last updated at 2004-05-18 22:31 UTC.

W3C Atom Meeting Minutes

Developers' Day info

jhendler: schedule in the conference program is wrong - be sure to check here instead

Enabling Open Source Blogspheres

interesting faceted browser from museum domain

Minutes from Atom discussion, New York City, 18 May 2004

Interaction Design and the Semantic Web - XForms and the Semantic Web by Steven Pemberton

DanC: wish I were there
pshab: demoing xforms examples - follow links in presentation
pshab: FOAF and Google

Blogs as "Protected Space", Michelle Gumbrecht

karlcow: 23 bloggers only
karlcow: karlcow: I have the feeling that the conclusion are a bit too cultural centric

Mapping the Blogospere in America, Alex Halavais

karlcow: and Jian Lin
karlcow: What does it mean a blog lives in a city?
karlcow: Does it matter? yes. Cities still affect people, identities, social network.
karlcow: karlcow, hmmm... he's showing data for US, not for America...
karlcow: He has decided to limit his data to US, for a lack of resources, but once the process will be formalized, it might be easy to extend to other countries and languages.

Workshop on the Weblogging Ecosystem: Aggregation, Analysis and Dynamics

bkdelong: Personal Knowledge Publishing Suite with Weblog, Ikki Ohmukai & Hideaki Takeda
karlcow: Ikki Ohmukai is making a demo of Glucose
karlcow: Glucose gives the possibility if you give keywords to query RSS search engines for you

WWW2004 Workshop Beyond the Click: Interaction Design and the Semantic Web - Part II

pshab: Special Afternoon Session
pshab: Balancing Human Control and Automation: A Tale of Two Perspectives in a Case Study and Thought Experiment of Web-based Photo Annotation.
pshab: Lead by Web Ontologist, James Hendler and HCI Guru, Ben Shneiderman
pshab: focus on annotating and then repurposing photos in a Semantic Web context
pshab: jhendler - presenting... rdf/rdfs/owl built on existing web infrastrurcture... turn that to photos...
pshab: images diff to text - text tends to feel like it can be processed to extract meaning --> a discussion for another day
pshab: but images are different - not pre-indexed, ad hoc, etc...
pshab: demo of mindswap photos project
pshab: use it browse info on people - links to web based data
pshab: tool is powered by the ontology
pshab: question - how do we know what will happen when we click a link
pshab: same as the Web... info about this person, etc... generates a page about the thing
pshab: if lots of info e.g. 5000 bits of info/concepts?
pshab: you need a different way of displaying it - its a different issue [to that he is talking about]
pshab: presentation - e.g. can have different presentation scripts based on how many links...
pshab: shows example from demo - human readable lables v URIs can use either
pshab: if use a differetn ontology get different view...
pshab: can add/use ontology - extends foaf with conf. concepts. can use any that can be parsed by Jena
pshab: demo of system, anyone know the link?
pshab: drag and drop interface, color coded... interface, drill down the ontology. Filling in values in form boxes
pshab: can create region - and annotate the region, drag the concept 'keynote speaker' into the annotation box and add metadata...
pshab: gives another example from mindswap of papers - bibtex data added using same system
pshab: could do similar (much better) in lots of areas - but usign the same infrastrucre that takes different ontologies
pshab: Ben Shneiderman asks how new data is added - e.g. adding new data to bibtex data? can addt o the form easily. But of course not populate new data for existing records...
pshab: Ben Shneiderman presenation
pshab: so why are we doing this [annotation] stuff?
pshab: showind demo - 26000 image, photo library, last years www conf. - makes slide show...
