Semantic Web Interest Group IRC Chat Logs for 2001-12-06

01:51:54 * tim-fmly wsa in a hurry, meant: In what langauge are your personal profiles expressed?

01:52:08 <tim-fmly> P3P matched a apolicy against a profile.

01:52:13 <tim-fmly> i thought...

01:52:58 <sbp> aha! I thought it was some quip about defining definitions

01:53:36 <tim-fmly> as ian puts it, the inevitable meta joke.

01:53:42 <tim-fmly> not this time.

01:53:49 <sbp> :-)

01:54:15 * sbp notes that () as subject is currently broken in CWM...

01:54:31 <sbp> I want to try out the list generation stuff, perhaps in order to get statistics

01:54:47 <sbp> What is the current status of axioms.n3 (in /test/includes/)

01:54:49 <sbp> ?

02:52:28 <Invitado4> Invitado4 is now known as inkel

02:59:41 <mnot> sbp: are there any tools for working with EARL available (agents, filters, etc.)?

03:00:04 <sbp> Yes

03:00:38 <sbp> try "Implementations" in http://www.w3.org/WAI/ER/

03:00:47 <mnot> thanks

03:01:18 <sbp> we also had a little session at our F2F about it, although I'm not sure where the ref. is, or if it will be useful (just go through the ERT page: loads of fun for all the family!)

03:01:22 <mnot> ah, I was looking under QA

03:01:28 <mnot> cool

03:01:39 <sbp> It will be moved to QA at some point

03:01:59 * mnot loves bookmarklets

03:02:08 * sbp too :-)

03:02:35 <sbp> But I try to avoid them - my bookmarklet bar will be bigger than my browsing window soon

03:03:15 <mnot> do you use netscape/mozilla? if so, you can put directories in the personal toolbar...

03:03:33 <sbp> Jim Ley and Nick Kew are very busy working on EARL clients etc., but you'll need to ask them for details - they move pretty quickly

03:03:45 <sbp> I tend to use IE6, but am not adverse to Mozilla 0.9.6

03:03:49 <mnot> OK... I'm just generally interested at this point

03:04:14 <mnot> IE6 is nice 'cause it supports P3P... tho I *think* mozilla is moving in that direction...

03:05:33 <sbp> They seem to have moved in just about every other direction...

03:06:04 <mnot> heh

03:06:39 <sbp> BTW, the point about having the reserved URIs is that therefore people don't *have* to resort to changing the HTTP headers... some folks find it difficult (if not in HTML, on things like images)

03:08:28 <mnot> I know, I know... the thing is, I think there's going to be a lot more of this sort of thing as time goes on. It would be nice to make it as flexible as possible... perhaps have one well-known location that acts as a metadata map...

03:08:44 <mnot> HTTP headers are horrible that way.

03:10:11 <sbp> ah, I see what you mean... that's a good idea, but then it means you have to get loads of disparate groups to agree to a single syntax... and they won't

03:10:38 <mnot> yeah.

03:11:16 <mnot> but if the map was just "this entry is in this location" (i.e., a URI namespace to local URL mapping), it might work

03:11:59 <mnot> e.g., "http://www.w3.org/p3p/whateverns -> http://www.example.com/my/privacy/policy/reference/file/is/over/here"

03:12:39 <sbp> yeah, that'd be simple enough

03:12:59 <sbp> sorta like in the same way that everyone agreed to robots.txt because it's pretty simple

03:13:40 <mnot> yep. Performance may be an issue (extra round trip), but for site-wide metadata, I think it's worth the hit

03:14:18 <mnot> I have my problems with robots.txt ;)

03:15:49 <sbp> I agree/heh

12:20:06 <tim-fmly> SBP yesterday:

12:20:15 <tim-fmly> * sbp notes that () as subject is currently broken in CWM...

12:20:16 <tim-fmly> [20:58] <sbp> I want to try out the list generation stuff, perhaps in order to get statistics

12:20:16 <tim-fmly> [20:59] <sbp> What is the current status of axioms.n3 (in /test/includes/)

12:20:16 <tim-fmly> [20:59] <sbp> ?

12:20:23 <tim-fmly> (eastern time)

12:20:57 <tim-fmly> 1. If () as a subject is broken - I thought I had fixed that. Do you have a test case?

12:22:06 <tim-fmly> 2. The axioms stuff - conversion from KIF -- is very incomplete indeed. Maybe now we have teh math we could complete it -- the cardinality stuff was daunting before.

