WWW2003 Conference IRC Chat Logs for 2003-05-21
WWW2003 Conference IRC Chat Logs for 2003-05-21 |
|
This web site provides the chat logs for the WWW2003 Community Hub for attendee-generated notes and coverage of the Twelfth International World Wide Web Conference.
You are here: Logs Home / 2003-05-21
06:22:53 * dajobe waves from www2003 breakfast
06:27:09 <dajobe>http://www.w3.org/2003/05/12-director-patent-decision-public.html
06:27:10 <dc_www2003> A: http://www.w3.org/2003/05/12-director-patent-decision-public.html from dajobe
06:27:26 <dajobe> A:|W3C Announces (Royalty Free) Patent Policy
06:27:26 <dc_www2003> Titled item A.
06:28:31 <dajobe> A:[http://www.w3.org/2003/05/patentpolicy-pressrelease|press release announcement]
06:28:31 <dc_www2003> Added comment A1.
06:34:20 <dajobe>http://www2003.org/schedule.htm#21
06:34:20 <dc_www2003> B: http://www2003.org/schedule.htm#21 from dajobe
06:34:33 <dajobe> B:|Schedule for today (Wednesday)
06:34:33 <dc_www2003> Titled item B.
06:54:40 <edd> edd has changed the topic to: WWW2003, logged and blogged at http://www2003.xmlhack.com/ - Budapest, Hungary - May 20-24 - First day of main conference - Keynotes
06:58:08 <edd> hey ma
06:58:10 <edd> ttb
06:59:54 <dajobe> conference beginning now...
07:00:08 <mattb> morning all
07:00:08 <dajobe> we watch an animation of previous venues of the conferrence
07:00:15 <dajobe> looks like SVG / SMIL
07:00:23 <dajobe> yes, SVG
07:00:35 <dajobe> 09:00 UTC+2 Hungarian local time
07:02:39 <dajobe> welcome speeches
07:02:56 <dajobe> Minister of Information welcomes to the conference
07:03:04 <dajobe> edd can type the accents
07:03:19 <edd> Kálmán Kovács
07:03:23 <dajobe> Minister of Informatics and Communications
07:03:56 <dajobe> "century of Jon von Neuman's birth"
07:04:02 <dajobe> (spelling, yeah)
07:04:26 <dajobe> "we are proud of von Neumann"
07:05:16 <dajobe> discussion of IT in Hungary
07:05:26 <dajobe> description
07:05:47 <dajobe> new ministry to create info society
07:06:05 <dajobe> goals to increase internet penetration in HU
07:06:22 <dajobe> now increased +16% use in 4 months
07:06:31 <dajobe> developing HU language content
07:07:27 <dajobe> e-commerce act, dig sigs
07:07:49 <dajobe> internet has given great, revolutionary change in last 10 years
07:07:59 <mdupont> cool, the hungarians have always been proud of thier education system
07:08:01 <dajobe> birth of knowledge-based societ
07:08:02 <dajobe> y
07:08:21 <dajobe> "we are especially optomistic... a favourable area for us"
07:08:30 <mdupont> a HU. college of mine in germany wants to take his kids back to HU so that they get a "real" education
07:08:41 <dajobe> and want to keep and encourage our young bright people they have a great future in HU
07:09:00 <mdupont> i think they do, when is the planned entry into the EU?
07:09:08 <dajobe> "we wish you a successful week, and a seamless broadband connection"
07:09:10 <dajobe> ---
07:09:18 <dajobe> [aplause]
07:09:41 * mdupont shuts up
07:09:52 <edd> mdupont: i think he said next week
07:09:58 <dajobe> yeah
07:10:39 <edd> "Hungary will bring the brain to the Union"
07:10:50 <mdupont> edd, wha is next week? the entry? no way that quickly?
07:10:51 <dajobe> "have 12 nobel prize winners"
07:11:01 <dajobe> they've had the vote and said yes, IIRC
07:11:40 <dajobe> slides of HU nobel prize winners
07:11:53 <dajobe> live music
07:12:06 <mdupont> i wish there were some types of webcam or video of the proceedings
07:12:26 <dajobe> "the princess"?
07:12:32 <dajobe> three violinists
07:12:33 <mdupont> that is what we need, some type of online conference
07:13:23 <dajobe> classical music with a techno beat
07:13:26 <dajobe> la la la....
07:13:54 <dajobe> foot tapping stuff
07:15:30 <dajobe> [applause]
07:18:41 <edd> some colour: all but the most hardened geeks have looked up from their laptops :)
07:19:57 <dajobe> it's a bit liked "hooked on classics" for those who recognise that style
07:25:04 <sandro> It's awefully catchy.
07:25:17 <sandro> But the second peice was much better.
07:25:42 <sandro> anyone know the names of the peices?
07:26:43 <edd> Bebo White now talking
07:26:46 <dajobe> ---
07:26:49 <edd> Welcoming on behalf of the conference committee
07:27:17 <edd> Explaining who the IW3C2 is.
07:27:40 <mattb>http://princess.musichello.com/about.html
07:27:41 <dc_www2003> C: http://princess.musichello.com/about.html from mattb
07:27:43 <mattb> C:|Princess
07:27:43 <dc_www2003> Titled item C.
07:27:50 <dajobe> by the power of google :)
07:27:53 <mattb> :)
07:28:05 <mattb> C:musical opening keynote
07:28:05 <dc_www2003> Added comment C1.
07:28:37 <edd> Bebo is the conference co-chair
07:28:42 <edd> Explains some admin details
07:30:04 <edd> Over 700 papers submitted, over 300 reviewers.
07:30:07 <bijan> Lots and lots of admin details :)
07:30:11 <edd> 12% of papers accepted.
07:31:54 <edd> Over 200 posters.
07:34:32 <dajobe> </Bebo>
07:34:47 <edd> back to Gusztáv Hencsey, co-chair
07:35:34 <sandro> calls Conference Organizing possibly the oldest profession
07:35:39 <dajobe> 800 participants, 57 countries
07:35:40 <edd> 800 participants from 57 countries
07:36:07 <dajobe> compares to PPP10003 conference -re-historric people's conference
07:37:04 <edd> wireless net access in adjacent park
07:37:33 <edd> "the pluggest conference in the history of the WWW" - meeting complaints at lack of power outlets
07:38:00 <edd> explains "users pad" - the cushion in the reg packl
07:38:32 <edd> make it easier to sit on the marble floors
07:39:31 <dajobe> lol
07:39:37 <sandro> Downloading the demo princess mp3s at >200 megabytes/sec....
07:39:42 <dajobe> "invite to the stage, our 'Shaman' o the web, TBL"
07:39:46 * edd hands scribe to da
07:39:47 <edd> jobe
07:39:47 <dajobe> sandro: pls don't :)
07:39:54 <dajobe> timbl
07:39:58 <dajobe> plenary
07:40:02 <dajobe> grab that url edd
07:40:09 <dajobe> web services and semantic web intgegrating apps
07:40:15 <dajobe> ^title
07:40:20 <edd>http://www.w3.org/2003/Talks/05-www-keynote-tbl
07:40:21 <dc_www2003> D: http://www.w3.org/2003/Talks/05-www-keynote-tbl from edd
07:40:23 <dajobe> 2 things to announce
07:40:24 <sandro> dajobe, seriously, it took 2 seconds. I wonder if there's a transparent cache? Or it's just that we're very network near them. 4 hops.
07:40:25 <dajobe> 1)
07:40:33 <dajobe> socially logical andpolitical side of web devels
07:40:37 <dajobe> 2) technical
07:40:37 <dajobe> --
07:40:46 <dajobe> 1 that about the openeness of thed evelopment about which we do
07:40:53 <dajobe> everitything that is web infrastructure
07:40:55 <edd> D:|Tim Berners Lee keynote
07:40:56 <dc_www2003> Titled item D.
07:41:05 <dajobe> the w3c which may has started technically after the 1st conf
07:41:11 <dajobe> It must have started approx. sep 94
07:41:16 <dajobe> and 2nd conf oct
07:41:18 <edd> D:Slides a 404 now, but this URI was on his opening slide
07:41:18 <dc_www2003> Added comment D1.
07:41:37 <dajobe> edd: he usuallyu doesn't commit them to the web beforehand
07:41:49 <dajobe> ... dot com boom
07:42:11 <dajobe> was the question of whether companies should have developed closed web tech or open at that time
07:42:15 <dajobe> patenting stuff then
07:42:23 <dajobe> but that wasn't what jhappeneed
07:42:39 <dajobe> ... working towards a royalty free web standards
07:42:53 <dajobe> The w3c has tackled the question of royalty free web standards
07:42:57 <dajobe> ref to the PPWG
07:43:07 <dajobe> edd: can you find urls for PPWG and other things I mention?
07:43:14 <dajobe> ... first thing is
07:43:24 <dajobe> ... the PP of the W3c has now been approved
07:43:25 <bijan> TBL called W3C a "standards body"
07:43:25 <dajobe> seen A:
07:43:33 <dajobe> press event later today
07:43:43 <dajobe> tbl: well done to everyone involved in PPWG
07:44:05 <dajobe> now the technical work can be done with out worrying about patente
07:44:14 <dajobe> thanskaskedfort danny weitzer
07:44:18 <dajobe> thanks
07:44:24 <dajobe> [applause]
07:44:51 <bijan> jjc shouts "royalty free" from the audience
07:44:57 <dajobe> lol
07:45:02 <dajobe> timbl....
07:45:09 <dajobe> ... guided by questions peolea sk me
07:45:19 <dajobe> ... we've heard about semantic web (SW) and web services (WS)
07:45:23 <dajobe> ... does one replace the other
07:45:25 <dajobe> ?
07:45:32 <dajobe> so first I'm talking about one, then the other
07:45:42 <dajobe> mainly about WS and SW
07:45:57 <dajobe> timbl: corrects slide url
07:46:04 <dajobe> ... it's about integration
07:46:08 <dajobe> slide 2
07:46:35 <mattb> did anyone get that slide url correction?
07:46:46 <dajobe> integration across app boundaries and company boundaries drivees
07:46:49 * edd will do
07:46:53 <dajobe> WS and SW - program for WS and data for SW
07:47:41 <PaulShab> URL http://www.w3.org/2003/Talks/0521-www-keynote-tbl/
07:47:57 <edd> D:=http://www.w3.org/2003/Talks/0521-www-keynote-tbl/
07:47:59 <dc_www2003> Replaced URL of D.
07:48:03 <edd> D1:""
07:48:03 <dc_www2003> Deleted comment D1.
07:48:36 <sandro> lol
07:48:47 <dajobe> slide3
07:48:52 <dajobe> slide4
07:48:56 <dajobe> how do they work?
07:49:01 <dajobe> the trick is the interface
07:49:15 <dajobe> when you connect progs together you have a dangerous funcitonal point
07:49:39 <dajobe> such as "the .h file" - a file that defines the sets of calls and params
07:49:54 <dajobe> tradtiionally everyone has to update to new i/fs - everyone change
07:49:58 <dajobe> but try to avoid doing that
07:50:14 <dajobe> the inteegration problem could be said to making that i/f more flexible
07:50:43 <dajobe> may not be best thought of as an Remote Procedure Call (RPC)
07:50:49 <dajobe> it might take some days to reply
07:51:16 <dajobe> the params may now be more complex, so we use XML to send the op,q uery, params and get results
07:51:26 <dajobe> SOAP can be used in RPC style, "maybe the old style"
07:51:36 <dajobe> rather than calling a procedure...
