Semantic Web Interest Group IRC Chat Logs for 2002-01-11

00:50:16 <mnot> anyone aware of an example of using rules to orchestrate a pipeline? e.g., :c :requires :b . :b :requires :a . asserts (a, b, c)

01:47:03 <sandro> mnot, I don't know anyone doing that in cwm. Of course it's been done in other rule systems (eg Prolog).

01:52:45 <mnot> ah, well. working with sequences in cwm is something that I bang my head against regularly

02:01:18 <sandro> Prolog has the same facilities I think, and with a bit of experience, it goes very smoothly.

02:01:58 <sandro> Of course the basic list-append relation is a trinary relation, which is going to be ugly in n3.

02:03:39 * mnot goes off to learn prolog

03:22:34 <dmiles> sandro, do you know of any mature inference engines?

03:23:38 <sandro> XSB comes to mind. There are lots; depends on the kind of logic you have in mind.

03:24:01 <dmiles> Constraint based experts system

03:24:12 <dmiles> err expert system

03:24:36 <dmiles> i was ussing XSB for about 3 months.. it kept breaking on large datasets

03:25:03 <dmiles> now i am trying to code one in SWI-prolog

03:25:20 <dmiles> but after 1.2 yrs i am looking outside

03:25:44 <dmiles> i would really be happy not writting it myself at this pint ;\

03:26:12 <dmiles> someone suggested Jess

03:26:52 <dmiles> i am afraid to spend a month porting it to my problems to find out it's tioo slow or cant hadle alot fo data

03:29:20 <sandro> Yeah. Sorry, I don't know the territiory all that well. You might drop an e-mail to www-rdf-rules; there are some serious folks on there.

03:30:01 <dmiles> it's all good, that is a good idea

03:30:21 <DanC> I'm interested in anything you find out about jess.

03:30:29 <DanC> what I've heard about jess sounds too good to be true.

03:30:33 <dmiles> i am hoping someone will say some obscure thing that kicks :)

03:31:08 <dmiles> yeah .. i really like Jess's rule system.. i gave it a try

03:31:29 <dmiles> but i would need to see what happens when i throw in 60,000 rules

03:31:55 <dmiles> ok 1000 rules with 590000 facts :)

03:32:25 <dmiles> every rule is a possible backchain event...

03:32:34 <dmiles> 59k^1000

03:33:01 <dmiles> err well (59k*1000)^2

03:33:20 <sandro> What's the kb?

03:33:58 <dmiles> lots of disperse data mainly at http://ontology.teknowledge.com/cgi-bin/cvsweb.cgi/SUO/

03:34:18 <dmiles> but i would like to try the Cyc-IKB + all there microtheories

03:34:53 <dmiles> i tghink i'll leave out wordnet mappings ;\

03:39:18 <dmiles> i have a working content managent system that holds the KBs then turn it all into logical entailed if/then statements cyc would call it 'canonicalization'

03:40:23 <dmiles> so i have cached out contraints and predicates.. enough to produce a forward chain only based daml pages

03:41:09 <dmiles> now need a stable inference mechanism.. to stuff it into

03:41:21 <dmiles> dont we all

03:42:48 <dmiles> DanC: i am gouing to give Jess a few more hours tonight

03:42:59 <DanC> cool.

03:43:26 * DanC is winning with cwm tonight...

03:43:34 <dmiles> i have to output the ikb into its format.. wont take long

03:44:03 <dmiles> cwm is behaving?

03:44:36 <dmiles> not that i thought it wasnt but thought it might be runing into combinatoric problems

03:46:00 <dmiles> for example if you made a rule that results in a new :type or an inverse

03:46:20 <dmiles> or a consequenting in a new subProperty

03:46:50 <dmiles> it may have a searchspace that is defined by more searchspace

03:47:22 <dmiles> until a closure is found..

03:47:42 <DanC> sure, there are lots of ways to send cwm into the weeds... but tonight, my application isn't one of them, fortunately.

03:48:13 <dmiles> right on

03:49:49 <dmiles> in about an hour i should know whether i can use Jess on a mediam of 500 axioms

03:50:12 <dmiles> i am sure it can.. if that sth ecase i am going tio use it to cache out the existentials

03:50:50 <dmiles> to where the inference code will be adiquately forward chained

03:51:45 <dmiles> no more backtracking to find type/subclass

04:03:15 <DanC> woohoo! evolution imported the calendar I exported from RDF!