12:22:22 <tim-fmly> s/teh/the/g

14:35:54 <em> ping... testing connection?

14:59:31 * DanC waves

15:12:15 * DanC is back (gone 07:57:51)

15:18:38 <timbl> Proposed N3 changes discussion, anyone?

15:19:14 <timbl> 1) "forAll" and "forSome" keywords to replacae (over time!) "this log:forAll" etc

15:20:00 <timbl> keywords would have to be in the order forAll the forSome and would have to precede anything else in the formula (?)

15:21:08 <timbl> 2) "of" alowed as binary operator where foo:mother of :joe would have high precedence and would stanjd for [ is foo:mother of :joe ].

15:22:59 <timbl> 3) "with" keyword shorthand for two membered list. :a with :b short/long for ( :a :b ), eg: :x :outputString string:concatenmation of "foo" with "bar".

15:23:39 <timbl> "with" has precedence over "of"

15:26:52 <DanC> precedence: ugh.

15:27:09 <DanC> please keep the language LL(1)

15:27:11 <timbl> Ugh erxample: :F is log:inverse of log:semantics of os:environ of "TARGET".

15:27:15 <timbl> !!

15:27:22 <DanC> is?

15:27:29 <timbl> Well, exactly!

15:27:45 <DanC> where did "is" come from?

15:27:52 <timbl> You could define a rule to make sense of it.

15:28:04 <DanC> make sense of what

15:28:05 <DanC> ?

15:28:07 <timbl> "is" .. "of" didn't you inroduce that?

15:28:22 <DanC> oh... is/of. I remember now.

15:28:24 <timbl> Make sense of :F is log:inverse of log:semantics of os:environ of "TARGET".

15:28:53 <timbl> We have a lot of use cases of chains of properties, and I was sondering how to do them without nesting backets.

15:29:26 <timbl> Real example from roadmap/todot.n3: :F is log:inverse of log:semantics of os:environ of "TARGET".

15:29:29 <DanC> what's wrong with nesting brackets?

15:29:43 <timbl> oops

15:29:50 <timbl> Real example from roadmap/todot.n3: {[] has log:uri [is os:environ of "TARGET"]; has log:semantics :F} log:implies {:F a :target}.

15:30:54 <timbl> Well, for example in the list syntax, the () list convention is introduced because a the list case with first and rest gives you n nesting level for list length n.

15:31:15 <DanC> hmm... algernon had a shortcut for paths... (mother (brother (friend (dog Pete))))

15:31:37 <timbl> When you are just tracing through a graph without branches you aren't really saving state and so brackets are an extral mental overhead.

15:31:48 <timbl> right.

15:32:19 <timbl> joe mother whose sister whose dog spot.

15:32:41 <timbl> joe mother -> sister -> dog -> spot

15:32:44 <DanC> has would help there: joe has mother whose sister whose dog spot.

15:33:14 <timbl> joe has mother who has sister who has has dog spot

15:33:18 * DanC gets a bad feeling about english keywords... reminds me of COBOL

15:33:37 * timbl not to mentin periods at the end

15:34:11 <timbl> Well, pascal trod a balance, with {} and begin end being equivalent on same cases.

15:34:31 <timbl> You mean geeks wont use it unless it uses enough !@#$%^&* keys ?

15:34:32 <DanC> well, I'm serious about the LL(1) constraint. Blow that and it becomes much more difficult to implement a parser. But as long as you keep it LL(1), I guess I'm willing to give these a try.

15:34:50 <timbl> OK, I'll go along with LL(1)

15:35:11 <DanC> the bad feeling comes because english keywords trade on intuitions that might not apply in all cases; they can be misleading.

15:35:44 <timbl> Goal: to be able to describe a path through a graph involving forward and backward traversals without a bracket.

15:35:44 <DanC> COBOL gives you the impression you can just write english and the computer will execute it.

15:36:23 <DanC> is ' used up? have you considered 's?

15:36:42 <timbl> COBOL never gave me that impression much. Yes, you have to be really clear about what is happening and very clean. No little special shortcuts.

15:37:03 <DanC> joe's dog's owner's brother has height "10".

15:37:28 <timbl> For example, "is :mother of :joe" would have to use "of" oin exactly the same way as nay ohter production did. "is" woul dhave to be defined to work with that. Which it does.

15:37:37 <DanC> using that/who is appealing, since we tell people "read [...] as 'that which..."