07:51:42 * DanC cheers on dajobe
07:51:46 <dajobe> ... you get more flexibility with XML tools
07:51:51 <dajobe> [hmm, lost that thread]
07:52:06 <dajobe> slide5
07:52:15 <dajobe> a slide of acronyms :)
07:52:36 <dajobe> SOAP now Proposed Rec stage (PR)
07:52:39 <dajobe> .. a few days ago
07:52:44 <dajobe> SOAP 1.2 now very stable
07:52:51 <dajobe> if you are using 1.0, 1.2. "now is the time to upgrade"
07:53:05 <dajobe> not just polished, it's changed - can be used in new ways - better
07:53:14 <dajobe> security &c
07:53:38 <dajobe> slide6
07:53:43 <dajobe> WS stack
07:53:50 <mattb> gosh, the colour repro on that projector is terrible
07:54:06 <mattb> maybe the red gun is dead
07:54:10 <dajobe> looks like it
07:54:35 <dajobe> there it is http://www.w3.org/DesignIssues/diagrams/ws-stack.png
07:55:07 <dajobe> dance of choreography of service dependndies in a transaction
07:55:18 <dajobe> when building large services from smaller WSes
07:55:26 <dajobe> most build on WSDL or XPath, as far as I've seen
07:55:39 <dajobe> to refer to a given part of a doc
07:55:43 <dajobe> it's just one part, it is a bit of a weak oint, scripting...
07:55:49 <dajobe> you talk about the sequence of things that you do
07:56:24 <dajobe> correlation of terms in different layers such as the same value of item purcashed
07:56:24 <dajobe> purchas order
07:57:05 <dajobe> matching telement contents in different docs
07:57:09 <dajobe> using xpath
07:57:19 <dajobe> some fragility on XML dtd or schema, design of the XML docs
07:57:23 <edd> D:[Start of notes on the talk|http://ilrt.org/discovery/chatlogs/www2003/2003-05-21.html#T07-40-09]
07:57:23 <dc_www2003> Added comment D1.
07:57:34 <dajobe> if you took the same info in a different dtd, you'd get a whole difernt thing you'd have to re-express
07:57:43 <dajobe> pointing at the syntax of an XML doc
07:57:55 <dajobe> purchase order is a document
07:58:00 <sandro> D:|Tim Berners-Lee keynote
07:58:00 <dc_www2003> Titled item D.
07:58:20 <dajobe> SW would come in to somewhat describe the contents [something like thios]
07:58:34 <dajobe> timbl: this pic not all the story
07:58:50 <dajobe> slide7
07:59:21 <dajobe> notes that there are other things oon slide6 above http such as reliable messaging
07:59:34 <dajobe> slide7 WS benefits
07:59:42 <sandro> tim points to things on his screen, but not using a mouse or laser pointer, so we can't tell what he's pointing to. :-)
07:59:50 <dajobe> hope is to have one rather thnn two, many sets of WS spec
07:59:53 <dajobe> s
08:00:15 <dajobe> tension between fast in a closed group, or in a more open and slower
08:00:26 <dajobe> always asking me about DCOM & CORBA again
08:00:32 <dajobe> but ok, computers are fast now
08:00:38 <dajobe> so hey, you can make the WS stacks twice
08:00:49 <dajobe> but as engineers, but we refer to that as "a pain"
08:00:57 <sandro> "it's what we actually refer to as 'a pain'"
08:01:02 <edd> [ Calls for one set of WS standards, stressed important that people are
08:01:02 <edd> willing to change and work in the open. Acks that open work will take
08:01:02 <edd> longer but you end up with one spec only ]
08:01:04 <dajobe> there is a small threshold to get involved in this - it's modular
08:01:35 <dajobe> slide8 what is the semantic web?
08:01:48 <dajobe> data integration across app, org boundaries
08:01:52 <dajobe> slide9
08:02:00 <dajobe> most data is relational data
08:02:14 <dajobe> rdf is just relational data, the base semweb standard
08:02:19 <edd> says should have called RDF relational data format :)
08:02:23 <dajobe> saying that when you have a tablel of data, thje column is a URI
08:02:39 <dajobe> when you use URIs, you move uip a level of abstraction
08:02:46 <dajobe> not talking about document syntax, you are talking about real things
08:02:48 <dajobe> slide10
08:02:50 <dajobe> real things
08:03:19 <dajobe> compares xml & rdf on slide10
08:03:34 <DanC> hmm... this slide could use a couple pictures
08:03:45 <DanC> likewise the purchase order XPath slide
08:03:47 <dajobe> the rdf data is much more reusable, since anybody can writes software can reusue this data
08:04:19 <dajobe> it can use the info that the number is unique and unambiguous
08:04:23 * DanC noodles on taking that "relational data format" idea seriously...
08:04:29 <dajobe> the piece from the XML schema only tells you about documents written to that schema
08:04:51 <dajobe> so we are coding a lot of arbitrary knowledge in how that schema & document wre designed
08:04:55 <bijan> DanC: see datalog
08:05:00 <dajobe> so wasting time dealing with info rhtat is irrelevant
08:05:03 <dajobe> slide11 SW features
08:05:10 <dajobe> treasts relational and semi-structured data uniformly
08:05:31 <dajobe> I get a lot of people asking about isng a relational or a new-fangled XMld atabase
08:05:34 * sandro imagines who would be really annoyed with that renaming -- experts in relational data who've been "left out" of the RDF development process.
08:05:39 <dajobe> from a semweb PoV, it's the samething pretty much
08:05:54 <dajobe> once you've taking out the arbitraryness and realise you are talking about real athings
08:06:00 <dajobe> you can treat them both the same
08:06:03 <dajobe> reules
08:06:06 <dajobe> rules
08:06:13 <sandro> TimBL makes "RDF Databases" sound vaguely off-the-shelf.
08:06:29 <dajobe> smart queries that go off to other places to answers
08:06:31 <dajobe> slide12
08:06:32 <dajobe> SW benefits
08:06:35 <dajobe> provably correct
08:06:35 <dajobe> (!)
08:06:47 <dajobe> much faster app integration
08:06:58 <dajobe> "what's the bang for the buck?"
08:07:09 <dajobe> answer is slide12
08:07:34 * sandro wonders how many people are able to follow TimBL's speech style with this material.....
08:07:36 <dajobe> at the moment semantic integration, and you have to think agaead when there is ahuge amount of public data out there you can integrate and link to
08:07:54 <dajobe> designed so that the langs evolve in a web-style way
08:08:14 <dajobe> can design apps in diff parts of the world that can be connected without making one huge system, can connect in a web-wide weay
08:08:16 <dajobe> slide13
08:08:35 * DanC imagines an instrument, ala a speedometer, that measures words-per-minute, or sylables-per-minute...
08:08:38 <dajobe> URIs ra ra ra
08:08:44 <dajobe> slide14
08:09:02 <dajobe> fundamental building block
08:09:03 <dajobe> slide15
08:09:14 <dajobe> slide16
08:09:23 <dajobe> 17, 18
08:09:29 <DanC> "timbl turns from slide 13 to slide 14 at 157 words per minute... blows past that on, on to slide 16, slowing down to 120..."
08:09:40 <dajobe> what you get isn't a tree or a rectangle, it's a semweb
08:09:42 <dajobe> slide19
08:09:52 <dajobe> what is cool with RDF you can't do with XML schema
08:09:59 <dajobe> is that you can concatenate the files to merge
08:10:01 <dajobe> slide19,20
08:10:14 <dajobe> nodes & identiers
08:10:17 <dajobe> saves you lots of time
08:10:18 <dajobe> slide21
08:10:24 * edd tries to get overarching themes from this talk: apart from patent policy announcement (hurray), think some attempt to show WS and SW as complementary.
08:10:24 <dajobe> joining the web using the same uri
08:10:27 <dajobe> for theproperty name
08:10:32 <dajobe> zip column names
08:10:40 <dajobe> happen to mean US zip codes but the computer doesn't know that
08:10:51 <dajobe> first turn into uris and then say they are equiv
08:10:52 <dajobe> slide22s
08:10:55 <dajobe> apps connect
08:11:02 <dajobe> sometimes I'm acused of making a huge global ontology
08:11:03 <dajobe> such as cyc
08:11:06 <dajobe> that isn't going to work
08:11:08 <edd> Dangling lots of SW concepts in front of folk to get them thinking.
08:11:10 <dajobe> we know society is web like
08:11:19 <dajobe> some concepts are shared, some dont
08:11:24 <dajobe> it's about partiallly sharing concepts
08:11:30 <dajobe> when you communicate with somebody
08:11:34 <dajobe> you start to share with them
08:11:40 <dajobe> there are some classes of apps that naturally s hare
08:11:48 <dajobe> if you look at slide22
08:11:53 <dajobe> in yuour addressbook you have addresses
08:11:58 <dajobe> maps about places
08:12:02 <dajobe> conferences have events
08:12:06 <dajobe> travel agents places
08:12:15 <dajobe> all those apps can uisefully be integrated by sharing info about place
08:12:24 <dajobe> after 9/11, they took a lot of info about maps
08:12:36 <dajobe> and recorded it via place
08:12:48 <dajobe> "he was jumping up and down and foaming at the mouth|
08:12:50 <dajobe> "
08:12:56 <dajobe> he found it so cool to connect things via place.
08:13:04 <dajobe> When you take a digi picture, it just knows a pic.
08:13:13 <dajobe> but your calendar knows where & when
08:13:18 <dajobe> so if they were integrated...
08:13:24 <edd> side ref to geo semweb work
08:13:31 * bijan sees the new semweb slogan: "Maying TIA easy for everyone"
08:13:33 <dajobe> in pic, the 'part number' item is not connected to everything
08:13:44 <dajobe> slide23
08:13:48 <dajobe> Web of Data
08:13:52 <dajobe> some item is very personal
08:13:59 <dajobe> s/items/items of data/
08:14:13 <dajobe> a lot of interesting data is global
08:14:22 <dajobe> but the personal is interesting and confidential
08:14:27 <dajobe> refgs to his tax data
08:14:37 <dajobe> worked using rules in rdf
08:14:41 <dajobe> but clearly can't put on the web
08:14:48 <dajobe> ddata from diff areas must be trteated very differently
08:14:53 <dajobe> access levels
08:15:03 <dajobe> like the web,
08:15:14 <dajobe> asks who's bneen there from 1st web conf - a few hands
08:15:29 <dajobe> at that point realised there was a group of stuff them connected
08:15:34 <dajobe> really was exponential growth
08:15:38 <dajobe> e are at the very early stage of that
08:15:41 <dajobe> slide24 connections
08:15:44 <dajobe> Your life is web
08:15:49 <dajobe> Your data is a web.
