Semantic Web Interest Group IRC Chat Logs for 2002-10-22

00:05:53 <MorbusIff> MorbusIff is now known as Morbus

02:40:09 * DanConn considers a silent power supply... $60 seems cheap for a quieter office.

02:40:25 <DanConn> silent pc: http://slashdot.org/articles/02/10/20/2322226.shtml?tid=137

09:16:02 <dajobe> hi graham

09:21:32 <gk> Hi Dave

09:22:15 <gk> I just popped in to see if anyone could answer Q's about DanCs RDF visualization stuff

09:23:48 <danbri> hello!

09:24:00 <danbri> I could try, though haven't used it much...

09:25:13 <gk> It's pretty simple -- I can't figure if the :digraph property is really needed, and if so does the complete list of nodes need to be specified?

09:29:02 * danbri out of my depth already, sorry

09:31:31 <gk> OK -- I'll poke around a bit & see what happens. Do you have a recommendation for a command-line XSLT processor to run on Windows?

09:32:35 <gk> I just tried Isaviz -- smooth piece of software, but not offering quite enough controkl over the graphics for what I want. DanCs stuff looks pretty cool.

09:32:56 <JibberJim> Xalan - http://xml.apache.org/xalan-c/

09:33:40 <danbri> yes, the isaviz zoom scrolly thing is rather nice

09:36:17 <gk> Ah yes, Xalan... still waiting for download of that. But if that's the one to use, I guess it's worth waiting for.

09:44:15 <maxf> gk, use Saxon. http://saxon.sf.net much better than xalan last time I checked

09:45:42 <JibberJim> - For windows http://jibbering.com/2002/10/xslt.wsf should work on any recent windows and is few hundred byte downloads.

10:10:25 <danbri> maxf, talk after lunch?

10:10:33 <maxf> right, ok.

10:10:39 <danbri> cool

12:28:15 <em-log> em-log is now known as em

12:35:06 <sbp-> sbp- is now known as sbp

12:58:27 <danbri> BLURB:ISBN-based metadata scraper?

12:58:28 <dc_rdfig> A: ISBN-based metadata scraper? from danbri

12:59:05 <danbri> A:I've started cataloguing my book collection, emphasis on grabbing the ISBN, plus a token effort re titles and a string representing author/editor etc

12:59:05 <dc_rdfig> Added comment A1.

12:59:30 <danbri> A:Idea is that one bit of crisp metadata, the ISBN identifier, should let me hook into better metadata than I can myself be bothered to create

12:59:31 <dc_rdfig> Added comment A2.

12:59:49 <bijan> Hmm. there was a free (for a while) bar code scanner. i think it would read book bar codes for ISBN.

12:59:51 <danbri> A:So I'm looking for ISBN-based metadata services, scraped, SOAP, RDF, whatever. URLs welcomed...

12:59:51 <dc_rdfig> Added comment A3.

13:00:32 <danbri> yeah, i had a look around those a while back, the free one (something -cat?) got given away as part of some business model whose details escape me

13:00:40 <bijan> Yep :)

13:00:58 <danbri> i don't mind typing in the ISBNs for now, at least to get a hundred or so, but i'm hoping i can avoid coding a screenscraper

13:01:15 <JibberJim> Does Amazons XML interface give you any info?

13:01:23 <bijan> Alas, ISBNs are notoriously unreliable.

13:01:40 <bijan> I don't know if they're *actuallY* unreliable, but librarians seems to scorn them.

13:01:43 <danbri> I don't know, re amazon, the restrictions on use made it not so interesting...

13:02:04 <bijan> I do know that the Squeak anthology I contributed to had the entirely wrong metadata page (copyright, etc.)

13:02:06 <danbri> isbns are from the publishing world. they get re-cycled, or used to, so they're not so clean as identifiers

13:02:23 <bijan> They just reused the copyright page from one of the editors other squeak book :)

13:04:20 <bijan> danbri, I think aaron was using some ISBN search service, but i forget what it was :)

13:05:26 <danbri> A:The [http://isbn.nu/|isbn.nu] search looks worth re-investigating...

13:05:26 <dc_rdfig> Added comment A4.

13:05:29 <danbri> maybe that one?