04:04:01 <Skyline> cool

04:04:05 <DanC> I'm writing toICal.py, which uses llyn.py as a KB and RDF import tool; then I do KB queries to walk the data, writing it out in ical syntax.

04:04:08 <Skyline> what format is used in the middle there

04:04:22 <Skyline> ical, ok

04:04:34 <DanC> I'm using Libby and Greg's ical schema. (with maybe a couple tweaks)

04:04:53 <DanC> the whole chain is: (1) email from the travel agent

04:05:10 <DanC> (2) thru grokNavItin.py into RDF, using the travel agent's terminology but RDF syntax.

04:05:44 <DanC> (3) then I use itin2ical.n3, an N3 rules file, to map from the travel agent vocabulary to ical vocabulary

04:06:03 <DanC> (4) toICal.py extracts ical vocabulary stuff from an RDF KB and spits it out in ical syntax.

04:06:22 <DanC> (5) import into evolution.

04:06:29 <DanC> TADA!

04:07:18 <DanC> actually, after step 2, I run rules to check constraints ala "make sure the itinerary gets me there in time for the meeting, but gets me back in the time agreed with my wife" and such.

04:08:00 <DanC> i.e. there are usually several itineraries proposed by the travel agent, and I use cwm to evaluate them to see which match my constraints.

04:09:23 <DanC> this is so cool... even the timezone stuff worked out.

04:14:28 <sandro> very cool. :-)

04:15:30 <DanC> yup... I suppose I should write it up... but I'm interested in the next steps too:

04:16:27 <DanC> converting ical syntax back to RDF, and doing consistency checking, ala "if my email says I'm busy at time T, but my evolution calendar doesn't, raise a flag!"

04:16:52 <DanC> ala... <evo-cal.rdf> log:semantics [ log:notIncludes ...]

04:17:56 <DanC> sandro, wanna noodle on how this connects to your federation ideas?

04:18:51 <DanC> eventually, I want my evolution calendar to act as a little agent/service. RDF that's available via HTTP naturally acts as a KB service...

04:19:18 <DanC> ... then I want little agents that go around checking consistency between, say, web pages of groups that I belong to versus my evolution calendar.

04:43:31 <DanC> logger_1, pointer

04:43:32 <DanC> See http://ilrt.org/discovery/chatlogs/rdfig/2002-01-11#T04-43-31

04:47:11 <DanC>http://www.w3.org/2002/01dc-nj/notes

04:47:11 <dc_rdfig> A: http://www.w3.org/2002/01dc-nj/notes from DanC

04:47:21 <DanC> A:|WebOnt NJ Trip: Semantic Web Travel Schedule Tools

04:47:22 <dc_rdfig> titled item A

04:53:09 <DanC> A:oops... gotta check in my fixes to llyn.py ...

04:53:09 <dc_rdfig> commented item A

04:54:18 <DanC> A:hm... that involves updating to the latest cwm... sometimes risky...

04:54:18 <dc_rdfig> commented item A

04:55:45 <DanC> A:ok... updated cwm with no problem

04:55:45 <dc_rdfig> commented item A

05:28:59 <DanC> A:sent a [msg|http://lists.w3.org/Archives/Public/www-rdf-calendar/2002Jan/0002.html] to www-rdf-calendar

05:28:59 <dc_rdfig> commented item A

07:42:46 <mnot> Thanks for the clue-by-four, BTW ;)

07:43:15 <mnot> oops, wrong window. But hey, thanks everybody anyway...

09:01:42 <tansaku2> tansaku2 is now known as tansaku

15:05:21 * AaronSw pokes danbri_ into #rdfcore

15:16:44 <AaronSw> Can some W3T member activate http://www.w3.org/2001/09/rdfprimer/primer.txt please

15:18:56 <AaronSw> and http://www.w3.org/2001/09/rdfprimer/01-primermodel12.html

15:30:43 <danbri_> danbri_ is now known as danb_awol

15:33:40 <em-mit> AaronSw, these are done

16:05:06 <larsbot>http://www.oasis-open.org/committees/tm-pubsubj/

16:05:08 <dc_rdfig> B: http://www.oasis-open.org/committees/tm-pubsubj/ from larsbot

16:05:24 <larsbot> B:|OASIS Topic Maps Published Subjects TC

16:05:25 <dc_rdfig> titled item B

16:05:50 <larsbot> B:The topic maps crowd is starting work within OASIS to define URIs to indicators for common subjects

16:05:50 <dc_rdfig> commented item B

16:06:22 <larsbot> B:This TC makes guidelines for such indicator sets, but TCs for the area of XML and languages/geography are on the way

16:06:22 <dc_rdfig> commented item B

16:06:33 <larsbot> B:The goal is that the URIs should be usable in RDF as well...