15:38:40 * DanC wonders what obligations I'm neglecting...

15:39:16 * timbl ... yes project reviwe time is over.

15:39:28 <DanC> I was building a TODO list for today... went to send mail... decided I'm tired of manually typing email addresses... started thinking about how to automatically build my address book from my list of people I've sent mail to...

15:40:07 <dajobe> DanC: can you visit #webont please?

15:40:10 <DanC> ... started thinking about what format my one-line-per-message log should be in.

15:41:38 <timbl> {"TARGET" is os:environ of thatWhich is log:uri of thatWhich has log:semantics :F} log:implies{:F a :target}.

15:42:24 <timbl> {"TARGET" is os:environ of % is log:uri of % has log:semantics :F} log:implies{:F a :target}.

15:43:03 <DanC> tim, RDF collection issues came up in modelling P3P... should I use { _:policy :purpose :admin, :telemarketing } or { _:purpose (:admin :telemarketing) }?

15:43:25 <timbl> {"TARGET" os:environ % is log:uri of % has log:semantics :F} log:implies{:F a :target}.

15:43:36 <DanC> in P3P/xml, it's <purpose><admin/></telemarketing></purpose>

15:44:02 <timbl> Do I assme that you conclued that the semantics were that it was an exclusive list?

15:44:13 <timbl> s/Do/Should

15:44:25 <DanC> I haven't really decided, I guess.

15:44:53 <timbl> Well, if the semantics are not exclusive, then the answer is easy - _:policy :purpose :admin, :telemarketing .

15:44:55 <DanC> I do a lot of syntactic/closed-world sort of stuff, using log:notIncludes.

15:45:27 <DanC> and the :prop :o1,:o2 syntax is easier when doing notIncludes stuff.

15:45:34 <timbl> Yes, is the semantics are exclusive, you have a choice, to use lists or to just have hte assumtion that the policy is definitive and exclusive.

15:46:43 <DanC> with the definitive/exclusive assumption, the value of using RDF goes way down... i.e. the ability to merge pretty much goes out the window.

15:46:47 <timbl> Well, if you have a list, you don't need not:includes. You use :includes to get the list, and then you need notIn, which someone could put math to make it faster than by hand.

15:47:06 <DanC> notIn is really tricky

15:47:36 <timbl> Ah.

15:47:48 <DanC> you can only tell that { :x :notIn (:a :b :c) } if you know that { :x :noteq :a. :x :notEq :b. :x :notEq :c }

15:48:05 <timbl> Yes. lots of underlying assumptions that #admin and #telemarketting are in fact not equivalent, and stuff.

15:48:43 <DanC> turns out to be easier to use strings than URIs for enumerations.

15:48:46 <timbl> Well, daml:notEquivaltTo not log:notEqualTo

15:49:12 <DanC> daml calls it differentIndividualThan, I think

15:49:31 <timbl> Could have a rule that for any candidates for enumerations, they have the property that they are not equivalent to anything else?

15:50:06 <DanC> er... what do you mean by "anything else"? you could state they're all different from each other.

15:50:15 <timbl> No I mean "anything else".

15:50:25 * timbl is nervous but continues

15:50:29 <DanC> I don't know how to communicate "anything else" to cwm.

15:50:54 <timbl> You declarethat they have the property that if anything is equivalent to htem then it is URI-equal.

15:51:04 <DanC> all I could say is: if it's not in this list, then it's not in this list. duh.

15:51:37 <timbl> The question is, what happens if someone says "is <my:nurdle> on hte list?"

15:51:43 <DanC> er.. you mean that nobody is allowed to use any other URI to denote this thing? hmm...

15:52:04 <timbl> Yes ... breaking all the normal rules ... but we have to make the properties of an enumeration type.

15:52:18 <timbl> They are as you say much more like those of strings.

15:52:33 <DanC> btw... log:uri shouldn't work on variables.

15:53:09 <DanC> i.e. it shouldn't return variable names.

15:53:37 * timbl didn't think it did work on variables because variables have to be bound before it will run

15:54:12 * timbl on phone

15:54:32 <DanC> sure it will... e.g. if you say { :x a :Person } log:implies { :x :head [] }.

15:55:16 <DanC> then you ask { [] :head [ log:uri :headName ] } log:implies { :headName :shouldNot :beHere }.

15:55:38 * DanC makes a test case...

15:59:13 <DanC> hmm... test passes.