13:06:17 <bijan> Maybe

13:07:26 <danbri> I only found 3 from 45 on the bookshelf that lacked isbns

13:08:02 <danbri> grep \s books-danbri.txt|grep '^\?' |more

13:08:03 <danbri> "Reinventing Comics", Scott McCloud

13:08:03 <danbri> "Hiroshima", John Hersey

13:08:03 <danbri> "The Amazon", Time life books

13:08:20 <bijan> Try Library of Congress too.

13:08:23 <danbri> the 1st was a suprise. the 2nd was ancient. the 3rd was coffee table tat, not a real book.

13:08:36 <danbri> I think they charge for access (and make you use z39.50...)

13:08:46 <bijan> Uocl/

13:10:26 <bijan> Or, rather, "Yick!"

13:12:09 <JibberJim> - 1863202722

13:12:11 <JibberJim> oops...

13:12:50 <JibberJim> - http://blpc.bl.uk/ (British Library) but doesn't look easily scrapable.

13:14:03 <JibberJim> ah no, they have a text version, which seems okay.

13:14:19 <danbri> wanna chump the url?

13:14:42 <JibberJim> I'll let you...

13:15:13 <JibberJim> you need session id's in the url...

13:15:49 <danbri> ick

13:17:23 <JibberJim> A scraper would be easy though.

13:17:37 <JibberJim> s/scraper/RDF conversion/

14:59:23 <akuchlin> danbri: I have RDF that points to entries in my book diary.

14:59:35 <akuchlin> See http://www.amk.ca/books/index.rdf for an example file.

14:59:56 <akuchlin> (I keep meaning to come up with an application that uses this, and then

15:00:03 <akuchlin> try to convince people like Danny Yee to use it.)

15:17:50 <sbp> [lightbulb goes on] you're the Py cryptography guy, right? nice package

15:31:42 <DanConn>http://www.w3.org/2002/Talks/swandwai/

15:31:42 <dc_rdfig> B: http://www.w3.org/2002/Talks/swandwai/ from DanConn

15:31:51 <DanConn> B:|Semantic Web and Web Accessibility

15:31:51 <dc_rdfig> Titled item B.

15:31:58 <DanConn> B:by Marja-Riitta Koivunen

15:31:58 <dc_rdfig> Added comment B1.

15:34:49 <DanConn> B:Oct 2002, Finish W3C Office opening

15:34:49 <dc_rdfig> Added comment B2.

15:35:02 <DanConn> B2:Oct 2002, Finnish W3C Office opening

15:35:02 <dc_rdfig> Replaced comment B2.

15:37:35 <qmacro> qmacro is now known as q|away

16:00:12 <maxf> Hi there. Does anyone have example files that use the 'hybrid' RDF calendar vocabulary?

16:06:58 <shellac> um, I think retsina uses the hybrid schema

16:07:08 <shellac> try examples at http://www.daml.ri.cmu.edu/site/projects/RDFCalendar/

16:07:22 <shellac> (under 'Schedules')

16:07:41 <maxf> That's it. cool, thanks.

16:20:52 <DanConn> max, what are you up to?

16:21:15 <DanConn> if you're interested in converting from/to existing ical-based-stuff, I have tools for that.

16:24:09 <maxf> DanConn: yeah? I was looking at your palmagent stuff, but couldn't find any hybrid in there.

16:24:35 <DanConn> palmagent isn't the most recent work; I should make a link to the newer stuff...

16:25:10 <DanConn> -> http://www.w3.org/2000/10/swap/pim/travel.html

16:25:28 <DanConn> sigh... that write-up is still pretty crappy...

16:25:45 <DanConn> the most relevant tool is http://www.w3.org/2000/10/swap/pim/ical2rdf.pl

16:26:10 <DanConn> and to go the other way (though it only groks a few fields), http://www.w3.org/2000/10/swap/pim/toIcal.py

16:26:25 <maxf> ok, thanks.

16:26:50 <DanConn> those 2 scripts don't use the hybrid namespace, btw... at the calendar workshop, libby and I agreed to merge our stuff into a new namespace...

16:27:24 <DanConn> ... meanwhile, somebody else had already used the name we agreed to at the workshop! (/2002/10/calendar) I didn't reserve it right away.

16:27:38 <maxf> ha!

16:28:37 <maxf> I imagine the result you'll come up with won't be too different from the hybrid schema.

16:29:02 <DanConn> true

16:29:09 <DanConn> thought it's sorta hard to say how different it will be.