08:16:03 <dajobe> slide25
08:16:10 <dajobe> if you hagve XML based apps
08:16:17 <dajobe> for sucking data out of one app and pushing into another app
08:16:27 <dajobe> you may notice that you are using programmer time to move data, query data
08:16:34 <dajobe> (the yellow circualr blobs on slide25)
08:16:39 <dajobe> want to try to minimse that
08:16:41 <dajobe> slide26
08:16:45 <dajobe> RDF idea is that uyou have a hub
08:16:51 <edd> eek, the word "Enterprise" appears :)
08:17:05 <dajobe> not necess one file, can be one conceptual file
08:17:09 <dajobe> so when you pose a question
08:17:11 * zool just popped in to enjoy some budapest-envy
08:17:17 <dajobe> I guess that is how ew have to seell the semweb at this tage
08:17:19 <dajobe> hi zool
08:17:36 <dajobe> the motivation EIS
08:17:38 <dajobe> slide27
08:17:44 <dajobe> they are part of a big global bus
08:17:56 <dajobe> so all they have to do is put a filter in to allow outsidcers to access
08:18:07 <dajobe> so they don't have to ask humans about the details that are recorded in the data
08:18:13 <dajobe> maybe somebvody is asking a complex query
08:18:29 <dajobe> maybe joining the query with other things, you just need to expose the data yuou have in a common format
08:18:33 <dajobe> and allow others to common queries
08:18:44 <dajobe> slide ? Case Tudies
08:18:56 <dajobe> Sun GKE
08:19:07 <edd> That's Uche's work, isn't it?
08:19:14 <dajobe> yes
08:19:33 <dajobe> many sources of data, need to convince them to convert & make avail
08:19:35 <dajobe> into RDF
08:19:39 <dajobe> put it together and repurpose
08:19:46 <dajobe> lots of diff web sites with different uses
08:19:57 <dajobe> custoekmr care are just one
08:20:05 <dajobe> ptr to the paper
08:20:09 <dajobe> HP and MIT library - simile
08:20:26 <dajobe> Dan Connolly's trip to Budapest
08:20:35 <dajobe>http://www.w3.org/2003/Talks/0520-www-tf1-d3-travel/
08:20:35 <dc_www2003> E: http://www.w3.org/2003/Talks/0520-www-tf1-d3-travel/ from dajobe
08:21:03 <sandro> lol
08:21:03 <dajobe> E:| Dan Connolly's trip to Budapest
08:21:03 <dajobe> E:| Dan Connolly's trip to Budapest
08:21:03 <dajobe> E:|Dan Connolly's trip to budapest
08:21:03 <dc_www2003> Titled item E.
08:21:04 <dc_www2003> Titled item E.
08:21:05 <dc_www2003> Titled item E.
08:21:18 <dajobe> points to slide4 where the data is
08:21:19 <dajobe> of E:
08:21:36 <dajobe> danc wants it in his caeldnar
08:21:43 <dajobe> shows the map that comes out of it
08:22:02 <dajobe> timbl refers to using google to find danc
08:22:12 * bijan notes that tbl just used a google ref instead of a uri for danc's home page
08:22:16 <dajobe> back to slide28
08:22:37 <dajobe> the cool thing is integrating apps of things they know
08:22:39 <dajobe> without pain
08:22:43 <edd> DanC: can you suggest a better title for E: ?
08:22:45 <dajobe> DAML
08:22:55 <dajobe> mike dean - nice registry of ontologies and ata
08:23:00 <dajobe> also the data, not just the schemas
08:23:04 <dajobe> for example US states
08:23:18 <dajobe> there are only 50
08:23:26 <DanC> E:|Semantic Web Application Integration: Travel Tools
08:23:26 <dc_www2003> Titled item E.
08:23:35 <dajobe> when you use a terms uch as us state code in a query
08:23:40 <dajobe> you can operate a semweb query in a model
08:23:41 <dajobe> mode
08:23:56 <dajobe> then if something is a name of a city, ina state, bordering something with state code of "MA"
08:24:02 <dajobe> then print it out
08:24:06 <dajobe> you can give that query to the machine
08:24:09 <sandro> This must be very strange for people who attended yesterday's tutorial, where Dan and I talked about this stuff.
08:24:18 <dajobe> but if the machine can use the web, and the schema on the web, the uIR
08:24:31 <dajobe> URI on the web - (this is a web conference) - it can go look it up
08:24:39 <dajobe> it also finds a pointer to say, this is a definitive doc
08:24:55 <dajobe> so the query doc can go resolve that knowlege
08:25:00 <dajobe> "that's pretty neat"
08:25:15 <dajobe> "the concept of the self-answering question"
08:25:16 <dajobe> slide29
08:25:21 <dajobe> Stack of expressive power
08:25:25 <dajobe> [sigh]
08:25:36 <libby> how did it go sandro?
08:26:02 * dajobe notices alberto on Rendezvous
08:26:14 <dajobe> semantic web wave slide
08:26:21 <dajobe> ^slide30
08:26:25 <alberto> hi dajobe
08:26:28 <mattb> time for a semantic web mexican wave
08:26:40 <mattb> then we can have a webservices one, in the other direction
08:26:43 <mattb> we'll meet in the middle
08:26:55 <libby> heheh
08:26:58 <sandro> I think it went well, libby, all things considered. But maybe you should ask Andy or someone in the class.
08:27:05 <libby> I will :)
08:27:07 <dajobe> a bit of review of stds appearing and moving on to deployent
08:27:32 <libby> are there any slides sandro?
08:27:34 <dajobe> [at this rate he has 7 slides in 3 mins to go?]
08:27:51 <dajobe> OWL the ontology language, more expressive
08:28:00 <dajobe> giving useful info on vocabs, fairly mature in std process
08:28:02 <dajobe> rules and query
08:28:08 <sandro> thousands of them, libby. :-) one sec
08:28:10 <dajobe> we have xml query, we have lots rtdf query langs 8-,9,10\
08:28:17 <dajobe> looks like it is right for standardisation
08:28:19 <libby> thanks sandro
08:28:21 <dajobe> looks like we're going to start that
08:28:24 <dajobe> ontologiy
08:28:38 <sandro>http://www.w3.org/2000/10/swap/doc/
08:28:38 <dc_www2003> F: http://www.w3.org/2000/10/swap/doc/ from sandro
08:28:42 <dajobe> not sure of speed, maybe dfaster than many people think
08:28:43 <edd> hi makoto
08:28:45 <dajobe> slide31
08:28:50 <dajobe> W3C standardisation
08:29:02 <libby> ta
08:29:04 <makoto> Hi edd.
08:29:07 <sandro> F:| Tutorial, with text and slides, covering much of what Tim's talking about
08:29:08 <dc_www2003> Titled item F.
08:29:39 <dajobe> w3c ra ra ra
08:29:48 * dajobe grins
08:30:18 * edd would like to see mention of tension between research and standardization within W3C
08:30:38 <edd> s/tension/balance/ if you like
08:30:55 * nmg nods at edd
08:31:05 <dajobe> slide32
08:31:10 <dajobe> peole ask what yoould I do?
08:31:17 <dajobe> don't jump on the sw bandwidth too early
08:31:24 <dajobe> mixture of acadamic, industrial, somewhere in between
08:31:34 <dajobe> somet of these things are exiting, still not much shrinkwarp
08:31:40 <dajobe> action -a ske products input and ouput rdf
08:31:42 <dajobe> get involved
08:31:51 <dajobe> s/a ske/ask for/
08:32:04 <dajobe> model in rdf, not the xml level
08:32:05 <edd> he's talking about what the IT sector should do to get ready for the Semantic Webn
08:32:06 * DanC wonders how long he's supposed to talk for
08:32:06 <dajobe> slide33
08:32:10 * bijan shudders at adding "tutorial/teaching" to the wg mandate
08:32:14 <dajobe> customer, demand rdf
08:32:19 <dajobe> "I want my data back"
08:32:27 <dajobe> DanC: he's 2 mins over, IIRC
08:32:33 <dajobe> Jena - HP
08:32:38 <dajobe> 1.0 stand, 2.0 rewrite out soon
08:32:40 <dajobe> in beta
08:32:55 <edd> And also whatthe customer should do. Demand "I want my data back". amen to that
08:32:56 <bijan> Hmm. Isn't that a bit of a winner picking?
08:32:59 <dajobe> "almost said bayta, but I'm in surope"
08:33:03 <dajobe> slide34
08:33:09 <dajobe> connections WS-SW
08:33:15 <dajobe> discovery is just SW
08:33:24 <dajobe> I hope all the work in UDDI will be xposed in the semweb
08:33:36 <dajobe> you can take a WS query and service it
08:33:46 <dajobe> realy interesting in services you use, things you buy, not WS themselves
08:33:59 <sandro> bijan - it might be, but it goes a huge way to making it real to many folks just to hear there is a solution from HP, even if they never use it.
08:34:01 <dajobe> WSDL not an rdf app, some folks think it should be
08:34:13 <dajobe> ... some worked to do that [that would be bijan]
08:34:19 <dajobe> RDF as SOAP
08:34:38 <dajobe> sending a well defined semantically enmabled message
08:34:47 <dajobe> RDF query plugs in wherever mlquery or xpath is used
08:34:48 <dajobe> [hmm!]
08:35:04 <dajobe> sw busnines engines should be a ble to call out using SOAP to use WS
08:35:18 <dajobe> s/sw bus/ws/
08:35:21 <dajobe> s/use WS/use SW/
08:35:29 <dajobe> slide35
08:35:33 <dajobe> connetions interactive semweb
08:35:44 <dajobe> interactive
08:35:51 <dajobe> voice browses
08:36:06 <dajobe> semweb data out there in a abstract level means you can represent it in many opther ways
08:36:13 <dajobe> in communicationw ith a computer
08:36:18 <dajobe> extra reusability of semweb data
08:36:31 <dajobe> may be the only practical way to builda practical system of diverse devicesg
08:36:38 <dajobe> great for accessibility
08:37:02 <dajobe> empowering pizza
08:37:09 * edd searches for PizzaML
08:37:19 <dajobe> top level of semweb
08:37:26 <dajobe> web of trust
08:37:34 <dajobe> gating factor on collab work
08:37:44 <dajobe> can't do it until we can say who's allowed, privacy aspects
08:37:47 <dajobe> slide36 conclusions
08:37:50 <dajobe> ws and sw connect together
08:37:54 <dajobe> bults separately
08:38:01 <dajobe> because it's too hard to do in one lump
08:38:08 <dajobe> connections exist such as daml-s, there will be be more
08:38:19 <dajobe> we aren't constraining the standards of them should all be ionterocnnected
08:38:22 <dajobe> just by pragmatics
08:38:36 <dajobe> WS meet an immediate need, build a acool app quickly
08:38:49 <dajobe> SW you tend not to see the gain, since it's a web app, until there is a lot of data out there
08:38:52 <dajobe> [the network effect]
08:39:04 <dajobe> out of things such as "why should I use your stupd synjtax|" - phase 1
08:39:20 <dajobe> phase2 now we can use common tools, formats, we can see stds are useful
08:39:27 <mattb> ouch
08:39:30 <dajobe> pharse3 is "WOW! The whole workld is onm my laptop"
08:39:40 <mattb> phase3 is loud
08:39:43 <dajobe> we've seen these three phraes for the web
08:39:43 * edd 's eardrum bursts
08:39:45 * dajobe stands back
08:39:49 <dajobe> Thank you
08:39:50 <dajobe> [applause]
08:40:01 * DanC low on battery...