15:59:28 <DanC> I wonder why I was seeing gensyms returned by log:uri

15:59:49 <DanC> maybe it had to do with gensyms from RDF/xml parsing

16:05:21 <DanC> yup. somehow, RDF/xml anonymous nodes are not as anonymous as N3 anonymous nodes.

16:05:58 <dajobe> lol

16:09:37 <DanC> no, the problem is in the N3 parser too.

16:10:00 <DanC> somehow, stuff coming from conclusions are different from stuff that's parsed in.

16:11:11 <dajobe> what do you think of Jeremy's naughty suggestion to make all blank node identifiers live below http://www.w3.org/2001/12/RDF/bnode/ and add to the language

16:11:33 <dajobe> he said in http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Dec/0008.html

16:11:45 <DanC> I'm not sure...

16:12:11 <DanC> if he's suggesting that we go back on our decision that RDF can express "there exists..." then I'd say NFW.

16:12:46 <dajobe> it's more to allow rdf/xml to encode all legal rdf models like ntriples can

16:13:09 <DanC> on the other hand, if he's saying "if you're an RDF serializer, and somebody passes you one of these URIs thru the API, what he really meant is for you to write it out anonymously" then I might be able to go along with that.

16:13:24 <dajobe> that's it

16:13:48 <DanC> but that doesn't allow RDF/xml to encode as much as n-triples. ["model" is a misnomer there]

16:14:23 <dajobe> s/model/graph/ - well moreso, few things about property URIs and qnames

16:14:35 <DanC> it doesn't allow RDF/xml to say: <a> <b> _:x. <c> <d> _:x.

16:15:35 <dajobe> why not?

16:15:47 <DanC> well, how would you say that in RDF/xml?

16:16:05 <DanC> he's not suggesting that his special identifiers actually occur in RDF/xml files, is he?

16:16:32 <dajobe> yes; things like <a:b rdf:resource="http://www.w3.org/2001/12/RDF/bnode/x"/>

16:17:24 <dajobe> unsure of terminology, an algorithm could be replace all _: with "http://www.w3.org/2001/12/RDF/bnode/

16:17:25 <DanC> that's silly. It means parsers have to peek into URIs. If we want to fix RDF/xml syntax, let's fix it. his localID proposal made a lot more sense.

16:17:36 <dajobe> peeking: yes

16:17:47 <dajobe> if we can add localID

16:18:28 <DanC> RDF 1.0 already says what <a:b rdf:resource="http://www.w3.org/2001/12/RDF/bnode/x"/>. If we want to express something different, we need new syntax.

16:18:39 <DanC> says what <...> means, that is.

16:19:26 <dajobe> heh, we could snarf any namespace startign with http://www.w3.org/TR/REC-rdf-syntax namespace for that - reserved in M&S

16:19:46 <dajobe> actually no, forget that

16:20:36 <DanC> he's correct that the inability to express two properties with the same anonymous value is a design flaw; but if we open the door to fixing design flaws, we start quite a long process.

16:47:41 * chaals waves to Sean

16:48:01 <chaals> That "I don't agree" thing was cool BTW...

16:48:17 * chaals wonders if sbp this time is someone else...

16:48:50 <chaals> it is really frustrating to send a msg to someone and discover they aren't the person normally associated in one's mind with that nick.

16:49:38 <dajobe> the sbp nick is registered, it must be him or someone who has his password

16:51:52 <sbp> Oop, hi, I'm here :-)

16:52:01 <sbp> And it's the "real" me, but I can't prove it...

16:52:38 <sbp> I don't agree: thanks, but I don't agree :-)

16:53:06 <sbp> 'twas good, and showed what we could all do if we had CWM installed (ahem)

16:55:12 <sbp> Of course, it runs as a Webservice too, so we could set it up with filters such that one only has to type in the address of some EARL, and it will give you the reports

17:31:35 <DanC> "I don't agree thing"? pointer?

17:39:49 * sbp finds...