16:29:45 <maxf> I suppose it'll still be quite close to iCalendar anyway.

16:30:56 <DanConn> yes, the hybrid stuff was mapped to RDF as a sort of intellectual excercise... a certain amount of judgement here and there. my approach is entirely automated (by ical2rdf.pl)

17:49:51 <JibberJim> The RDF documents in http://www.w3.org/2001/Annotea/User/Protocol.html should have r: on the attributes about/resource etc. yeah?

17:51:16 <DanCon> so say the RDFCore WDs, yes. I'm personally reconsidering that decision.

19:26:27 * danbri scribbles some draft faqs into the web

19:26:36 <danbri>http://www.w3.org/2002/10/rdf-faq-edits.html

19:26:37 <dc_rdfig> C: http://www.w3.org/2002/10/rdf-faq-edits.html from danbri

19:26:54 <danbri> C:|RDF FAQ Development Page (collecting faq ideas...)

19:26:54 <dc_rdfig> Titled item C.

20:03:31 <ronwalf> dajobe: Is there a way to encode or decode a redland node into straight ascii?

20:04:20 <dajobe> not at present

20:04:31 <ronwalf> I'm finding myself jumping through hoops just to find all the places where the parka code requires a null terminated string

20:04:59 <ronwalf> Hm, Perhaps I can munge it...

20:05:39 <ronwalf> Hm...

20:06:07 <ronwalf> Is anything besides the length feild ever going to have a null sitting in it?

20:06:18 <dajobe> there are flags fields too

20:06:24 <dajobe> (and 2 lengths)

20:07:18 <ronwalf> What are the flags for?

20:08:08 <dajobe> I forget

20:08:19 <ronwalf> Wait, I thought the encode for a node ignored flags...

20:08:22 <ronwalf> wait, that's statements

20:08:33 <ronwalf> Harrumph

20:08:43 <ronwalf> Hm...

20:08:45 <dajobe> literal_is_well_formed_XML for example

20:10:09 <ronwalf> Two sizes for a node, or just two bytes?

20:10:23 <dajobe> I don't think this is a good approach

20:10:41 <dajobe> making sure there are no null bytes is possible but I never coded to that assumption

20:11:03 <ronwalf> True

20:11:40 <dajobe> you could do some lower level fiddle if this was a pain

20:11:44 <ronwalf> The problem is decoding your node_as_string back into an object

20:12:02 <dajobe> you mean librdf_node_encode

20:12:20 <ronwalf> That doesn't give you a string though

20:13:09 <dajobe> librdf_node_to_string is for debugging really

20:13:17 <dajobe> I make no guarantee that it won't change format

20:13:24 <dajobe> although, true, I haven't documented that

20:13:59 * danbri realises that all Adobe XMP RDF isn't RDF according to current RDFCore RDF/XML syntax

20:14:10 <danbri> they use unqualified 'about="..."'

20:14:15 <dajobe> yeah

20:14:31 <dajobe> but if MikeD gets the namespace change, it'll all be broken

20:14:33 <danbri> How would you feel about revisiting that decision in the light of implementor feedback?

20:14:41 <danbri> I don't want a namespace change...

20:14:45 <dajobe> I'll wait for the feedback

20:15:02 <dajobe> there's been too much predicting of potential complaints

20:15:08 <DanCon> er... we have the feedback; danbri just checked their shipped product.

20:15:46 * DanCon will send it to rdf-comments if that's easier for dajobe

20:15:55 <danbri> I'll do it.

20:16:33 <DanCon> nodeID is cool, but I wonder if it's realistic to deploy in this version.

20:16:43 <danbri> logger, pointer?

20:16:43 <danbri> See http://ilrt.org/discovery/chatlogs/rdfig/2002-10-22#T20-16-43

20:16:44 * dajobe rm -rf ~/rdfcore

20:17:06 <danbri> nodeID I can live with. It makes a huge deployability leap

20:17:55 <DanCon> at least the nodeID decision doesn't involve W3C saying that an example in one of its own RECs, along with content from shipped products, is broken.

20:18:24 <DanCon> it just involves saying that parsers are buggy/incomplete.

20:18:31 <ronwalf> dajobe: Generically, I could run it through a base64 encoder

20:18:46 <ronwalf> It would make debugging ugly, though

20:19:20 <ronwalf> Is it base64? Something like it anyway

20:21:41 <ronwalf> I'm looking at your node encoder though, and there isn't any notion of flags that I can see

20:21:51 <ronwalf> (unless it is in librdf_uri_as_string?)