08:40:13 <dajobe> timbl exits stage left
08:40:15 <dajobe> and so do we...
08:40:24 <edd> --- break ---
08:40:33 * bijan notes that tbl didnt' use a uri on his last slide for the w3c home page
08:40:42 <bijan> "w3.org"
08:48:58 * DanC wonders if my net connection survived...
08:55:36 <DanC> TimBL: when you hit phase III, don't invest. think!
08:58:16 <DanC> ......
08:58:19 <DanC> pub 1024D/6E52C29E 2002-05-07 Dan Connolly (2002-05) <connolly@w3.org>
08:58:19 <DanC> Key fingerprint = D3C2 887B 0F92 6005 C541 0875 0F91 96DE 6E52 C29E
08:58:21 <DanC> ......
09:00:00 <DanC> hmm... now what would it take for an attacker to screw that up? (a) there are a bunch of people here who can pretty much tell that this is me,
09:00:13 <DanC> there are public logs of the channel, as well as google caches etc.
09:00:54 <DanC> you'd have to attack a lot of stuff to conduct a man-in-the-middle attack
09:01:02 <DanC> hmm... I should put my key in a keyserver somewhere...
09:03:37 <edd> power outage in the Bartok room laptop supply :)
09:04:40 <dirkx> Aye whole room lost power
09:06:00 <edd> Google talk in Bartok room on generating related news from closed captions on TV
09:06:28 <edd> Some challenges: captions ungrammatical, mixture between content and metacontent, frequent topic changes
09:07:27 <dirkx> edd: but at the same time significant structure and hints as to what is up, what the format is; i.e. weather will be last usually :-)
09:09:03 <ks7687_> ks7687_ is now known as kateS
09:09:06 <edd> Discussion of algorithms used to extract queries from closed captions.
09:09:16 <DanC> "foundations" session is packed.
09:09:35 <edd> Yeah, Google session packed too.
09:13:11 <edd>http://www.hackdiary.com/archives/000032.html
09:13:12 <dc_www2003> G: http://www.hackdiary.com/archives/000032.html from edd
09:13:30 <edd> G:|First WWW2003 blogger: Matt Biddulph's running account of Day 1
09:13:30 <dc_www2003> Titled item G.
09:13:46 <libby> wow, that's fast!
09:14:12 <darobin> edd: any mention of TT in the caption stuff?
09:14:25 <libby> [[Bah, I find a powersocket and then the room loses power to its plugstrips. Puzzled people jiggle their cables all over the room]] bah
09:14:27 <edd> darobin: TT?
09:14:36 <libby> timedtext?
09:14:42 <darobin> Timed Text, new WG, part of SMIL Activity
09:15:01 <edd> darobin: no, the context is that they're using the closed captions as source material for query generation
09:15:09 <edd> the talk is about search strategies
09:15:15 <mattb> caption stuff not really being mentioned except as a source of freetext to autoextract queries from
09:15:22 <darobin> ah ok
09:15:29 <darobin> gotcha, thanks
09:15:33 <mattb> processing text segment by segment, keeping a running track of similarity to track topic change
09:15:36 <mattb> and weight terms over time
09:15:55 <edd> i'd be interested to know how much of this relies on the US style of captioning
09:16:00 <darobin> it would be nice if the video also had MPEG-7 metadata to be correlated with the captions :)
09:16:12 <edd> if they'd get significantly different results from British TV, for instance
09:16:30 * danbri waves from front/floor of Google news talk
09:16:37 <sandro> has anyone seen Pat Hayes here?
09:16:40 <darobin> that's likely, captioning changes a lot from place to place, and even from channel to channel sometimes
09:16:41 <libby> are you using your mat?
09:16:42 * mattb waves from about 3 metres away from danbri
09:16:45 <danbri> not yet
09:16:56 * danbri grins x2
09:17:00 <alberto> hello
09:17:07 <libby> heya alberto
09:17:35 * darobin wonders who to complain to about limitations in HTTP now that it's a dead spec
09:18:14 <alberto> being at the google preso - some numbers.....looks interesting
09:18:47 * DanC is getting crummy wireless signal
09:18:53 <DanC> PatH... he's not coming, is he?
09:19:03 <DanC> Hayes isn't in the (hardcopy) list of WWW2003 participants
09:19:05 <DanC> HTTP dead? in what sense?
09:19:28 <dajobe> I'm pretty sure PatH isn't coming here
09:19:35 <libby> boo
09:19:40 <DanC> perhaps in the sense that it's so widely deployed that its momentum is so large that it's practically impossible to change its direction?
09:19:54 <darobin> DanC: dead == stable
09:20:10 <darobin> yes, that and no one working on furthering it
09:21:18 <darobin> I just want parameters to content-codings the way there are parameters to mime types
09:24:03 * DanC doesn't expect complaining about HTTP is constructive
09:24:04 <DanC> if Hayes were here, he'd be pretty annoyed that this guy is repeatedly saying that there are differences between first order model theory and RDF model theory. The RDF model theory *is* first order.
09:26:18 <darobin> DanC: yes, that's why I'm annoyed :)
09:26:39 <sandro> who is talking (in sw) ?
09:37:16 <dirkx> Has anyone seen the W2003 conference as vCards or as RDF ?
09:39:23 <DanC> BLURB: Model-theoretic Semantics for the Web
09:39:23 <dc_www2003> H: Model-theoretic Semantics for the Web from DanC
09:39:28 <DanC> IanH is presenting now
09:39:36 * DanC wonders why dc_www2003 hasn't responded to my BLURB request
09:39:44 <DanC> in the same slide, he says "the relationship isn't really clear" and then says exactly what the relationship between XML and RDF is. odd.
09:39:53 <DanC> IanH: "may facilitate SemWeb to use more standard KR thesis"
09:40:07 <DanC> goodness... lots of words on this slide.
09:40:39 * danbri notices loggger isn't logging /me'd actions, just main text
09:40:43 <edd> DanC: it did
09:41:30 * DanC thought he saw the conference program in RDF-calendar
09:41:30 <DanC> libby?
09:41:30 <DanC> yes, just 13 minutes later
09:41:55 <edd> DanC: i'd ascribe it to flaky wireless :/
09:42:53 <sandro> Cons of FOL which Ian omits: it can't do metedata. search engines are out in the cold!
09:44:27 <danbri>http://esw.w3.org/topic/RdfQueryTestingBudapestMeeting
09:44:28 <dc_www2003> I: http://esw.w3.org/topic/RdfQueryTestingBudapestMeeting from danbri
09:44:55 <danbri> I:|RDF Query Testing BOF, thursday 12.30 - 14:00
09:44:55 <dc_www2003> Titled item I.
09:45:16 <danbri> I:I've not checked out the room etc yet (Brahms) or found BOF signup sheets.
09:45:17 <dc_www2003> Added comment I1.
09:45:32 * dajobe looks at logging of /me's
09:46:06 * danbri thinks its a w3c-ism; ralph's logger doesn't log /me, so folk use that for [off]topic chat
09:46:17 <dajobe> it's my logger
09:46:24 <dajobe> here that is
09:46:44 <danbri>http://www.w3.org/2003/Talks/0521-sb-wwwintro/
09:46:45 <dc_www2003> J: http://www.w3.org/2003/Talks/0521-sb-wwwintro/ from danbri
09:47:28 <danbri> J:|W3C: Introduction, Process, Accomplishments, Steve Bratt
09:47:28 <dc_www2003> Titled item J.
09:47:32 <danbri> J:(just finished)
09:47:32 <dc_www2003> Added comment J1.
09:47:36 * sandro thinks RRSAgent should WARN about /me statements for the next 30 days, then start logging them.
09:48:12 <dajobe> it's an xslt issue here; seems not to be matching them properly. they are being logged & recorded
09:48:58 <dajobe> hmm, it is in the HTML
09:48:59 <dajobe> \
09:49:02 * dajobe stops lokking
09:54:01 <dajobe> i wish Ian would not use "elements"
09:57:25 <DanC> hmmm... "benefits are not that clear"... stating synonym relationships isn't a clear benefit?
09:58:37 <danbri>http://www.w3.org/2003/Talks/0521-BudapestW3CTrack-IH/
09:58:38 <dc_www2003> K: http://www.w3.org/2003/Talks/0521-BudapestW3CTrack-IH/ from danbri
09:58:45 <dajobe> I wouldn't know where to start with respect to Ian's talk
09:58:59 <danbri> K:|High Level Overview of W3C Technologies, Ivan Herman (W3C).
09:59:01 <dc_www2003> Titled item K.
09:59:27 <danbri> Ivan giving brief Web Service use case, travel agent app
10:00:20 <danbri> ...now talks about data-based automation. metadata. rdf. Ontologies as 'metadata vocabularies'. Mentions transitivity, symetricity of properties, hence Web Ontology language under development.
10:00:26 <danbri> ...search engines as a use case.
10:01:12 <danbri> ...example: SW app, W3C's RSS newsfeed
10:01:52 * danbri crashes IE on macosx by loading w3c's rss url :(
10:02:01 <danbri> ....same data picked up by news aggregators
10:02:52 <danbri> ...
10:03:04 <danbri> ...sometimes noise in press trying to present SW and WS as in conflict
10:03:29 <danbri> ..."A widely deployed Web Services infrastructure may become the most compelling business case for the Semantic Web"...
10:03:31 <danbri> ...but also
10:03:42 <danbri> ..."the synergy is important for both"
10:03:44 <danbri> eg
10:04:05 <danbri> ...searching for best equation solver
10:04:28 <danbri> ....ontology services, ("Semantic Web Services"), getting logical consequences of your questions
10:04:38 <danbri> ...difficult to implement, hence using a web service for this is of value
10:04:48 <danbri> ...both areas (WS/SW) are represented at w3c
10:05:20 <danbri> Ivan now talking about 'the web is for everybody', accessibility, ...
10:05:44 <danbri> ...Horizontal Activities at W3C
10:06:08 <danbri> ...accessibility, internationalization
10:06:15 <danbri> eg: International Text
10:06:23 <DanC> hmm... so to state that two properties have the same extension, we'd have to use universal quantification; we couldn't just state it as a simple fact. Hmm... maybe that is just a stylistic difference.
10:06:24 <danbri> ...not just character sets
10:06:38 <danbri> ...also selection of text in browser, right-to-left (eg. Hebrew)
10:07:28 <danbri> ...eg. of adapted presentations; eg. of address lists. same data on phone, ipaq, laptop, desktop
10:07:38 * danbri stops taking notes
10:11:13 <dajobe> I saw guha in the corridor outside, should have dragged him into horrocks' talk
10:11:37 <danbri> guha was looking for eric miller, btw.
10:12:51 <dajobe> he's in here
10:12:56 <dajobe> sw-1 track
10:23:42 <edd>http://www.oreillynet.com/pub/wlg/3213
10:23:43 <dc_www2003> L: http://www.oreillynet.com/pub/wlg/3213 from edd
10:23:51 <edd> L:|Edd's collected reportage from day one
10:23:52 <dc_www2003> Titled item L.