17:41:33 <sbp> logger, grep pleased.*w3c-wai-er-ig

17:41:37 <logger> I'm logging. I found 2 answers for 'pleased.*w3c-wai-er-ig'

17:41:38 <logger> 0) 2001-12-06 17:41:33 <sbp> logger, grep pleased.*w3c-wai-er-ig

17:41:39 <logger> 1) 2001-12-03 17:28:04 <sbp> * sbp emerges from ERT "call", pleased at http://lists.w3.org/Archives/Public/w3c-wai-er-ig/2001Dec/0003

17:42:26 <sbp>http://lists.w3.org/Archives/Public/w3c-wai-er-ig/2001Dec/0003

17:42:27 <dc_rdfig> A: http://lists.w3.org/Archives/Public/w3c-wai-er-ig/2001Dec/0003 from sbp

17:42:32 <sbp> A:|I Do Not Agree

17:42:32 <dc_rdfig> titled item A

17:43:45 <sbp> A::A filter experiment - uses [http://www.w3.org/2000/10/swap/|CWM] to sniff out disagreements between parties in [http://www.w3.org/2001/03/earl/|EARL] documents

17:43:45 <dc_rdfig> commented item A

17:44:33 <sbp> Just a simple little thing, but shows some value to drafting EARL in RDF :-)

17:47:07 * sbp searches for *.n3 on his HD...

17:47:35 <sbp> 996 files, 6.78 MB (7,115,381 bytes)

17:47:54 * sbp wonders if he should try merging them with CWM :-)

18:08:25 <AaronSw> Thinking about URIs for controlled vocabularies reminds me of the grant-matcher problem.

18:08:32 <AaronSw> .google erights grant matcher problem

18:08:36 <xena> erights grant matcher problem: http://www.erights.org/elib/equality/grant-matcher/history.html

18:08:44 <AaronSw> cf. http://www.erights.org/elib/equality/grant-matcher/

18:13:47 <sbp> Hmm... no point having them sitting on my hard drive; I'll compile an anthology for the Web

18:29:26 * timbl wonders how many of sbp's files are http://www.w3.org/2000/10/swap/test/includes/check.n3 valid

18:30:54 <timbl> Sean, how's the crypto going?

18:31:49 * timbl duh [13:20] *** sbp has quit IRC (Ping timeout: 181 seconds)

18:46:40 * sbp uploads 192KB of N3 to the Web

19:00:59 <timbl> Sean, how's the crypto going?

19:01:44 * timbl ,one day, will make a recursive N3 script to find all unschemavalid N3 files he i sresponsible for

19:06:14 * chaals today, decides it is time to sleep

19:06:19 <chaals> 'night all

19:16:31 <AaronSw> Why it's MarkB! Good to see you here.

19:58:23 <AaronSw> WebOnt is doing Web Services stuff?!

20:01:55 <AaronSw>http://groups.yahoo.com/group/WOLWebServices/

20:01:55 <dc_rdfig> B: http://groups.yahoo.com/group/WOLWebServices/ from AaronSw

20:02:09 <AaronSw> B:|WebOnt Web Services Requirements Group

20:02:09 <dc_rdfig> titled item B

20:02:26 <AaronSw> B:"A group with the goal to publish a requirements documenent which determines the requirements from a Web Services perspective"

20:02:26 <dc_rdfig> commented item B

20:03:05 <AaronSw> B:On YahooGroups, closed archives, restricted membership -- this makes me worry.

20:03:06 <dc_rdfig> commented item B

20:05:33 <AaronSw> B:On www-archive, Dan Connolly explains [why this is "a little tricky"|http://lists.w3.org/Archives/Public/www-archive/2001Dec/0016].

20:05:33 <dc_rdfig> commented item B

20:11:53 <AaronSw> B:I'm not even sure what Web Services has to do with WebOnt...

20:11:53 <dc_rdfig> commented item B

20:14:25 <sandro> The connection is problem that Web Services is a computer-related problem (how to make open distributed applications?) and WebOnt is a computer-related solution (model with description-logics) in search of good problems.

20:15:17 <AaronSw> I think I just got more worried.

20:15:44 <AaronSw> Hmm, I guess that makes sense./

20:16:33 <AaronSw> I guess I just thought that WebOnt was better scoped than that (approve DAML, not search for problems).

20:28:17 <sandro> Yeah. I dunno the story there.