20:22:40 <dajobe> there are four

20:22:47 <dajobe> fields

20:22:50 <dajobe> that can be null

20:22:55 <ronwalf> It's <type><size><string>, where type is a letter, size is a short, and then string

20:23:03 <dajobe> er, no

20:23:04 <ronwalf> Ok, looking again

20:23:25 <dajobe> see around line 924

20:23:43 <ronwalf> ah, literal has more

20:24:30 * danbri writing up rdf:about issue reopening request for www-rdf-comments

20:24:55 <dajobe> so I should stop editing the doc for another WD then

20:25:00 <DanCon> why?

20:26:40 <ronwalf> Hm, base64 would get rid of my other escaping problems, too

20:28:21 <danbri> no

20:28:32 <danbri> do you always stop editing when people send feedback? eek...

20:28:48 <dajobe> no

20:28:58 <dajobe> and that's unfair

20:29:08 <danbri> I guess such a change could have ripple through effects on lots of bits of the doc

20:29:20 <danbri> sorry, didn't mean it meanly

20:30:03 <DanCon> the change looks localized to me... just change URI==rdf:about to URI==rdf:about or local-name=about in a few places, and duplicate the syntaxTerms list with the unprefixed versions.

20:30:08 <danbri> I've heard a few 'is rdf stable / back compatible' etc worries, and generally shrugged them off, but today I thought thru this one, then on top of having changed my mind discovered XMP would be broken...

20:30:21 <dajobe> on this particular issue

20:30:26 <dajobe> I don't care

20:30:32 <dajobe> but it'll need more than 1 line

20:31:21 <dajobe> the non-prefixed xml attributes issue is very hard to explain clearly

20:32:20 <DanCon> hmm... does the syntax spec try to explain stuff? it seems to be just an exact specification... a reference work.

20:33:18 * ronwalf looks around to make sure he doesn't need to write his own base64 function

20:34:37 <dajobe> can I ask for changes in the form of a patch?

20:34:48 <danbri> we can probably just cite the issues list (or test cases) rather than put all the explanatory work into the syntax spec

20:34:50 <DanCon> yes...

20:34:52 <danbri> you can always ask ;-)

20:35:05 <dajobe> argh

20:35:07 <DanCon> am I looking at the source you maintain here? http://ilrt.org/discovery/2001/07/rdf-syntax-grammar/

20:35:12 <dajobe> it would change rdf's dependency on xml

20:35:21 <dajobe> i.e. non-namespaced XML would be ok, sort of

20:35:24 <DanCon> how would it change RDF's dependency on XML?

20:35:25 <danbri> how?

20:35:32 <DanCon> "non-namespaced XML"?

20:35:55 <DanCon> RDF with unqualified about attributes is 100% namespace-happy.

20:36:16 <dajobe> maybe that's ok

20:36:21 * ronwalf could do it in one easy, evil call to perl

20:36:22 <DanCon> you don't produce that HTML using XSLT or xml-spec or anything, do you?

20:36:29 <danbri> infosets that have some content that isn't in any namespace (but in our case that content's meaning can be divined by looking at the meaning of the tags higher up the infoset tree, eg. rdf:RDF)

20:36:33 <dajobe> DanCon: no

20:36:43 * DanCon fires up emacs to generate patch...

20:36:55 <danbri> wohoo, my work here is done...

20:36:57 <dajobe> a word-based patch is fine

20:37:16 <dajobe> I mean change "these words" to "these new words"

20:37:33 <danbri> not MS Word? I guessed somehow... :)

20:37:35 <dajobe> the only delaying thing for me

20:37:39 <dajobe> is the datatypes ntriples format

20:37:53 <dajobe> which we discussed yesterday

20:38:09 * DanCon wonders how n-triples is critical path for the syntax spec

20:38:29 <dajobe> I can't emit any ntriples datatyped statements without it

20:38:43 <dajobe> i.e. use of rdf:datatype is currently flagged with FIXME

20:40:33 * DanCon discovers use of n-triples-ish notation in the syntax spec

20:41:09 <danbri> -ish?

20:41:16 <DanCon> #

20:41:16 <DanCon> e.subject.string-value <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <e.URI> .