10:26:32 <edd> almost break time. i'm heading to the patent policy press conference.
10:26:53 <edd> in the Brahms room, for any other interested parties.
10:26:59 <dajobe> ok
10:27:03 <DanC> BenG coins "OWL feather", a subset of OWL lite corresponding to the DLP KR that he's analyzed
10:27:28 <dajobe> he talks of mapping the conclusions - ground facts - to DL/LP
10:27:42 <DanC> this DLP stuff sounds pretty cool.
10:28:36 <mattb> phew, lunchtime
10:30:52 <sandro> BenG used the "OWL Feather" term at DAML PI in Miami.
10:31:21 <sandro> "SweetOnto" is the demo DLP implementation
10:32:03 <sandro> yeah, I'd sure like to use this for working with OWL. Being able to go back and forth would help my brain.
10:32:57 <sandro> BenG says slides to be avail by next week.
10:33:22 <sandro>http://ebusiness.mit.edu/bgrosof
10:33:22 <dc_www2003> M: http://ebusiness.mit.edu/bgrosof from sandro
10:33:22 <dajobe> on his web site
10:33:46 <sandro> M:|Benjamin Grosof (location of slides to come)
10:33:47 <dc_www2003> Titled item M.
10:34:00 <sandro> M: slides for Descriptive Logic Programming to appear
10:34:01 <dc_www2003> Added comment M1.
10:35:09 <DanC> I wonder if there's any way to do the signature/trust stuff in this neighborhood.
10:37:07 <edd> patent policy press conf open to public, says janet daly, in case anyone's interested
10:37:14 <DanC> I wonder why the wavelan was so flakey during that session. it'f fine here/now
10:37:46 <edd> maybe contention?
10:38:58 <DanC> yeah, it did feel like an overloaded ethernet segment
10:39:40 <DanC> maybe they shouldn't assign more DHCP leases than they can service or something
10:40:10 * DanC installs wavemon, per clue from EricM
10:40:36 <edd> I'd rather have intermittent than get locked out.
10:40:48 <DanC> I suppose
10:41:22 * DanC tries to remember exactly what "dB" means
10:41:30 <DanC> I know it's short for "decibel"
10:42:25 <DanC> hmm... "power management: off"; edd, do you know if I can turn that on without risking my machine going kerflewey?
10:42:43 * DanC sees danny; thought he'd be in the press conference; is it over?
10:43:11 * DanC gets more interested in food than anything else...
10:44:44 <edd> press conf not started yet: 2 mins
10:44:48 <edd> power management, no idea
10:47:31 <edd> hardly anyone here. :-(
10:47:44 <edd> This means I'll have to figure out questions to ask.
11:11:24 <sandro> How do we schedule a BOF on RDF Social Meaning? It looks like RDF Query Testing is the *only* bof here.
11:12:59 <mattb> is there a bof signup and info area?
11:50:07 * maxf waves
11:52:25 <DanC>http://swordfish.rdfweb.org/discovery/2003/04/www2003/
11:52:25 <dc_www2003> N: http://swordfish.rdfweb.org/discovery/2003/04/www2003/ from DanC
11:52:43 <DanC> N:|toward WWW2003 conference proceedings in RDF calednar
11:52:43 <dc_www2003> Titled item N.
11:52:55 <DanC> N:libby's not proud of it, but I convinced her to share it
11:52:55 <dc_www2003> Added comment N1.
11:57:21 * sandro finds the BOF info on http://www.www2003.org/read.htm
11:59:06 <nmg> looks like a perfectly principled hack to me - nice work, libby!
12:03:46 <mattb> yikes, semweb panel room way full
12:04:07 <bijan> Yep
12:04:19 * maxf gave up on it, went to W3C track instead.
12:04:33 <maxf> But Pemberton declared it a joke-free zone :-(
12:05:13 <bijan> Wow, fiercely hostile initial talk
12:05:35 <edd> Well, directed to produce debate.
12:05:48 <edd> Lol at introducing jhendler as Al.
12:06:11 * danbri in "SciAm article: crap or not" panel sesh
12:06:16 * edd sat on the floor behind a screen, can't see who's whjo
12:06:16 <sandro> to counter the article, show how it might be perceived as harmful....
12:06:23 <edd> hopefully somebody who can see can scribe
12:06:35 <nmg> false expectations (as I suspect Nigel may say)
12:06:47 * dajobe in same room, can't see either
12:06:57 <bijan> Guha gets credit for RSS :)
12:07:20 <bijan> Call for audience particpation
12:07:27 <bijan> "STORM THE STAGE"
12:07:45 * edd gets shifted along further. It's like being cattle in here.
12:07:47 <mattb> saucy
12:07:55 * sandro doesnt get the "Al" joke
12:08:09 <nmg> "Berners Lee et al"
12:08:35 <nmg> (where presumably Ora is "et")
12:09:09 <bijan> et al
12:11:20 <bijan> Technically, it makes 50 or more service calls
12:11:42 <edd> I don't get what showing one particular focused implementation has to do with the general point
12:11:53 <edd> It's like the chess playing automaton
12:12:24 <danbri> It's interesting: how can folk who've implemented this stuff persuade skeptics that this is more than canned demo, smoke'n'mirrors etc...?
12:12:34 <bijan> Not 50 types of service
12:12:34 <bijan> Well.
12:12:34 <bijan> It's reasonably generic
12:12:40 <bijan> yes
12:12:45 <bijan> That's the challenge of the dev day
12:12:48 <bijan> session
12:13:12 * danbri wonders who amongst us make daily productive use of rdf/owl in our lives
12:13:24 * bijan gives congrats to evren
12:13:31 <edd> can someone paste the URI for jim's slide?
12:13:33 <edd> I can't read it from here
12:14:00 <danbri> try http://www.mindswap.org/2003/wwwpanel03/side1-0.html or similiar
12:14:13 <edd> BLURB:Semantic Web: SciAm article considered harmful - panel session
12:14:14 <dc_www2003> O: Semantic Web: SciAm article considered harmful - panel session from edd
12:14:22 <nmg>http://www.mindswap.org/wwwpanel03/slide1-0.html
12:14:23 <dc_www2003> P: http://www.mindswap.org/wwwpanel03/slide1-0.html from nmg
12:14:39 <edd> ah
12:14:42 <edd> i was going to put it under O
12:14:51 <bijan> He worked hard under pressure and the demo was reasonable, understandable, and smooth
12:14:51 <bijan> Whether it establishes the point is a different matter :)
12:14:51 <edd> O:[Jim Hendler's slide|http://www.mindswap.org/2003/wwwpanel03/slide1-0.html]
12:14:52 <dc_www2003> Added comment O1.
12:14:52 <danbri> ta
12:14:53 <dirkx_again> danbri: you can easily convince by showing that it is easy to implement; minimize company or country wide workflow alterations and is relatively robust. There is a lot you can do with jut 20% of the stack which can be instantly justified by just letting the old authoritative data sources do what they always did.
12:15:02 <nmg> thanks, edd
12:15:32 * danbri nods, agrees re gradualist deployment
12:15:49 <nmg> yes, but what about provisions for scalability?
12:15:55 <bijan> Not one way, but many to the semantic web
12:15:55 <bijan> The Buddist Web
12:15:55 <bijan> Or the Bahai web :)
12:16:03 <nmg> if you don't build for scalability, you won't have scalability
12:16:36 <edd> O:[participant list|http://www2003.org/t_panels.htm#PN-2]
12:16:36 <danbri> Jim, 'ecommerce did not create the web; the web created ecommerce'
12:16:36 <dc_www2003> Added comment O2.
12:16:41 <danbri> jim++
12:17:08 <libby> I totally agree re making the data open....
12:17:09 <libby> can anyone see the url?
12:17:27 <edd> libby, it's on the www2003 chump already
12:17:44 <edd> hi uche
12:18:00 <uche> hello edd
12:18:09 <danbri> danc?
12:18:12 <uche> hello all lucky folks in Bdapest
12:18:16 <edd> Amit Sheth up now
12:18:20 <uche> Budapest
12:18:23 <edd> University of Georgia
12:18:31 <libby> yeah!
12:18:37 <edd> and Semagix, Inc
12:18:42 <libby> hey uche!
12:18:46 <uche> hey libby
12:18:56 <danbri> uche, are you in budapest?
12:18:57 <uche> Hallo matt
12:19:01 <uche> I wish :-(
12:19:02 <danbri> ah, no
12:19:05 <DanC> aha.
12:19:05 <danbri> shame :(
12:19:07 * bijan would MUCH rather be a critic here.
12:19:07 <bijan> It's the far easier task
12:19:07 <bijan> Amit Sheth speaking
12:19:07 <bijan> Of Semag!x (Power * Through * Relevance)
12:19:08 <bijan> Amit is friendly to the article
12:19:10 <bijan> "Not harmful, but needs more context"
12:19:14 <dajobe> hi uche
12:19:17 <uche> hey ave
12:19:20 <uche> Dave
12:19:35 <dajobe> uche: you got a namecheck from tim re Sun
12:19:43 <uche> Uh oh
12:19:48 <dajobe> :)
12:19:48 <mattb> heya uche
12:19:50 * uche looks up witness rotection service :-)
12:20:11 <bijan> Big green cube slide
12:20:11 <bijan> It's about building ontologies
12:20:11 <bijan> Aha! It's a mapping of "agreement space"
12:20:11 <bijan> x axis == Scope of Agreement
12:20:12 <bijan> y axis: Degree of agreement
12:20:53 <dajobe> semagix?
12:21:15 <bijan> z axis: Subject of agreement
12:21:15 <bijan> Semantic web foucs is high y, all x, little z
12:21:15 <bijan> And the slide moved on before the scribe could figure it all out
12:21:15 <bijan> Next slide: Observations of practical ontology development
12:22:19 <bijan> Observed that targeted ontologies are practical to build and maintatin
12:22:19 <bijan> Next slide
12:22:19 <bijan> Use of these practical ontologies
12:22:19 <bijan> Classification
12:22:19 <bijan> Automatic semantic metadata
12:22:29 <maxf> W3C track: Bert Bos now up, on CSS3
12:23:24 <DanC> yo hugo
12:24:12 * hugoh waves, now that he managed to grab an IP address from the UDP-flooded network
12:24:36 <DanC> UDP-flooded?
12:24:44 <maxf> yeah, UDPflood-- :(
12:25:00 <hugoh> I keep seeing UDP packets and I had a 95% packet loss as a result
12:25:34 <sandro> automatic obituaries.
12:25:35 <sandro> wow.
12:25:57 <sandro> that's definitely the scariest SemWeb app I've heard of. (well, okay, second scariest.)
12:26:04 <DanC> I say: any information that isn't used rots.
12:26:10 <dirkx_again> Hugoh: the ICMP packet losses are not indicative of any overload by UDP - but more a result of the bridges loosing state to the wired network. This is a common problem - next rev of the firmware solves it.
12:26:28 <DanC> so if you can get lots of eyeballs on the data, it's much more likely to be maintained
12:28:06 <bijan> Er.. how much was lost?