pshab: annotations added to query via drag and drop interface
pshab: shows query of images of him and Jim... shows import of images from today... and adds metadata e.g. people in images... can do bulk annotation...
pshab: Jim asks were the additional info of e.g. that they were eating at Lindy's ... well... discussion about [potential] adding ways of adding new data e.g. dates, diaries,...
pshab: also has free text search... new tools will data from GPS, image analysis e.g. colours...
pshab: What do we want to support? Jim says' everying, I don't thnk you can... [Ben] is taking... I'm not the side of control, what we know, ...
pshab: shows lots and lots of examples of all the kind of info that could be given, ... goes to very extream, e.g. temp difference when they were speaking in presentation...
pshab: Ben thinks we need to go incremental, adn then build and grow... e.g. papers, sessions, people. good place to start, but needed soem time on interface even for that...
pshab: modest goals and incrementally grow... feedback from user... user interface drives it wiht the ontologies etc. beign importnat but secondary
pshab: motivated by questions/peruposes we have...
pshab: shows exampel fo photo system - with zooming interface - designed for rapid exploration...
golbeck: The software Ben is demoing is called photofinder, and available at
pshab: Wendy Hall - asks about ethics of putting photos up wiht annotations...
golbeck: The sofware jim is showing is at the mindswap website
pshab: do get requests to be removed or aspects of them removed...
golbeck: Jim's demoed markup tools are available in the downloads section of
pshab: perhaps need to say this will be captured in the conference registration documentation?
pshab: interesting issue. parked.
pshab: note from audience. legal infrastructure is not keeping up - not just photos
pshab: Jim - Common aspects? drag and drop. yes. incremental? both agree grow functionality of interface...
pshab: both interested in linking, sorting, organisation based on links...
pshab: most difference - traditional interface adn Web - interface try to accoutn for the needs of as many users as possible. Web more open e.g. better e-books than the web but the Web is more open...
pshab: different communities need different customisation - sorting, viwing... too many needed to do it by hand so...
pshab: from audience - the properties are of people not of images
pshab: shouldnt' been in the interface about images(?) did I get that right?
pshab: Jim wants all the people info about the people in the image...
pshab: Jim gives example of chaning needs in indexing, e.g. have images indexed by age - now want where it was taken...
pshab: thougth - Jim's problem here is similar to my problem when doing background work on ARKive see dlib article
pshab: Ben - if you say you can ask any question...
pshab: ... what is a photo feature? esp. about the content of the photo e.g. number of people? what concept space?
pshab: ... there are no facts on the Sem Web - only assertions
pshab: general agreement(?)
pshab: Jim, would like to be able to take data from one applicaion area to others. e.g. put it in an RDF database to be re-used
pshab: as more sem web data ... have x hundred thousands images... need new ways of displaying that...
pshab: ... if conference badges had embeded visula data, link to other data when a photo is taken
pshab: could join data automatically...
pshab: Jim ical data + photo data from camera another example of potential automatic/semi-automatic annotation
pshab: Ben gives example of defn. of 'year' in government stats - footnotes as long as the doc... can't just join data on 'year'. Doign it automatically gets you in to trouble...
pshab: break for coffee.
pshab: Ben notes that photofinder is free to download