20:51:20 * timbl wonders whether anyone has looked at http://www.w3.org/TR/2001/WD-semantic-interpretation-20011116/

20:51:29 <timbl>http://www.w3.org/TR/2001/WD-semantic-interpretation-20011116/

20:51:30 <dc_rdfig> C: http://www.w3.org/TR/2001/WD-semantic-interpretation-20011116/ from timbl

20:51:45 <timbl> A:|

20:51:45 <dc_rdfig> titled item A

20:52:07 <timbl> A:|Semantic Interpretation for Speech Recognition

20:52:08 <dc_rdfig> titled item A

20:53:38 <timbl> A:"the syntax and semantics of semantic interpretation tags that can be added to speech recognition grammars to compute information to return to an application on the basis of rules and tokens that were matched by the speech recognizer" from abstract

20:53:38 <dc_rdfig> commented item A

20:53:58 <timbl> A: I wonder whether anyone has looked at whether this should be/could be generating RDF

20:53:58 <dc_rdfig> commented item A

20:54:53 <timbl> A: I think the idean is that if you ahve a converation to order a pizza, the output is details of that order and of the pizza.

20:54:53 <dc_rdfig> commented item A

20:57:58 * sbp wonders what's going on

20:58:55 <sbp> I think you mean C, Tim

20:59:05 <sbp> C:|Semantic Interpretation for Speech Recognition

20:59:06 <dc_rdfig> titled item C

20:59:12 <sbp> C:"the syntax and semantics of semantic interpretation tags that can be added to speech recognition grammars to compute information to return to an application on the basis of rules and tokens that were matched by the speech recognizer" from abstract

20:59:12 <dc_rdfig> commented item C

20:59:24 <sbp> C:TimBL: I wonder whether anyone has looked at whether this should be/could be generating RDF

20:59:24 <dc_rdfig> commented item C

20:59:36 <sbp> C:TimBL: I think the idean is that if you ahve a converation to order a pizza, the output is details of that order and of the pizza.

20:59:36 <dc_rdfig> commented item C

20:59:51 <sandro> Sometimes it seems like we're all in orbit on the same problems.

20:59:58 <sbp> A:|I Do Not Agree

20:59:58 <dc_rdfig> titled item A

21:00:09 <sbp> A:Er... labelling mix up. Ignore the comments above

21:00:10 <dc_rdfig> commented item A

21:08:22 <DanC> A:nifty

21:08:22 <dc_rdfig> commented item A

21:10:43 <sbp> thanks

21:11:38 <DanC> A:sorta begging to use {} and such... hmm...

21:11:38 <dc_rdfig> commented item A

21:12:09 <bijan> A: *Very* interesting!

21:12:09 <dc_rdfig> commented item A

21:12:40 * sbp hopes they're talking about A A, and not C A :-)

21:13:02 <sbp> A:The problem is, we sorta decided on using XML RDF normatively, and there isn't a strict {} => XML RDF conversion (just log:Quote)

21:13:02 <dc_rdfig> commented item A

21:14:38 <DanC> A:quite.

21:14:39 <dc_rdfig> commented item A

21:15:28 <bijan> A:Be nice if there were an XML serialization for {} in the spirit of the RDF/XML (N3/XML anyone?).

21:15:28 <dc_rdfig> commented item A

21:16:19 <sbp> A:cf. [http://www.w3.org/2000/03/rdf-tracking/#rdfms-contexts|rdfms-context] in RDF Issues Tracking

21:16:19 <dc_rdfig> commented item A

21:19:39 * timbl slaps MarkB around a bit with a large trout

21:19:48 * sbp wonders if QNames should map to URI-views, or just stay as QName pairs using an undefined structuring

21:19:52 <sbp> lol, Tim

21:19:52 * timbl what?

21:20:17 <MarkB> ow

21:20:20 * timbl apologizes to markb - must have dropped a trout on kis keybotrad

21:20:35 <MarkB> set beep_on_trout

21:20:36 * timbl was going to apologize for hte A/C screwup....

21:21:15 * AaronSw jumps into action at the sound of trouts

21:21:20 * sbp wonders what's going on in the TimBL household/office :-)

21:21:42 * timbl has to move from office to household

21:22:23 * sbp notes he meant "QNames in log:parseType", but was interruped by the sound of fish-slapping

21:22:23 * timbl hmm. MarkB happened to be the one selected in the list of people, and then there must be a key shortcut.

21:22:27 * timbl slaps timbl around a bit with a large trout

21:22:45 * AaronSw slaps AaronSw around a bit with a large trout

21:23:05 * sbp wonders who's going to fall in the canal

21:23:28 <timbl> I have updated http://www.w3.org/2001/04/roadmap/todot.n3 to include URLs on graphnodes ... they had stopped working. Subgraphs still don't work.

21:24:31 * timbl later

21:24:42 <sbp> cool, cutting the XSLT out of the circle...