12:28:20 <bijan> Same as first app, ontologies and rules for modleing domain and presenation
12:28:30 <bijan> Genearte multimedia presentation from this
12:28:30 <edd> bijan, practically all i think
12:28:38 <bijan> Oh damn.
12:28:46 <edd> last line was "Automatic semantic metadata"
12:29:07 <bijan> should I repaste?
12:29:24 <edd> Why not, from that line I mention.
12:30:50 <DanC> that seems like a near-sighed conclusion.
12:31:35 <nmg> live version of this demo at http://triplestore.aktors.org/~swh/xAktiveSpace-wd/
12:31:50 <nmg> works in mozilla
12:31:54 <danbri> dc_www2003:view
12:31:54 <dc_www2003> L: Edd's collected reportage from day one (http://www.oreillynet.com/pub/wlg/3213)
12:31:55 <dc_www2003> M: Benjamin Grosof (location of slides to come) (http://ebusiness.mit.edu/bgrosof)
12:31:56 <dc_www2003> N: toward WWW2003 conference proceedings in RDF calednar (http://swordfish.rdfweb.org/discovery/2003/04/www2003/)
12:31:57 <dc_www2003> O: Semantic Web: SciAm article considered harmful - panel session (blurb)
12:31:58 <dc_www2003> P: Repeat of Jim Hendler's slide (http://www.mindswap.org/wwwpanel03/slide1-0.html)
12:32:03 <danbri>http://tux.w3.org/~danbri/www2003/34KC0124.jpg
12:32:03 <dc_www2003> Q: http://tux.w3.org/~danbri/www2003/34KC0124.jpg from danbri
12:32:16 <reBijan> Info and app integration
12:32:17 <reBijan> Next slide: Example applicaitons
12:32:17 <reBijan> Lots of TIA stuff
12:32:17 <reBijan> Switches to demo
12:32:17 <reBijan> Whoops, Screenshot
12:32:17 <reBijan> Airport app
12:32:21 <danbri> Q:|Jane Hunter on SemWeb panel, just now.
12:32:21 <dc_www2003> Titled item Q.
12:32:25 <bijan> Oh I give up
12:32:44 * edd abandons notion of collecting panel stuff under 'O' :)
12:32:51 <dajobe> rdql he developed?
12:33:01 <danbri> sorry!
12:33:07 <nmg> it's the implementation that swh wrote
12:33:14 * danbri skimmed looking for Jane's entry
12:33:33 <edd> Danbn: it hardly matters. hardly. :-)
12:34:08 <mattb> he nearly said "dark matter of the internet", i'm sure
12:34:10 <mattb> shirky alert
12:34:16 <dajobe> lol
12:34:20 <danbri> he said 80% of it
12:34:20 <nmg> ;)
12:34:22 <edd> beer's on him. or mattb. opr
12:34:22 <dirkx_again> Ouch
12:34:24 <mattb> heh
12:34:26 <dirkx_again> That is very bad.
12:34:52 * danbri wonders who _does_ claim "one size fits all"... strawman opposition...
12:34:56 <sandro> "I'm not an ontological imperialist. I don't believe one size fits all"
12:35:21 * edd votes for a time-out soon, so we can change the air in the room and stretch my legs.
12:35:22 <bijan> danbri: cyc
12:35:33 <sandro> "subtley, as you move from task to talk, your ontology changes"
12:35:34 <mattb> yes, let's stretch edd's legs
12:35:36 <nmg> and how
12:35:58 * sandro tries to remember if edd is tall.
12:36:20 <danbri> fair point re cyc, though it is a bit modular and encouraging folks to use it to develop separate kbs...
12:36:54 <edd> It seems that most of the difficulties being talked about here are the "404 analogue". Trying to get to grips with messy imprecision.
12:37:03 <bijan> danbri: I often get students facing the problem of integrating multiple ontologies to "oh, why can't we all agree! Use Cyc!"
12:38:23 <bijan> Umm.
12:38:23 <dajobe> "more thing the web told us" <- talk about a simplification
12:38:30 <dajobe> s/more/one/
12:38:32 <sandro> "Dont kill the system by insisting on an absurdly obtuse bit of logical machinery"
12:38:33 <bijan> Jim was pretty right about what this panel should be called
12:38:46 <edd> bijan: ?
12:38:46 <sandro> "Scruffy Works. The web proves it works"
12:38:53 <bijan> WE had negative: moderator, pos: jim, pos: amit, neg: jane, superpos: nigel
12:39:02 <bijan> This is not a hostile to the article panel
12:39:10 * danbri nods
12:39:10 <nmg> who was that applauding the "web works because links are allowed to break" comment at the back of the room?
12:39:24 <danbri> R.V.Guha on stage now
12:39:44 <bijan> Giving the superlove treatment to the article!
12:39:56 <dajobe> guha: call to arms to build something
12:40:03 <bijan> I'll note, too, that Jane's critique was not particularly powerful
12:40:07 <DanC> 1959 McCarthy on AI
12:40:12 <dajobe> guha: article states the kind of bold vision that drives fields
12:42:22 <sandro> "70-80% of the pages on the web today are CGI out of some relational database. The right thing is to export the data!"
12:44:46 <maxf> Dean Jackson now up in main room. Cool SVG demos expected.
12:44:54 <sandro> "Keep things simple! Big stacks don't succeed!"
12:45:03 <libby_> oop
12:45:17 <danbri> O:[http://tux.w3.org/~danbri/www2003/34KC0126.jpg|R.V.Guha talking on SW adoption] in panel session.
12:45:17 <dc_www2003> Added comment O3.
12:45:20 * DanC is dismayed to be missing DeanJ on stage
12:46:04 <bijan> Oh dear, blogs
12:46:16 <danbri> speaking from floor: Colas Nahaboo, Ilog SA, France.
12:46:26 <danbri> (asked about weblogs)
12:46:46 <danbri> "the most important phenomenon in web today is appearance of weblogs/bloggers
12:46:59 <danbri> ...success is its simplicity, not needing to know much about links, structure
12:47:14 <danbri> ...when you look at recently successful systems, they're based on monkey-like immitation
12:47:34 <danbri> ...as long as sw relies on assumption on human knowledge of style/structure distinction
12:47:58 <danbri> jim: how many ppl wrote their first www page pre mosaic?
12:48:00 <danbri> (a few hands)
12:48:11 * danbri not trying to scribe, feel free fill in...
12:48:20 <libby_> I'm in deanj's talk....dean is wearing foafnaut tshirt :)
12:48:45 <sandro> Jim: the semweb will be easy and simple, when things settle down and tools emerge
12:48:54 <dajobe> jimH: what's simple to do now, wasnt's simple a while ago
12:49:15 <maxf> libby_, I'll be wearing mine friday ;)
12:49:28 <libby_> :)
12:49:29 <danbri> anyone seen eric miller? we keep missing each other...
12:49:40 <mattb> he was in here briefly earlier
12:49:42 <edd> he's behind you
12:50:15 <sandro> Guha: things need to keep simple UNTIL YOU KNOW IT'S GOING TO SUCCEED. *THEN* it's in the interest of the vendors to make things complex to keep high the barriers to entry. Netscape and Microsoft did this.
12:50:25 <bijan> "If things are too complex, they may be too hard to build and not successful"
12:50:42 <bijan> Bijan: "If thigs are too simple, they may be too hard to build and not successful"
12:50:49 <libby_> libby_ is now known as libby
12:50:59 <bijan> bijan: "If things are too hard to build and not successful, they may be successfully built"
12:51:02 <libby> oooh!
12:51:21 <maxf> wow!
12:51:22 <dajobe> libby: you tease!
12:51:31 <libby> wheee!
12:51:38 <danbri> libby, photos pls!
12:51:53 <danbri> took me 2 mins to get photo from camera -> www
12:51:59 <libby> dean demos text rotating round and round in a box
12:52:11 <mattb> is this svg 1.2 stuff?
12:52:19 <libby> yeah, think so
12:52:20 <mattb> antoine demoed some nice textbox-wrapping stuff at xmleuope
12:52:27 <maxf> mattb, yes
12:52:29 <libby> i missed that
12:52:29 <mattb> also video clipped to polygons
12:52:38 <mattb> all dynamically resizable of course
12:52:48 <danbri> semweb panel, floor speaker(who?)... how much data before enough to do useful merging?
12:52:56 <mattb> antoine claims svg 1.2 will be sufficient to implement an xhtml renderer :)
12:53:02 <sandro> "We spend 80% of our efforts scraping the data to RDF!"
12:53:09 <danbri> nigels: same in our experience
12:53:10 <libby> wow, this is so cool
12:53:11 <DanC> hmm... maybe we need a "what's new in the semantic web" ala the NCSA what's new page ;-). Hmm... maybe that's not a joke.
12:53:31 <DanC> what's so cool, libby?
12:53:42 <libby> showing multiple videos in same screen
12:53:42 <danbri> yup... i sorta do that myself by watching foaf crawler behaviour, but no automated output
12:53:49 <libby> rotating, transparency
12:54:06 <libby> stunning
12:54:09 <maxf> two videos playing at the same time, overlapped with transparency, and rotated.
12:54:19 <mattb> what client is rendering this?
12:54:21 <libby> smil in svg
12:54:22 <danbri> photos!
12:54:28 <libby> he has a special pre-alpha
12:54:32 <mattb> antoine had an acrobatreader plugin iirc
12:54:34 <DanC> pre-alpha of what?
12:54:37 <danbri> a full smil player? or the smil subset used for svg animiation?
12:54:39 <mattb> aha, super secret svg powers
12:54:40 <libby> not sure...
12:54:42 <maxf> and now dean moves on to explain how to be able to render XHTML ;)
12:54:56 <libby> heh
12:54:56 <maxf> (including the browser controls)
12:54:58 * mattb yawns, seen it all before then ;)
12:55:00 <edd> libby: if question time, petition for linux svg support :)
12:55:06 * mattb seconds
12:55:25 <edd> there's only one practical impl, adobe's
12:55:30 <edd> that's so not a good situation
12:55:33 <DanC> "how many people use meta tags?" I don't use those, but I use <title> and <address> all the time.
12:56:05 <danbri> edd, have you investgated svg on p800? i heard it possible...
12:56:15 <libby> I have!
12:56:21 <edd> danbri: i downloaded opera with it in, just not got a URI for a page with it on yet
12:56:32 <dajobe> emiller: blogs -t rackback, foaf, semantic weblet
12:56:34 <libby> the nice opera guy says you can email a company for their plugin
12:56:42 <libby> not opera...
12:56:44 <dajobe> ... interesting tying info together - lots of communities
12:56:54 <dajobe> emiller: not using ontologies (v large), content, network effect
12:56:55 <maxf> libby, zoomOn?
12:57:05 <libby> that sounds right
12:57:06 <dajobe> emiller: ... more content, sharing, exciting
12:57:06 <edd> libby: oh right, i misunderstood. too many hoops to jump through. the compant who makes the plugin had a Flash-full site that i couldn't navigate properly
12:57:21 <libby> I'm not sure it's even up there....
12:57:27 <dajobe> emiller: ... very pleased to see it, network effect close. Lots of talks here are bout bootstrapping.