20:41:22 <DanCon> e.subject.string-value isn't n-triples

20:41:25 <danbri> ah

20:42:18 <DanCon> ugh; you maintain this sort of reference by hand? <a href="#section-baseURIs">section 5.3</a>

20:44:31 <DanCon> I don't suppose explanations in running text need to say "rdf:ID or ID", e.g. [[ constructed from XML Namespace-qualified element and attributes names

20:44:31 <DanCon> (QNames) or from rdf:ID and rdf:bagID attribute values. ]]

20:44:41 <dajobe> I'm thinking

20:44:46 <dajobe> maybe one sentence could do it

20:44:57 <DanCon> I think we can treat unqualified ID as a 2nd-class citizen.

20:46:03 <dajobe> if the it was something like "If a node element is in the RDF namespace, unprefixed attributes are given the rdf namespace" somewhere in the infoset section

20:46:11 <dajobe> s/node element/element/

20:46:35 <DanCon> no, about= should be happy on typednodes too... <my:Class about="abc"> too

20:46:57 <danbri> 'given'? can we say things that re-assign namespaces according to our own rules?

20:47:04 <dajobe> why not

20:47:18 <dajobe> rdf/xml's use of namesapces has been cursed far and wide

20:47:24 <dajobe> and by wg members

20:47:28 <DanCon> why not: cuz typednodes should work with unqualified about attributes too.

20:47:41 <dajobe> ok, I missed that

20:47:48 <dajobe> so all elements then

20:47:50 <ronwalf> dajobe: Will you be wanting to incorporate the parka bindings directly?

20:48:07 <dajobe> ronwalf: most likely

20:48:19 <danbri> isn't the namespace of an element or attribute determined entirely by the Namespaces REC?

20:48:39 <dajobe> "If an element has non-prefixed attributes, they are given the rdf namespace" somewhere in 6.1.2

20:48:46 <DanCon> implementors are gonna see a SAX2-like interface, mostly. They're not gonna remap about= to rdf:about= en mass; they'll have to check for both. XSLT scripters will have to check for @rdf:about or @about. I don't think it helps them to try to localize this change.

20:48:47 <danbri> the fact that we treat about= within RDF as being functionaly equivalent to rdf:about= is a different layer...

20:48:48 <dajobe> you can't have it both ways

20:49:21 * ronwalf waves the magic wand to rename his files/functions

20:49:26 <danbri> I can't? Or DanC can't?

20:49:35 * danbri missing the cake'n'eat it problem

20:51:03 <dajobe> either the 1 line solution or replace each mention of all dozen or so attributes with (rdf:foo | foo)

20:53:06 <DanCon> I'm replacing each mention in the formal syntax stuff with rdf:foo | foo. But in the discursive text, I think rdf:foo is sufficient. maybe a "foo is a deprecated form of rdf:foo, for foo=ID,about, ..." note is in order.

20:53:08 <danbri> I'm not sure. Maybe one line, but if it says 'is given the rdf namespace' i'm kinda squeamish.

20:53:47 <danbri> they're not namespace-same, just similar constructs that rdf parsers treat the same. A one liner might be able to say that...

20:54:11 <dajobe> don't expect polished word forms on irc

20:55:04 <dajobe> of course if you take the M&S grammar strictly

20:55:13 <dajobe> you'll have a mixture of allowed/forbidden rdf:foo/foo names

20:55:23 <dajobe> which is where we came in

20:56:12 <DanCon> where do you explain that rdf:ID isn't literally the URI, in stuff like [[ * If there is an attribute a with a.URI == rdf:ID ]]

20:56:33 <dajobe> rdf:ID there is notation for the URI

20:56:45 <dajobe> 5.1

20:56:47 <DanCon> I know that; in what part of the spec do you explain that, though.

20:56:53 <dajobe> [[Throughout this document the terminology rdf:name will be used to indicate name is from the RDF namespace and it has a URI of the concatenation of the ?RDF Namespace URI? and name. For example, rdf:type has the URI http://www.w3.org/1999/02/22-rdf-syntax-ns#type]]

20:57:26 <DanCon> hmm... one sentence right there might do it.

20:57:38 <DanCon> hmm... no... not quite...