16:06:34 <dc_rdfig> commented item B

16:16:20 <DanC> B:does this include things like airports, citities, days of the week, that sort of thing?

16:16:20 <dc_rdfig> commented item B

16:16:51 <larsbot> B:In the long run it could; depends on what there's interest in

16:16:51 <dc_rdfig> commented item B

16:17:15 <larsbot> the idea is that vertical industries can run their own shows within or without OASIS

16:17:34 <larsbot> some core things are being done by the TM-heads themselves, though

16:17:36 <DanC> how is that different from the present situation?

16:17:46 <DanC> i.e. folks can already make up URIs for things. I guess I don't get it.

16:18:00 <DanC> I'm interested to know what these core things are, then.

16:18:38 <larsbot> what's different is that there are guidelines for how to publish, document, and update such URI sets

16:18:46 <larsbot> and also how to define the meaning of each URI

16:19:04 <larsbot> the core things being done now are languages and countries

16:19:09 <DanC> hmm... guidelines like "hold the dang things still"?

16:19:19 <larsbot> rather more than that :)

16:19:29 <larsbot> what do you do when a code set has 3 codes for every subject?

16:19:40 <larsbot> what assertions should you make about the subjects?

16:19:52 <larsbot> how should you express equivalences across code sets?

16:20:05 <larsbot> how do you define the meaning of each code?

16:20:14 <DanC> using DAML+OIL/WebOnt, esp UnambiguosProperty.

16:20:19 <larsbot> how do you package and publish a set of codes?

16:20:24 <AaronSw> thanks em-mit

16:20:55 <larsbot> we're TM-heads, so we're not thinking in that direction

16:21:07 <larsbot> RDF-heads may correct our thinking

16:21:46 <DanC> I was being a bit flip, taking the "you" in "what do you do..." to refer to me personally.

16:22:02 <larsbot> ah, comprendo

16:22:07 <DanC> I can now see why the PubSub folks are trying to get the attention of the WebOnt WG.

16:22:29 <larsbot> how so?

16:22:31 <AaronSw> pubsub folks? publish-and-subscribe?

16:22:47 <larsbot> published subjects

16:23:03 <AaronSw> ah

16:23:12 <larsbot> see the URI I just chumped

16:23:13 <DanC> how so: because a lot of the things that folks in the PubSub group are trying to encode/express are much like what folks in the WebOnt WG are trying to do.

16:23:30 <larsbot> I should look at that, then; do you have a URI?

16:23:39 <AaronSw> .google webont

16:23:40 <xena> webont: http://www.w3.org/2001/sw/WebOnt

16:23:44 <larsbot> thx

16:23:57 <DanC> B:see also [W3C WebOnt WG|http://www.w3.org/2001/sw/WebOnt/]

16:23:57 <dc_rdfig> commented item B

16:24:35 * larsbot is reading

16:25:08 <dmiles> DanC: jess is cool.. but its definately a programming language more then a declaritive expert systems language

16:25:17 <DanC> re "not thinking that way"... do the TM-heads use math/logic to express their thinking at all? I found the TM spec (as of Mar 2001) hard to understand because it didn't relate to math/logic.

16:25:36 <larsbot> not very much, I'm afraid

16:25:41 <DanC> bummer.

16:25:41 <larsbot> I think we need to do that more

16:25:50 <DanC> glad you think so.

16:25:59 <larsbot> the webont people seem to be doing more than the pubsubj people

16:26:09 <dmiles> I dont mind it.. its just would be hard to get new ussers to write Jess based rules w/o alot of pretraining

16:26:19 <DanC> er... yeah, WebOnt might be a little *too* into math/logic.