12:57:34 <edd> emiller: ra ra excitement
12:57:46 <dajobe> emiller: Guha - what we've done is building a simple notion
12:58:24 <sandro> emiller: Guha, you *can* write this down on a page. (ie triples + URIs)
12:58:27 <dajobe> emiller: ... you said we might be doing this wrong. I'm interested to what are we doing wrong? Q2 Suggesting a service, challenge you to do that and make available
12:58:36 <DanC> what service is he talking about?
12:58:56 <libby> svg: zoomable UI - map say xforms to svg....
12:59:35 * libby considers career change
12:59:41 <maxf> :)
13:00:05 <sandro> Guha: we got to the RDF model by taking KR and chopping it down until we'd have to chop off a limb to go farther.
13:00:30 <sandro> Guha: XML, XLink, X*..... it's nuts.
13:00:44 <DanC> "blind watchmaker" by hawkins [did I get that right?]
13:00:45 <sandro> Guha: Read Richard Dawkins _Blind Watchmaker_
13:01:00 <DanC> .google blind watchmaker
13:01:07 <dajobe> not here
13:01:33 <libby> authoring tool for svg in svg!
13:01:37 <danbri> good book
13:01:43 <danbri> s/good/great/
13:01:44 <sandro> someone: SMS as an example of brute simplicity
13:01:52 <DanC>http://www.amazon.com/exec/obidos/tg/detail/-/0393315703/danconnollyA/
13:01:52 <dc_www2003> R: http://www.amazon.com/exec/obidos/tg/detail/-/0393315703/danconnollyA/ from DanC
13:02:03 <maxf> svg authoring tool is an 8k svg file
13:02:06 <DanC> R:|The Blind Watchmaker: Why the Evidence of Evolution Reveals a Universe Without Design
13:02:07 <dc_www2003> Titled item R.
13:02:19 <danbri> someone (from Canon; Jos Bullock)
13:02:21 <DanC> R:by Richard Dawkins
13:02:21 <dc_www2003> Added comment R1.
13:02:33 <DanC> R:buy it at amazon and DanC gets a kickback :)
13:02:33 <dc_www2003> Added comment R2.
13:02:42 <danbri> R:No design but _lots_ of testing... :)
13:02:43 <dc_www2003> Added comment R3.
13:03:18 <DanC> R:ironic... I've been reading a lot about intelligent design; how evidence in cosomology reveals a universe *with* design
13:03:19 <dc_www2003> Added comment R4.
13:03:40 <sandro> ?: Users cannot see the differenece between this stuff and their apps in Java.
13:04:07 <sandro> ...: As in Expert Systems which were 10 times better, and still not accepted.
13:04:24 <sandro> Nigel: SHaring of ontologies and data will help
13:04:38 <sandro> JimH: I like your example....
13:04:46 <danbri> R:Interesting... would be fun to follow this up in #rdfig sometime
13:04:46 <dc_www2003> Added comment R5.
13:05:16 <sandro> ...: Early browser demos couldn't touch hypertext systems of the day. BUT they were different. You do need the explanation
13:05:27 <sandro> ...: What breaks the AI analogy...
13:05:47 <sandro> ...: It's the web of semantics, it's not semantics on the web. Its scruffy.
13:06:22 <sandro> Amit: I hope this is a multidiciplnatry field, not just AI.
13:06:24 <maxf> dino prepares to show foaf, runs out of time :-/
13:06:30 <libby> noooo!
13:06:59 <sandro> ...: GOogle baught Applied Semantics, wihch uses text analysis to build ontology.
13:07:54 <sandro> Guha: We have a serious problem if the people watching the demo can's see the difference.
13:08:11 <maxf> dino does foafcorp
13:08:15 <libby> fatcats!
13:08:36 <sandro> Q(france): How much of your demos really realies on RDF?
13:08:55 <maxf> applause for Dino!
13:08:59 <libby> pepsi to coke in 2 people! [applause]
13:09:13 <maxf> Timbl throws him out of stage!
13:09:28 <maxf> Anti-corporate riot!
13:09:38 * mattb notes http://google.blogspace.com/archives/000927 re google and metadata/semweb
13:09:51 <edd> mattb: so chump and add somewhere?
13:09:59 <mattb>http://google.blogspace.com/archives/000927
13:10:00 <dc_www2003> S: http://google.blogspace.com/archives/000927 from mattb
13:10:03 <alberto> yes that was cool libby! :-)
13:10:05 <mattb> S:|Google myths
13:10:05 <libby> dino didnt have time to do foafnaut :(
13:10:05 <dc_www2003> Titled item S.
13:10:25 <mattb> S:*He thought meta data was fairly useless. Users don't enter it. He thought the semantic web project would fail. He said users around the globe will not work to enter descriptive data of any kind when there is no user benefit.*
13:10:25 <dc_www2003> Added comment S1.
13:11:03 <sandro> Carol: this panel... turf war.... If you can't do X, it's not the "Semantic Web"
13:11:15 <sandro> ...: If a little bit is a little bit useful, it's useful!
13:11:59 <danbri> O:[http://tux.w3.org/~danbri/www2003/34KD0127.jpg|depiction of semweb panel] during q'n'a.
13:11:59 <dc_www2003> Added comment O4.
13:12:11 <sandro> ...: I thought of this panel because there's *already* a backlash, which is odd. It's only been 2 years!
13:12:18 * danbri disagrees with the '2 years' thing. W3C doing RDF since mid 1997.
13:12:28 <sandro> Nigel: I disagree. THere's very little turf war, very little backlash.
13:12:36 <edd> I don't think the turf war idea is atypical at all.
13:12:40 <edd> Happens everywhere now.,
13:12:44 * dajobe saw MCF at 1995 web conf with Guha
13:12:46 <sandro> (france): Maybe because it's the old AI backlash
13:13:04 <sandro> Bijan: I worry it's going too fast, that it's making things go too fast.
13:13:30 <danbri> SW articles dates the rollout of the label/brand/bandwagon 'Semantic Web', not the technology-family that underpins it.
13:13:32 <sandro> ...: None of the promised democratization of the web
13:13:53 <sandro> ...: Fast capital flows are devestating to local communities
13:13:57 <DanC> "fast capital flows are devastating on local communities"
13:14:04 <libby> interesting
13:14:18 <sandro> ...: I worry about the policy implicatins, about how this will change human society.
13:14:32 <sandro> ...: I worry about TIA apps
13:14:50 <sandro> Amit: Every tech is good and bad
13:14:51 <libby> TIA?
13:15:00 <danbri> total information awareness (darpa / us mil)
13:15:05 <libby> oh aye
13:15:21 <DanC> how is TIA connected to DARPA?
13:15:27 <sandro> Amit: Tia was cut back -- it wont cover US citizens. [ Oh Yay. ]
13:15:43 <sandro> TIA is a DARPA program like DAML is a DARPA program.
13:15:45 <danbri> just us foreigners? ;)
13:15:45 <libby> oh just everyone else then?
13:16:27 <DanC>http://www.darpa.mil/iao/TIASystems.htm
13:16:27 <dc_www2003> T: http://www.darpa.mil/iao/TIASystems.htm from DanC
13:16:38 <sandro> And of course the the US Govt can contract with the UK to monitor its citizens. (I gather that's the usual eschelon practice.)
13:16:42 <DanC> T:|Terrorism Information Awareness (TIA) System
13:16:43 <dc_www2003> Titled item T.
13:17:02 <nmg> DanC: they've renamed it, have they?
13:17:16 * bijan whews
13:17:30 <DanC> this is way too metametametameta for me. This is navel-gazing about navel-gazing. I'd rather be in the SVG demo room where folks are building new interesting stuff.
13:17:35 <sandro> Guha: The "data web" is easy, an "intelligent agent" is harder.
13:18:01 <sandro> ...: The web has mininimal constraint on social behavior
13:18:24 <sandro> ...: Moms and pops putting up pages of their dogs and cats squashed MSN and AOL. Distributed Extensibility.
13:18:38 <bijan> DanC: no kidding
13:18:39 <libby> DanC++
13:18:46 <bijan> They also might have air
13:18:47 <libby> was the same last year
13:18:51 <bijan> WE've used ours up
13:19:00 <mattb> at least you have a chair :)
13:19:04 <sandro> DanC, it was advertised as a discussion about a discussion about the SemWeb.
13:19:06 <bijan> Yep
13:19:10 * mattb is sitting on the power cables
13:19:15 <libby> comfy!
13:19:20 * edd also concurs with danc
13:19:24 * DanC had reservations about coming in here; should have followed them
13:19:30 <libby> man, http://www.zoomon.com/start.htm site is awful
13:19:37 <edd> i had hoped for some revealingn insights to tell the world
13:19:43 <libby> I really want to try SVG on my P800
13:19:49 <edd> "SemWeb has big navel"
13:19:56 <libby> heheh
13:20:26 <sandro> DanC prefers first-order knowledge. :-)
13:21:46 <DanC> it seems to me that JimH does "I won't do X. [X]" where X="make a speech", "interrupt you", etc. *very* frequently.
13:21:48 <sandro> Guha: MS CTO says Web will fail for 6 (wrong) reasons -- no security, no caching, .... But those things were added later.
13:21:58 <sandro> s/says/said once/
13:22:19 <sandro> Guha: Killer app turns a bystander into a believer.
13:22:38 <sandro> ...: As in the <img> tag in mosaic
13:22:45 <edd> semweb needs an <img> tag :)
13:22:50 * edd ponders
13:23:00 <edd> guha definitely source of this afternoon's memorable quotes
13:23:04 <libby>http://www.w3.org/2003/01/monsterWIntro.svgz
13:23:05 <dc_www2003> U: http://www.w3.org/2003/01/monsterWIntro.svgz from libby
13:23:05 * DanC recalls the greatest insight I learned from high school... "all ambiguous references refer to sex." e.g. "doing it" etc.
13:23:26 * bijan wonders how many years the semweb will be delayed if we all expire of oxygen deprivation
13:23:36 * sandro turned guha's words into a sound bite there.
13:23:36 <edd> or accelerated...
13:23:39 <libby> U:|cute SVG 'monster' demo from zoomon, works on phones...if I can get the damned plugin
13:23:40 <dc_www2003> Titled item U.
13:23:47 <bijan> Given my comment, I'm not sure if I shouldn't endorse that
13:23:50 <nmg> small hot stuff rooms considered harmful
13:23:56 <bijan> Given how I feel, dying quick seems attractive
13:24:05 * DanC chuckles at nmg
13:24:13 * libby in deliciously cool interesting room
13:24:23 * bijan will hunt libby down
13:24:47 * edd sighs, decides to do something other than semweb next session
13:25:35 <libby> these things are a ploy to stop people getting too excited abotu the semweb ;)
13:25:52 * DanC wishes Roger were at the N3 tutorial TimBL/Sandro/I gave. N3 can be scribbled in 10 min.
13:25:59 <sandro> Q(cutler): you need to be able to teach someone how to do this in 20 minutes.
13:26:14 <sandro> danc:Roger=sandro:cutler
13:26:47 <sandro> ...: We never will have one format/ontology for invoices.
13:27:08 <edd> UBL will be a strong contender
13:27:14 <edd> it will never have 100% adoption...