teefal: PacManhattan ... play pacman as real people in greenwich village NYC

bkdelong: | - a mobile social network for people in the NYC area.

WWW2004 Workshop Beyond the Click: Interaction Design and the Semantic Web

pshab: lots of interesting discussion about how to go from SemWeb data to display esp. questions about query
pshab: talking about the people/community side of Sem Web being important - not just "data data data..."
pshab: talking about FOAF in that context - foaf good basis but need to have a mechanism for finding extensions where necessary?
pshab: good discussion about how to find stuff about/related to/etc. a sem web concept e.g. 'person'
pshab: ... markup of doc as authoring e.g. sematnic marker for tony blair for prime minister kinda thing
pshab: different from an end user - uses
pshab: Jim H. pointed to a project
pshab: ... problem of concept e.g. coffee defined in 10 different places with 10 different perspectives...
pshab: Is a problem with english conflating the 10 differetn concepts of 'coffee' e.g. bevergae/crop
pshab: Jim H. saying could/need have a way to make it clearer what particualr concept a query is related to...
pshab: ... difficult to talk about this stuff as the Semantic Web content is not yet very extensive...
pshab: Ben Shneiderman - asking. is there a project that takes a text qury and returns concepts? (I think?)
pshab: Ben and Jim are arguing about if there should be a smaller number of ontologies - so people learn them (or not) muhc like people learn classification schems like dewy
pshab: ... have now gone to FOAF projects to show example of browsers... foafnaut
pshab: going very slowly...
pshab: talking about how it (foafnaut) could be improved...
pshab: foafexplorer is down?
pshab: Session breaks for coffee.
pshab: session restarts
pshab: Haystack - MIT project
pshab: allow end users to add semanticly meaningful markup - based on idea of identifiying types of data
pshab: doing demo... shows a context menu generated from the data
pshab: ... creates display based on 'wrapper induction' - e.g. identifying patterns via feedback from user(?) and/or DOM of the page
pshab: can also map properties of things identified to properties in external ontologies (?)
pshab: ... this is done in practice via highlighting a node and saying this is a dc:title...
pshab: jen talking about IRC bots... kinds of things they do
pshab: talking about FOAFbot - not currently running?
dajobe: no, it's not running on #foaf at the moment
pshab: the first interesting way to get to FOAF data interactively
pshab: created Trustbot... not sure what it did - missed that bit
pshab: going to give examples - currency/weather/translation
pshab: semantic web adds... ?
pshab: showing wsbot. as a demo of three services. based on exising web services
pshab: notes that if the services go down so does the bot service - question, could add services if one goes down? yes but not done here
jhendler: trustbot extended foaf to include a trust value, and allowed a computed trust value between two users
pshab: have help for each function to show how to use it
edd: (FOAFbot is currently resting)
dajobe: "restoring the DOS command line to us all"
pshab: question/note - just giving us back the dos command line?
pshab: yes but... part of conversations...
pshab: question - why apple/MS not tried to do this at the OS level?
pshab: but why not taken it from the chat communities and added it to the OS level?
pshab: They have - been doing it for 30 years but hasn't taken...
pshab: need an economic model - BBC doing [something like this] this internally? or did I missunderstand that comment?
pshab: have a canned list of queries ... how going to know/find how to use them
pshab: ... about communities (members) evolvign/learning it together
pshab: ... example insult bot - sends you a random insult
pshab: question - semantic web relation?
pshab: Sem Web Backed - FOAFbot/Sem Web Web Serivces...
pshab: using web services makes it a bit easier than HTML scraping...
pshab: e.g. 6 lines of code using APIs that already exist
pshab: could be used to add webserivces (on the fly) by having a standrd syntax for it...
pshab: presentation ends
pshab: is this like the MS Office paper clip? - no, less intrusive, not polluting what's goign on the channel
pshab: there if you want it... is it more than a comfort factor for users?
libby__: re bot web service thing - I sorta did this for lightweight query routing, using very simple lists of predicates based on whois++ ideas. but won't scale as it is if there are many data sources with similar predicates available
pshab: a way for humans to interact with web srvices without writing own code...
pshab: there are examples from xforms?
pshab: more complex inputs? e.g, more complex say direction data, need two addresses.
pshab: is a problem - but does do lots, even with farily complex sercies so long as they take only a small number of inputs
pshab: there are examples of using IRC to fill in xforms...
pshab: is that useful?
pshab: compare to phone bot interfaces...
pshab: beyond basic e.g. weather --> weather tomorrow needs a more complex environment...
pshab: discussion about visualisation/interaction possibilities ... with or without complex conditions
pshab: need middle ground between human/machine driving
pshab: e.g. lists + visualisation more potent than either alone
pshab: computer is a visual --- brain dead idea that machines interact well based on human/human interactions
pshab: last comment by Ben Shneiderman
pshab: ... discussion about how to do command line interactions...
pshab: ... example of want for google command line (launch at least)...
pshab: mozilla can do some of that... apple has 'launchbar'(?)
pshab: session move on to poster session
pshab: discussion based on quesion about when 'documents' become so richly embeded with semantics and conditionality in display - is there a difference between a document and interfaces
pshab: Nigel Shadbolt. poster... defnece funded project but on activities not concerned with war (or similar phrase) e.g. relief efforts coordination
pshab: integration of information in the field
pshab: ... sets of view points
pshab: currently info is feed in but not integrated
pshab: ... need to present and swap info sources in and out quickly and meaningfully
pshab: need to understand the needs of different user groups etc...
pshab: seems to be based on AKTIVE space framework?
pshab: question - trust and provanence issues..?
pshab: interesting to see what is done now e.g. tacit processes etc.
pshab: not just about trust, but also about how its passed up/along the chain to other stakeholders with different goals
pshab: Next presentation
golbeck: Greg Elin
pshab: picturee annotation - very active area, inc. FOAF related stuff, codepiction...
pshab: demo
pshab: shows example of adding metadata into a form... and linking to rectangles over image
pshab: have defined new rdf about image regions. e.g. imageregiondepicts cna't see the syntax
pshab: WWW pics being used to demo e.g. people
libby: the rdf imageregions schema should be somewhere around here; we have a greement for it to go on the w3 site at
pshab: had to deal with issue of what is the minimum elve of metadatata - originally zero but then realised need creative commons data... to allow reuse
golbeck: conference ontology is
pshab: blog model is very powerful - posting image + annotations under the image
pshab: pointing to moblog site on
pshab: annotated images - some with 25(ish) comments
pshab: Example from Aktivespace photo demo taking photonotes and integrating it with activespace
pshab: this looks like a cool example of conference data integration
pshab: session ends for lunch