16:26:26 <larsbot> pubsubj is about establishing identity, nothing more

16:26:44 <DanC> as if "establishing identity" were a simple thing to do ;-)

16:26:53 <larsbot> true :)

16:26:59 <AaronSw> yeah, pubsubj is a subset of webont requirements, I think.

16:27:38 <larsbot> in the language PSI set you'll basically find ISO 639 codes recycled, with English and French names assigned to each code + the assertion that the identities are instances of language (as defined by ISO 639)

16:27:47 <larsbot> aaronsw, seems that way

16:28:13 <larsbot> not very much, but, like XML, very valuable if everybody uses it

16:28:42 <DanC> the real trick would be to get ISO to publish their dang stuff in the web. 1st-party data is so much more valuable.

16:29:15 <AaronSw> yeah, but if they did that, what are the chances they'd have persistent uris ;)

16:29:28 <DanC> I meant stable-publish.

16:29:49 <larsbot> that's true, but I think the chances of them doing so are essentially nil

16:30:29 <DanC> eventually, they will or they'll become irrelevant. But this "eventually" is on a timescale of years or decades, unfortunately.

16:31:33 <larsbot> exactly, and we need a solution *now*

16:31:38 <DanC> actually, ISO might already be irrelevant... i.e. it's not clear that if they decided to change their country codes, it would have real-world impact. Perhaps something like the international postal union is more relevant.

16:32:10 <DanC> their country codes are relevant every day: use the wrong code and your package doesn't get there.

16:32:11 <larsbot> for example, I'd like our software to be able to use the HTTP language header to select what names to show for topics in web applications

16:32:30 <larsbot> they are relevant since they are very widely used

16:32:40 <larsbot> if you want to integrate data properly you'd better be able to use them

16:32:43 <AaronSw> Hmm, i didn't realize they had country codes... i just write the name of the country on my packages

16:32:52 <DanC> well, a historical snapshot of the ISO codes is widely used.

16:33:09 <DanC> *the* name of the country? you mean *a* name, right?

16:33:12 <larsbot> oh, they are used in *many* places, even in Norwegian encyclopedia systems :)

16:33:31 <AaronSw> yeah, a name.

16:33:36 <larsbot> ISO has codes for languages, countries, regions, currencies, scripts, human sexes, and probably more that I don't know of

16:34:02 <larsbot> dunno how often they update the last one :)

16:34:17 <DanC> the 1st-party/3rd-party distinction is largely irrelevant if you're talking about a historical snapshot. It's only if the data can change that it matters.

16:34:34 <larsbot> true, and these codes do change every now and then

16:34:46 <larsbot> we intend to update our PSI sets to follow them, in as near-realtime as possible

16:35:13 <AaronSw> So you mean that you're going to change the meaning of your URIs?

16:35:23 <AaronSw> (or let ISO change their meaning)

16:36:07 <larsbot> I've yet to see ISO change the meaning of one of their codes,

16:36:16 * DanC noodles on consistency and context...

16:36:18 <larsbot> they tend to remove old ones and introduce new ones

16:36:33 <DanC> ISO change: ask Misha wolf. They do change them.

16:37:14 <larsbot> ok, I've missed something, then

16:37:54 <DanC> Misha wrote a huge rant on how W3C should grab the country codes and hold the dang things still, cuz ISO isn't doing it, and Reuters can't afford for them to change.

16:38:12 <larsbot> well, most people probably can't

16:39:11 <AaronSw> interesting: http://lists.w3.org/Archives/Public/www-international/1999OctDec/0033.html

16:39:12 <DanC> actually, there's a huge tension between having short identifiers that everybody knows and keeping the meaning of a symbol fixed across time.

16:40:10 <DanC> ooh... AaronSw, shall we scrape that list, ala the way we scraped the IANA registry of URI schemes?

16:40:13 <larsbot> there is, since in the end you'll be bound to run out of codes; there are, after all, only 676 two-letter codes

16:41:01 <AaronSw> they can start going into symbols then

16:41:08 <larsbot> danc, what do you mean "scrape"?

16:41:28 <larsbot> (I can give you the list in XML form, BTW)

16:41:37 <AaronSw> scraping doesn't look too hard

16:41:44 <AaronSw> ooh, where's the xml form? is it official?

16:41:49 <DanC> "This list is updated whenever a change to the official code list in ISO 3166-1 is effected by the ISO 3166 Maintenance Agency." <-- I'd like to see a guaranteed maximum latency there. e.g. "updated within 6 months of any change..."

