 you do <silly-document> :specifies [:schemeName "foo"]
23:30 DanC: I'm reasonably happy to refer to the mythical scheme named "afs".
23:31 sbp: oh, Dan, does N3 need to have a trailing ; in an anon node, like [:x :y; :a :b; ]
23:31 AaronSw: danc, ok i used is :reserves of <...iana doc...>
23:31 DanC: but what <aDoc> :reserves [ :schemeName "afs" ] says is "there's some scheme named 'afs' that is reserved." I think what we mean is "*every* scheme named 'afs' is reserved."
23:31 sbp: its something I haven't figured out yet...
23:31 AaronSw: danc, hmmm
23:31 DanC: trailing ;: no, I don't think so.
23:32 sbp: cheers
23:32 sbp: but I could reserve my own afs scheme
23:32 AaronSw: i think what it's saying is that it reserves a scheme with a name of "afs", thus the n3 is right
23:32 sbp: Aaron: exactly
23:32 sbp: and the contxt of that URI resrving it is different form another reserving it
23:32 AaronSw: other wise the description would make no senmse.
23:33 sbp: Plus dates come into play: what document reserved a scheme first?
23:33 DanC: well... if we've said that :schemeName a daml:UnambiguosProperty, then there's only one scheme named "afs", I guess.
23:34 DanC: ... so saying "there is some scheme named 'afs' that is reserved" is the same as saying "all schemes named 'afs' are reserved."
23:34 sbp: agreed
23:34 DanC: and indeed, we said { :schemeName a daml:UnambiguosProperty } in schemes.n3 several revisions ago.
23:37 DanC: -> http://www.w3.org/2001/02pd/rfc65.n3 an event-based[INDECS, Harmony] model of Internet RFC publication.
23:37 AaronSw: ahh, it's on dan's machine
23:38 DanC: -> http://www.w3.org/2001/02pd/rfcIndexGrok.pl rfc-index.txt to n3 converter
23:38 DanC: sigh... not even a makefile to help folks find the bits and pieces of the RFC stuff.
23:40 DanC: ok... committed schemes.n3 r1.6 with pointer to aaron's thingy