21:24:47 <sbp> c'ya Tim

21:25:31 <AaronSw> MarkB, are you still on the XMLPWG?

21:26:28 <AaronSw> It appears you are.

21:31:57 <sbp> Hmm... the full capability of tobot.n3 has just come to me

21:32:03 <sbp> s/tobot.n3/todot.n3/

21:32:53 <dtv1> hi

21:34:52 <mnot> IIRC MarkB was on a while, left, and very recently came back w/ a shiny new logo ;)

21:35:09 <MarkB> right-o

21:36:11 <sbp> Hmm... so, how do I get log:outputString to... well, output

21:37:06 <sbp> Using all.n3 --think, I got:-

21:37:07 <sbp> [[[

21:37:07 <sbp> :result0 log:outputString """

21:37:07 <sbp> /* transformed by: */

21:37:07 <sbp> /* $Id: todot.n3,v 1.6 2001/12/06 21:03:53 timbl Exp $ */

21:37:07 <sbp> digraph my__Graph {

21:37:09 <sbp> label="W3C Technology Roadmap";

21:37:11 <sbp> """ .

21:37:13 <sbp> ]]]

21:37:59 <sbp> and a ton of other crud... could purge [ log:notEqualTo log:outputString; a log:Chaff ] . perhaps

21:39:10 <sbp> well, that totally didn't work

21:40:43 <sbp> aha:-

21:40:44 <sbp> [[[

21:40:44 <sbp> @prefix log: <http://www.w3.org/2000/10/swap/log#> .

21:40:44 <sbp> this log:forAll :x , :y , :z .

21:40:44 <sbp> { :x :y :z . :y log:notEqualTo log:outputString }

21:40:44 <sbp> log:implies { :y a log:Chaff } .

21:40:46 <sbp> ]]]

21:41:04 <sbp> and it works, hooray!

21:42:24 <sbp> tsk, it needs to concat result0 and result9

21:43:46 <sbp> Hmm... no new check-ins to CVS. I'll try grep outputString

21:44:38 <sbp> aha, --strings

21:45:22 <sbp> great!

21:45:23 <sbp> [[[

21:45:23 <sbp> /* transformed by: */

21:45:24 <sbp> /* $Id: todot.n3,v 1.6 2001/12/06 21:03:53 timbl Exp $ */

21:45:24 <sbp> digraph my__Graph {

21:45:25 <sbp> label="W3C Technology Roadmap";

21:45:27 <sbp> }

21:45:29 <sbp> /* End of generated .dot file. */

21:45:31 <sbp> ]]]

21:46:09 <sbp> BLURB:String Output in CWM

21:46:10 <dc_rdfig> D: String Output in CWM from sbp

21:46:34 <sbp> D:CWM can now output strings, using log:outputString, and the --string command line

21:46:34 <dc_rdfig> commented item D

21:47:07 <sbp> D:cf. [http://www.w3.org/2001/04/roadmap/todot.n3|todot.n3] and [http://www.w3.org/2001/04/roadmap/all.n3|all.n3]

21:47:07 <dc_rdfig> commented item D

21:47:41 <sbp> D:Run: TARGET=all.n3; cwm todot.n3 --think --strings > out.n3

21:47:41 <dc_rdfig> commented item D

21:47:52 <sbp> D:It's *strings* plural, not singular

21:47:52 <dc_rdfig> commented item D

21:48:25 <sbp> D:Great stuff: might start hacking on an RDF => XHTML conversion thing, and perhaps EARL => XHTML... the possibilities!

21:48:26 <dc_rdfig> commented item D

21:49:12 <sbp> logger, where am I?

21:49:12 <sbp> See http://ilrt.org/discovery/chatlogs/rdfig/2001-12-06#T21-49-12

21:49:44 <sbp> D:cf. Discussion just above [http://ilrt.org/discovery/chatlogs/rdfig/2001-12-06#T21-49-12|2001-12-06 21-50] in #rdfig

21:49:44 <dc_rdfig> commented item D

21:50:03 <sbp> s/-/:/

23:35:17 <sbp> Hmm... can't get log:concatenation to work

23:35:30 <sbp> oh, heh: string:concatenation

23:55:54 <AaronSw> Heh heh heh, DanC really liked my %23 proposal.

23:56:05 <bijan> Which is not my fault.

23:56:21 <AaronSw> Of course it is.

23:56:37 <bijan> Nope nope nope nope.