16:42:05 <larsbot> nah, I scraped it from the din.de site :-)

16:42:23 <AaronSw> by scraping we mean software to turn it into rdf

16:42:39 <AaronSw> (i think)

16:42:54 <AaronSw> like http://blogspace.com/rdf/iana-uri-schemes

16:43:00 <DanC> AaronSw, we could record the whole provenance of the din.de thing: din.de is the official registrar because ISO WG2l3kj423 decided, on 15Jan1997 that they are the registrar for the next 10 years ...

16:43:13 * em-mit is more familiar with the iso639-2 codes -> http://lcweb.loc.gov/standards/iso639-2/langcodes.html

16:43:54 <larsbot> actually, the pages say that din.de is handing that function over to someone else (I forget whom), and that the site is going up on iso.org

16:44:00 <danb_awol> danb_awol is now known as danb_desk

16:44:10 <DanC> EricM, I'm torn about this. When W3C started the RDF WG, we promised not do do any real-world schemas. We promised to leave that to other parties. But can we afford to wait for ISO to publish their codes in RDF?

16:44:27 <danb_desk> Is that promise documented?

16:44:36 <DanC> I expect so...

16:44:46 <AaronSw> i think it's cool that w3c is hosting schemas, like danbri's copy of the RSS 1.0 schema

16:45:14 <larsbot> aaronsw, what's that file actually saying?

16:45:25 <AaronSw> the blogspace one?

16:45:28 <danb_desk> I'm in a similar situation on RSS-DEV (as Aaron notes); but am wary of folk thinking these things *need* to be at w3.org to be important etc.

16:45:29 <larsbot> yeah

16:45:31 <em-mit> DanC, yes... i feed the same way.

16:45:56 <danb_desk> www.very-important-centralised-schema-registry.org isn't the RDF way.

16:46:06 <AaronSw> it says: this code broke cuz iana broke their uris

16:46:15 <AaronSw> yick

16:46:17 <larsbot> well, Danc, if you republish the codes in RDF, and we do it in XTM, let's at least make sure we use the *same* URIs...

16:46:21 <em-mit> LOC a couple years back was discussing hosting the ISO codes in XML/RDF, but i stopped going to the MARBI meetings and this fell off their radar... th

16:46:41 <DanC> FRAGLKJESLIjwel! we can't even get the IETF to do stable-publishing in the web. Is there any hope? Sometimes I dispair.

16:47:13 <AaronSw> it's funny tho: it broke cuz they renamed url-schemes to uri-schemes

16:47:19 <AaronSw> ah, ironiy

16:47:21 <AaronSw> err irony

16:47:29 <DanC> they didn't leave a redirect?

16:47:35 <AaronSw> a textual one

16:47:39 <DanC> whee!

16:47:54 <AaronSw> it's ftp: ftp://ftp.isi.edu/in-notes/iana/assignments/url-schemes

16:48:00 <AaronSw> --

16:48:01 <AaronSw> This registry has moved to the following:

16:48:02 <AaronSw>http://www.iana.org/assignments/uri-schemes

16:48:02 <dc_rdfig> C: http://www.iana.org/assignments/uri-schemes from AaronSw

16:48:02 <AaronSw> For all registries, please see the following:

16:48:02 <AaronSw>http://www.iana.org/numbers.htm

16:48:03 <AaronSw> Updated 2001 September 6

16:48:03 <dc_rdfig> D: http://www.iana.org/numbers.htm from AaronSw

16:48:03 <DanC> isi.edu. great.

16:48:06 <AaronSw> oops.

16:48:07 <danb_desk> OH, em-mit, DanC: you asked about the video camera I had working with Debian. It's a USB, Ezonics EZCam. I modprobed a bunch of likely looking kernel modules until gqcam was happy. Might've had to make /dev/video too.

16:48:21 <AaronSw> C:|IANA's new location for the uri schemes list

16:48:21 <dc_rdfig> titled item C

16:48:36 <DanC> C:everybody say a prayer that they hold it still for a while.

16:48:36 <dc_rdfig> commented item C

16:48:38 <em-mit> danb_desk, what tools are you using to drive this?