20:58:22 <DanCon> aha; here's the place to localize the change: http://ilrt.org/discovery/2001/07/rdf-syntax-grammar/#eventterm-attribute-local-name

20:58:51 <DanCon> or rather: http://ilrt.org/discovery/2001/07/rdf-syntax-grammar/#eventterm-element-URI

20:59:55 <DanCon> around [[ Set to the string value of the concatenation of the value of the namespace-name accessor and the value of the local-name accessor. ]] just add before: In the case of namespace-qualfied element and attribute names, ...

21:00:43 <DanCon> and after, add: ... in the case of unqualified attributes named ID, about, resource, bagID [others?], set to the local name prepended with ...-syntax-ns#.

21:00:50 <dajobe> that's for elements

21:01:02 <DanCon> oops; quite.

21:01:05 <dajobe> you mean http://ilrt.org/discovery/2001/07/rdf-syntax-grammar/#eventterm-attribute-URI

21:01:22 <DanCon> yes, and never mind elements in the "in case of...".

21:01:54 * DanCon reviews use of syntaxTerms to see if this works...

21:02:34 <dajobe> the list is ID bagID about resource parseType type

21:03:06 <DanCon> syntaxTerms still seems to need to be doubled, as do occurences of rdf:ID,about,etc, in productions 7.2.20 thru 7.2.30

21:03:29 <Morbus> eric miller reminds me of a skinny steve jobs.

21:03:37 <Morbus> or that guy who played steve jobs. and that guy on ER.

21:03:47 <Morbus> who are the same person. but I rudely digress.

21:03:56 * DanCon wonders if that's close enough to a patch

21:04:53 <dajobe> a new syntax terms thing is ok

21:05:21 <dajobe> the actual list of rdf:foo | foo doubles would be corresponding to the list above

21:06:41 * dajobe will look at this tomorrow

21:07:14 <DanCon> the list [ID bagID about resource parseType type] should include datatype, the way I read the syntaxTerms production

21:07:36 <dajobe> if you are expanding to include new terms rather than just the old ones

21:07:50 <dajobe> type/rdf:type is an anomoly here

21:07:55 <dajobe> since M&S says use bare type

21:08:02 <DanCon> oh fart.

21:08:07 <dajobe> and the new grammar has no special case for it

21:08:26 <DanCon> new/old: good point.

21:11:53 <DanCon> hmm... production 7.2.20 and so on are about URIs, not attribute names, yes?

21:12:05 <DanCon> no, they're about attributes.

21:12:05 <dajobe> yes

21:12:21 <dajobe> er, attribute - yeah

21:12:33 <dajobe> 7.2.2 is about URIs

21:12:54 <dajobe> actually they don't need to change

21:13:01 <dajobe> if you added the rule earlier?

21:13:25 <DanCon> indeed, 7.2.2 needn't change.

21:13:43 <DanCon> but 7.2.20 and so on do need to change.

21:14:03 <DanCon> where's the production for the type attribute?

21:14:11 <dajobe> there is none

21:14:14 <DanCon> ?

21:14:15 <dajobe> not a special case

21:14:26 <DanCon> we took that out?

21:14:36 <dajobe> it's just a property attribute

21:14:51 <dajobe> like rdf:subject, rdf:predicate, rdf:object - just follow standard rules

21:14:54 <DanCon> but isn't the value treated as a URIref rather than as a literal in the '98 spec?

21:15:31 <DanCon> i.e. isn't type="foo" short for <rdf:type rdf:resource="foo"/> rather than <rdf:type>foo</rdf:type>?

21:15:45 <dajobe> yeah

21:15:56 <DanCon> so where's that specified?

21:16:32 <dajobe> it's a special case

21:16:41 <dajobe> see 7.2.9 action

21:16:51 <dajobe> [[If there is an attribute a in propertyAttr with a.URI == rdf:type then the following statement is added to the graph: ]] etc.

21:17:37 <DanCon> ok, got it.

21:19:38 <DanCon> hmm... none of the productions need to change; they all go via the URI accessor; changing that does all the work.

21:20:00 <dajobe> that's the 1 line

21:20:07 * DanCon is slow, sorry

21:20:20 <dajobe> s'okay, worth a second pair of eyes

21:22:05 <DanCon> ok, 1 line sent to www-rdf-comments

21:28:57 * danbri sends more evidence (mozilla 1.2; rss 1.0) that about= has been deployed

21:29:38 <danbri> adobe + mozilla + rss are a fair slice of RDF as currently deployed...