13:27:21 <sandro> ...: Agreeing on business semantics; ... bringing down cost of entry into business relations.
13:27:28 <edd> but it doesn't mean that for controlled subsets it won't be done
13:27:36 <edd> sounds like speaker is ignorant of UBL work
13:27:36 <sandro> ...: please help us with this stuff
13:28:10 * DanC is getting tired of JimH putting words in timbl's mouth
13:29:21 <danbri> (sw panel closes)
13:32:42 <DanC> hoot! "For those of us who do not watch daytime television there is now a worthy alternative. Follow Dave Winer & Bill Kearny in their mudslinging and backstabbing adventures." -- http://weblogs.cs.cornell.edu/AllThingsDistributed/
13:34:59 <DanC> so what's anybody doing for the next session?
13:37:45 <dajobe> semweb paper track
13:39:33 * DanC sees he's to appear in the W3C track... http://www.w3.org/2003/03/w3c-track03.html
13:59:45 <sandro> wow, the degree of counter programming here is rather annoying.
14:03:03 <dajobe> semweb track
14:03:09 <dajobe> ibm guy
14:03:47 <dajobe> semweb bootstrap chicken & egg - apss, tags, adoptiona -> aps
14:03:49 <dajobe> apps
14:04:05 <dajobe> automatic large-scale bootstrapping
14:04:07 <dajobe> 1M pages
14:04:20 <dajobe> large enough scale toe ncourage app writers
14:04:31 <dajobe> then fostering user buy in, cycle of activity
14:04:40 <edd> is volume enough?
14:04:44 <dajobe> approach to generate the first tags across a braod corupus?
14:04:49 <dajobe> hmm?
14:05:00 <dajobe> ibm have a platform for data anaysis on the web
14:05:10 <dajobe> deeper than semantic tagging
14:05:35 <sandro> I think he said us:1 billion. ie 1e+09 pages.
14:06:00 <dajobe> "SemTag and Seeker: Bootstrapping the semantic web via automated semantic annotation"
14:06:37 <dajobe> data is being presented as an annotating server that is contacted by client browser
14:06:51 <dajobe> 3rd party annotatin service for web pages, dealing with RDF annotations
14:06:59 <dajobe> browser merges the content & annot into one page
14:07:08 <dajobe> ta da - they are using TAP
14:07:31 <dajobe> ok, 1B pages - ack
14:07:35 <dajobe> 250M pages with annotations
14:07:41 <dajobe> 50 types of annotations
14:08:02 <dajobe> (TAP nodes)
14:08:09 <dajobe> partial index keyed for nodes
14:08:33 <dajobe> perf, operates at 20K pages/sec working over 128 machines
14:08:36 <dajobe> ^spotting...
14:08:44 <shab> types of annotation?
14:08:51 <dajobe> learning at 8K/sec on 1 machine, disambig 50K windows/sec on 128
14:08:59 <dajobe> shab: he hasn't got to that yet
14:09:06 <dajobe> oshab: or it's in the pape
14:09:10 <dajobe> r
14:09:32 <dajobe> related work - web annotation tools
14:09:41 <dajobe> this is 3rd party, machine generated
14:09:43 <dajobe> annotations
14:10:16 <dajobe> taxonomy outline
14:10:16 <dajobe> types of ambiguity
14:10:43 <dajobe> either ambiguity in the taxonomy or out of it (to unknown terms)
14:10:54 <dajobe> the latter is the focus, there is more outside the taxo than in it
14:11:00 <dajobe> Alg: Automated Best Guess
14:12:18 <dajobe> concenterating on precision rather than recal
14:12:18 <dajobe> l
14:14:31 <dajobe> 1M pages giving 200K words, then 1100 judgements
14:16:52 * dajobe runs out of typing energy
14:17:23 <dajobe> "WebFountain end-to-end architecture overview" slide
14:21:21 <dajobe> server returns annotea applications
14:21:42 <dajobe>http://tap.stanford.edu/semtag/
14:21:43 <dc_www2003> V: http://tap.stanford.edu/semtag/ from dajobe
14:21:49 <dajobe> V:|Semtag site (to appear)
14:21:50 <dc_www2003> Titled item V.
14:21:59 <dajobe> V:404 now *while the data rsyncs*
14:21:59 <dc_www2003> Added comment V1.
14:22:38 <dajobe> V:=http://tap.stanford.edu/semtag/index.html
14:22:38 <dc_www2003> Replaced URL of V.
14:22:41 <dajobe> V1:@@
14:22:41 <dc_www2003> Replaced comment V1.
14:22:44 <dajobe> V1:""
14:22:44 <dc_www2003> Deleted comment V1.
14:22:55 <dajobe> V:|SemTag site
14:22:55 <dc_www2003> Titled item V.
14:26:50 * mattb wonders if the papers CD is available anywhere, via web or someone's laptop webserver on the lan?
14:26:57 <mattb> not got my drive with me
14:28:26 <shab> I have it on my machine here
14:29:02 * dajobe leaves sw-2
14:37:41 <sandro> Oooh, HiWe, a form crawler that runs through combinations of inpouts to forms. Ouch.
14:38:00 <sandro> (to get at the "hidden web" only seen through forms.)
14:41:08 <amy> amy is now known as amyv
14:43:21 <maxf> hi amy
14:43:28 <amyv> hey Max :)
14:49:59 <libby>http://rdfweb.org/people/damian/foaffinger/
14:50:00 <dc_www2003> W: http://rdfweb.org/people/damian/foaffinger/ from libby
14:50:21 <maxf> aha! worky this time libby?
14:50:35 <libby> W:|FoafFinger, by shellac (Damian Steer): a little application using rendezvous and rdf. Sort of a unix 'finger' replacement.
14:50:37 <dc_www2003> Titled item W.
14:50:38 <libby> hope so!
14:51:57 <maxf> libby, the jar file is served as text/plain
14:52:04 <libby> I know :(
14:52:27 <libby> I told him...
14:53:06 * alexbrn slaps eamonn_ around a bit with a large trout
14:55:06 * maxf runs foaffind, sees no one
14:55:46 <DanC> anybody know if there's any stuff on the WWW2003 proceedings CD that isn't on the www2003.org web site?
14:56:11 * mattb runs foaffinger, sees libby
14:56:35 <libby> ooh
14:56:50 * libby sees 3 mattbs, all offline!
14:57:01 <maxf> aha!
14:57:08 * maxf sees Libby, Matt, Jimbo
14:57:22 * mattb sees maxf too
14:57:25 <mattb> this is neat
14:57:26 * libby too :)
14:57:29 <mattb> i must look into jrendezvous
14:57:33 <mattb> good job damian
14:57:42 <libby> shellac had some problems with it, but it's cool alright
14:59:33 <pshab> DanC: I don't think the actual papers on the website?
15:00:18 <libby> show doesn;t seem to work quite right
15:02:42 <maxf> how do I see my data?
15:03:04 * DanC would like a pointer to the slides PaulC is presenting
15:05:33 <libby> maxf, dump shows it in rdf
15:05:46 <DanC>http://www2003.org/t_www.htm
15:06:52 <maxf> ah, not real name. I'll restart.
15:09:22 <dc_www2003> X: http://www2003.org/t_www.htm from DanC
15:09:37 <DanC> X:|W3C track at WWW2003
15:09:49 <dc_www2003> Titled item X.
15:11:45 <DanC>http://www.w3.org/2003/Talks/www2003-0521-TAG/all.html
15:11:49 <dc_www2003> Y: http://www.w3.org/2003/Talks/www2003-0521-TAG/all.html from DanC
15:12:04 <DanC> Y:|Technical Architecture Group Update
15:12:06 <dc_www2003> Titled item Y.
15:12:13 <DanC> Y:presented by Paul Cotton
15:12:21 <dc_www2003> Added comment Y1.
15:33:06 <amy> amy is now known as amyv
15:53:23 <maxf> maxf is now known as maxoff
16:41:04 <DanC> hmmm... can anybody here put geourl markup in http://www2003.xmlhack.com/ ?
16:44:05 <DanC>http://geourl.org/near/?lat=47.4&lon=19.240.&dist=40
16:45:06 <DanC> or http://geourl.org/near/?lat=47.51&lon=19.08&dist=40
16:46:45 <dc_www2003> Z: http://geourl.org/near/?lat=47.4&lon=19.240.&dist=40 from DanC
16:56:13 <uche> The chump has been a bit slow today
16:56:18 <uche> Heavy use, I wager :-)
16:56:36 <uche> But I don't think t's dropped anything entirely
18:03:53 <uche> uche is now known as uche|away
18:31:16 <edd> hmm, does being in the basement bar count as part of the conference?
18:32:16 <JibberJim> That's the most important place in the conference surely?
18:33:12 <edd> Well, it has wireless :)
18:36:39 <edd> hey DanC :)
18:41:33 <edd> DanC: so, i'm trying to sign your key. the difficulty is, i can't find it on the keyservers. can you let me know where it is?
18:48:53 * edd looks on DanC's home page, still can't find pubkey
19:23:23 * MarkB spots http://www.w3.org/People/Connolly/connolly-pgp-key.html
19:25:20 <edd> MarkB: not the same key he gave me a fingerprint for
19:28:33 <edd> DanC: i couldn't find your pubkey anywhere
19:37:10 <plh-hu> I found it at gpg.mit.edu
19:37:40 <plh-hu> s/gpg/pgp/
19:39:14 <edd> plh-hu: which keyid?
19:39:18 <edd> i'd tried that server
19:39:33 <edd> i'm looking for the key with ID 6e52c29e
19:41:42 <plh-hu>http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x6E52C29E
19:41:42 <dc_www2003> AA: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x6E52C29E from plh-hu
19:41:50 <plh-hu> oops
19:41:57 <plh-hu> not sure if this was intended
19:43:21 <edd> AA:|DanC's public key
19:43:22 <dc_www2003> Titled item AA.
19:43:24 <edd> may as well use it :)
19:43:57 <edd> heh, it works now for me too.
19:44:02 <edd> darn heisenbugs.
19:44:04 <edd> thanks, plh-hu
19:44:32 <plh-hu> wlc
19:45:11 * edd signs and re-uploads DanC's key
20:01:46 <edd> grr, how I hate RSS crawlers.
20:02:21 <bitsko> :)
20:03:50 * edd declares all lost, goes to cry into his beer
22:27:44 * eikeon wonders what timezone the conference is in.
22:28:36 <JibberJim> UTC IIRC
22:28:44 <JibberJim> Europe
22:28:44 <eikeon> Thank you.
22:28:51 <eikeon> Close enough ;)
22:28:51 <dajobe> utc+2
22:28:54 <dajobe> hi eikeon
22:28:57 <dajobe> and jim :)
22:29:04 * JibberJim went the wrong way...
22:29:57 <eikeon> Hi dajobe
22:32:43 <dajobe> wireless_in_the_bar++
22:39:44 <eikeon> Very nice ;)
22:44:11 * dajobe sitting with libby and mattb &c
22:46:21 * eikeon failed to find the means by which to make it to dev day ;( -- whould have been nice to pull up a spot next to y'all.
22:46:51 <dajobe> we re-watch dean's cool svg foafcorp demo
22:47:10 <dajobe> eikeon: :(