16:48:44 <AaronSw> otoh, http://www.iana.org/assignments/url-schemes is just broken

16:48:58 <AaronSw> D:|Stuff IANA keeps track of

16:48:58 <dc_rdfig> titled item D

16:49:40 <danb_desk> At the moment, just gqcam (a little thing that pops up a window with cam's contents, so I can see if I have bad posture. (I do)).

16:49:57 <danb_desk> My webcam, such that it is, is nothing more than: [[

16:49:58 <danb_desk> print "Content-type: image/png\n\n";

16:49:59 <danb_desk> print `killall gqcam`;

16:49:59 <danb_desk> print `/usr/bin/gqcam --dump - `;

16:50:00 <danb_desk> ]]

16:50:04 <danb_desk> ...as a cgi script.

16:50:23 <DanC> EZCam: can you give me the USB id? product names are, unfortunately, not unambiguous w.r.t. hardware/software interfaces in the PC hardware marketplace.

16:50:26 <AaronSw> i suppose i better update my script

16:51:17 <DanC> pls do

16:51:38 <dajobe> ooh, Henry Thompson writes an xml schema for dc in rdf/xml for me

16:51:41 <danb_desk> I don't know where it was bought. Where do I find USB id?

16:51:43 <danb_desk> anbri@debian:/proc/video$ more dev/video0

16:51:43 <danb_desk> name : CPiA Camera

16:51:43 <danb_desk> type : VID_TYPE_CAPTURE

16:51:43 <danb_desk> hardware : 0x18

16:51:45 <AaronSw> C:My [script to convert this to RDF|http://blogspace.com/rdf/iana-uri-schemes.tcl.txt] is pnow fixed|http://blogspace.com/rdf/iana-uri-schemes].

16:51:46 <dc_rdfig> commented item C

16:51:49 <DanC> did anybody see that I got an RDF calendar imported into evolution?

16:52:01 <AaronSw> ooh, no...

16:52:21 <DanC> I chumped it and sent it to www-rdf-calendar

16:52:21 <danb_desk> Yup, nice work! I'm more inclined to wrestle with installing evolution at some point...

16:52:31 <dajobe> I've switched to evo on new home pc

16:52:49 <dajobe> isn't quite there in terms of usability tho

16:52:59 <DanC> I also got evolution to import ldif contacts generated from python...

16:52:59 <danb_desk> Oh btw, happy birthday for yesterday dajobe!

16:53:06 <dajobe> cheers

16:53:20 <DanC> happy birthday, indeed!

16:53:28 <dajobe> 0x21 again as the joke goes

16:53:37 <DanC> so now I want to rdf-ize all of ldif, of course.

16:53:49 <DanC> not to mention RFC822

16:53:53 <danb_desk> DanC, can you advise me on how to find the USB ID for the camera, before I flip back to something else...

16:54:01 <DanC> er... lsusb?

16:54:19 * AaronSw adds this information to his database: dajobe turned 33 on 2002-01-10

16:54:19 * dajobe also interested in good linux webcams

16:54:28 <DanC> or look in /proc/bus/usb/devices

16:54:51 <DanC> sigh... yet another thing to RDF-ize.

16:55:12 <AaronSw> my contact book exports vCard

16:57:11 <dajobe>http://lists.w3.org/Archives/Public/xmlschema-dev/2002Jan/0108.html

16:57:12 <dc_rdfig> E: http://lists.w3.org/Archives/Public/xmlschema-dev/2002Jan/0108.html from dajobe

16:57:15 <danb_desk> lsusb: can't find such a thing (nor can apt-get). ooh, but auto-apt tells me to apt-get install 'usbutils'. So I try lsusb, lsusb -v etc., it returns nothing. And /proc/bus/usb/ is empty.

16:57:24 <dajobe> E:|XML Schema for Dublin Core in RDF/XML, Henry S. Thompson

16:57:24 <dc_rdfig> titled item E

16:57:25 <danb_desk> I'm suprised it's working!

16:57:33 <dajobe> E:attached tar.gz of XSD files

16:57:34 <dc_rdfig> commented item E

16:57:42 <DanC> but the camera is working, danbri? odd.

16:58:11 * DanC should be working on a URI CG charter...

16:58:13 <danb_desk> yes, just checked, still working.

16:58:43 <DanC> I see /proc/bus/usb on /proc/bus/usb type usbdevfs (rw)

16:58:47 <dajobe> danb_desk: apt-get install usbutils

16:58:51 <DanC> so can you try to mount it?

16:59:06 <DanC> I guess: mount -t usbdevfs /proc/bus/usb /proc/bus/usb ?

16:59:08 <danb_desk> thanks; I did that.

16:59:14 <em-mit> danb_desk, what kernel are you running?

16:59:16 <danb_desk> usbutils I mean.

16:59:29 <danb_desk> debian:/proc# uname -a

16:59:29 <danb_desk> Linux debian 2.4.16-686 #1 Wed Nov 28 09:27:17 EST 2001 i686 unknown

17:02:02 <danb_desk> mount: that worked. Odd, I don't recall having done that before, yet I've certaintly seen stuff in /proc/bus/usb. Maybe another module I've sometimes had loaded did it for me. Anyway:

17:02:03 <danb_desk> [[

17:02:05 <danb_desk> tav: Bus=01 Lev=01 Prnt=01 Port=00 Cnt=01 Dev#= 2 Spd=12 MxCh= 0

17:02:05 <danb_desk> DanC: Ver= 1.00 Cls=00(>ifc ) Sub=00 Prot=00 MxPS= 8 #Cfgs= 1

17:02:05 <danb_desk> P: Vendor=0553 ProdID=0002 Rev= 1.00

17:02:05 <danb_desk> sbp: Product=USB Camera

17:02:05 <dc_rdfig> Label P not found.

17:02:06 <danb_desk> ]]

17:02:15 <danb_desk> ...is I think what you need.

17:02:41 <sbp> I need a USB camera? :-)

17:04:02 <AaronSw> tav too

17:04:39 * tav cries for his wonderful camera somewhere in new mexico

17:05:28 <sbp> heh, heh

17:06:24 <DanC> thanks, danb_desk. " Vendor=0553 ProdID=0002" is what I'm after.

17:07:58 <DanC> crap.. the usb devices database hides the dang numbers. http://www.qbik.ch/usb/devices/showdev.php?id=979

17:08:13 <danb_desk> sbp: Wouldn't do any harm! we could do... SELECT ?mbox, ?pic from rdfweb WHERE (foaf:webcam ?who ?pic) (foaf:mbox ?who ?mbox) (foaf:receivesMail ?mbox mailto:www-rdf-interest@w3.org)

17:09:51 * DanC parses...

17:09:58 <AaronSw> we can see how people look when they get emails from us

17:10:00 <DanC> receivesMail is something like subscribedTo?

17:10:59 <danb_desk> I just made it up. Yeah. kindof like listMember but mailbox to mailbox instead of human/mailbox. Do you have subscribedTo in a SWAD vocab somewhere? happy to borrow...

17:11:29 <DanC> no, no SWAD-registered subscribedTo. do see cyc:groupMembers , though

17:12:02 <AaronSw> larsbot, where's the page with the country code uris?

17:13:14 <DanC> "The original Ezonics cam was reported to have worked with CPIA, but the Pro, Plus and now II models don't." -- http://www.qbik.ch/usb/devices/showdev.php?id=381

17:13:35 <larsbot> aaronsw, not sure what you mean; are you thinking of the XML file I scraped?

17:13:40 <AaronSw> yeah

17:13:49 * DanC is very tempted to noodle on mailing list schemas...

17:14:23 <larsbot> if you give me your email I'll send it to you

17:14:31 <larsbot> never put it on the web for reasons of internal politics

17:14:32 <AaronSw> so it's not on the web?

17:14:35 <AaronSw> hmm

17:14:40 <AaronSw> ok, me@aaronsw.com

17:15:26 <larsbot> sent

17:19:02 <danb_desk> mailing list schemas: not this week, for me. Maybe next...

17:45:02 <chaals> bye all...

17:45:03 <chaals> dinner time.

18:30:50 <DanC> hmm... how are cwm formulas like cyc microtheories...

18:30:56 <DanC> esp w.r.t. background assumptions.

18:31:15 <DanC> more noodling on context...

18:31:20 <DanC> log:semantics and indexicals...

18:34:09 <DanC> . http://www.cyc.com/cyc-2-1/vocab/mt-vocab.html

18:34:34 <DanC> I think daml:imports is like cycContext:genlMt

18:34:54 <DanC> ah.. there's the thing I'm looking for: http://www.cyc.com/cyc-2-1/vocab/mt-vocab.html#domainAssumptions

18:35:56 <libby> really leaving now!

18:36:24 <libby> oops

18:47:00 <DanC> hmm.. how is the HTTP Date of a message like a domainAssumption?

18:47:49 <DanC> e.g. "How old are you?" "I'm 10". <- in the reply, there's a domain assumption that the year is 2002.

18:48:48 <mnot> it's the date of the representation; the resource may change over time. nice.

18:49:20 <mnot> of course, some proxies may mistakenly update the date, etc., etc. *bluergh*, but nice.

18:50:47 <DanC> er... it's the date of the message.

18:51:12 <DanC> which turns into a background assumption for the assertions in the content.

18:51:51 <mnot> what I said ;)

18:51:58 <DanC> ok.

18:52:18 <DanC> this noodling is prompted by Massimo's call for formalism in TAG stuff.

18:52:43 <DanC>http://lists.w3.org/Archives/Public/www-tag/2002Jan/0077.html

18:52:43 <dc_rdfig> F: http://lists.w3.org/Archives/Public/www-tag/2002Jan/0077.html from DanC

18:52:53 <DanC> F:Foundational Web Model(s)

18:52:53 <dc_rdfig> commented item F

18:53:07 <DanC> F:from Massimo; a call for formalism in Web Architecture writings

18:53:07 <dc_rdfig> commented item F

18:58:41 <DanC> F:the [cyc context vocab|http://www.cyc.com/cyc-2-1/vocab/mt-vocab.html] seems like a good treatment of logic-meets-communicaiton-protocol

18:58:42 <dc_rdfig> commented item F

18:59:13 <DanC> F:i.e. a reasonable approach to "flattening" a model like [my HTTP larch model|http://www.w3.org/XML/9711theory/HTTP]

18:59:13 <dc_rdfig> commented item F

18:59:41 <DanC> F:meanwhile, cyc's logic is pretty far out in the weeds (higher order, nonmon, not much specification)

18:59:42 <dc_rdfig> commented item F

19:13:39 * DanC fixes Notation3 home page to point to my specification of N3 contexts in terms of KIF, i.e. FOPL and quoting

19:18:00 * DanC is tempted to take a serious whack at http://www.w3.org/DesignIssues/Logic.html

19:27:05 <DanC> F:|Foundational Web Model(s)

19:27:05 <dc_rdfig> titled item F

19:27:28 <DanC> F:hmm... "PRA seems adequate once we are committed to a syntactic approach to verifying translators." -- http://www-unix.mcs.anl.gov/qed/mail-archive/volume-2/0017.html

19:27:28 <dc_rdfig> commented item F

19:27:44 <DanC> F:yes... fundamentally, we communicate thru the web with documents. (i.e. syntax)

19:27:45 <dc_rdfig> commented item F

19:30:49 <DanC> F:cf my notes from 1998 on [learning about QED|http://www.w3.org/Math/QED]

19:30:49 <dc_rdfig> commented item F

19:32:35 <DanC> F:er... those notes date rom 1994, originally. cf [my msg to the qed list|http://www-unix.mcs.anl.gov/qed/mail-archive/volume-2/0013.html]

19:32:35 <dc_rdfig> commented item F

19:36:54 <DanC> F:updated QED stuff includes [The Root Logic---Some Technical Details|http://www.rbjones.com/rbjpub/logic/qedres05.htm]

19:36:55 <dc_rdfig> commented item F

19:37:43 <DanC> F:once again, Peter Suber to the rescue with [Recursive Function Theory|http://www.earlham.edu/~peters/courses/logsys/recursiv.htm]

19:37:44 <dc_rdfig> commented item F

19:52:39 <DanC> logger_1, pointer

19:52:39 <DanC> See http://ilrt.org/discovery/chatlogs/rdfig/2002-01-11#T19-52-39

21:05:32 <danb_desk> hey lib!

21:05:51 <libby> hia!

21:25:14 * DanC plays around defining PRA in N3...

21:54:58 <DanC> aha... seems to be working:

21:55:00 <DanC> [ a :PairOfNatural;

21:55:00 <DanC> :id21 "1";

21:55:00 <DanC> :id22 "3";

21:55:00 <DanC> :plus "4";

21:55:00 <DanC> :plusAux "4" ].

23:49:49 * DanCon is having net troubles...