TOPIA Demo - interface over semantic web info

HERA project Web site

WWW2004 Workshop on Content Labeling

mattb: Technical and Socio-Cultural Challenges and Solutions
mattb: papers
libby: agenda
mattb: in Japan, the mobile phone is currently the primary communication device for teenagers and 20somethings
mattb: so parental supervision is not an option for safe surfing
mattb: proposed new solutions have RDF as a basis
mattb: workshop includes representation from (amongst others) Internet Content Rating Association, BBC, Vodafone
mattb: ICRA label tester
loggerw: See discussion
mattb: arrests made for cases related to "dating sites" in japan (february 2003 figures, PDF)
mattb: more than 1000 per year
mattb: speaking now: Akio Kokubo
mattb: talking about the SFS filtering browser demo system
mattb: now talking: Daniel Appelquist from Vodafone, talking with a focus on European operations
mattb: in the context of Vodafone Live - not "Internet on your phone" or "WAP"
mattb: looks like a walled-garden with the platform provider aggregating content providers and providing micropayment services
mattb: platform provides trust relationship with the consumer
mattb: presence in more than one country means there's a patchwork of regulatory issues and cultural norms
mattb: eg some Dutch content would be considered unsuitable in other countries
mattb: content stored in Vodafone Content Markup Language, which supports addition of RDF metadata (eg DC, PRISM), transcoded into appropriate format for delivery to phones
mattb: they refer to the internet as a whole as 'Off-Net'
mattb: the vodafone live world is 'On-Net'
mattb: one possible model for parental filtering (assuming content is labeled): child attempts to view possibly-forbidden content, parent gets a yes/no SMS
mattb: note: this is not implemented, and would be opt-in by the child too
mattb: mobile users have different expectations from PC users: the phone is an appliance, and so is not expected to crash or be random in any way
mattb: break user trust just once and they'll give up on you
mattb: deai-kei sites very popular in japan
mattb: Empirical Analysis of Google SafeSearch
mattb: general architectural point about filtering based on content labeling: where to do it?
mattb: three basic options: on the device (phone, etc), in a proxy as part of the platform, or at the provider
mattb: general production point: getting content labeling 'right' is very difficult from an onto/vocab point of view
mattb: especially considering different viewpoints, cultures, etc
mattb: cf music genre labeling
mattb: seems to be a general agreement that RDF is the best way forward
mattb: PICS's fixed ontologies with numerical values of 'badness' didn't work
mattb: point from Phil Archer of ICRA: "commercial filtering companies have enormous proprietary databases but this is their commercial advantage and they won't translate them into shared formats"
mattb: "also, there is a liability issue if someone else is using their data with code they didn't write"
mattb: The RDF Trust Model using RDF bookmark and its Application
mattb: Nobuhisa Shiraishi of NEC about to speak
mattb: google query using RDF bookmark demo
mattb: Eric Prud'hommeaux speaking
mattb: slides (CSS file appears to be password-protected, hence prompt - content is open)
mattb: first talk to actually give examples of RDF and OWL technology applied to the rating and selection/filtering process
mattb: using RDF query for this sort of selection is tricky because of closed vs open world distinctions
mattb: Rachel Yould now speaking on socio-cultural issues
libby: swig mission

restaruants in the NY area

WWW2004 community coverage

What's this?

This web site provides a hub for attendee-generated notes and coverage of the World Wide Web Conference 2004.

The conference is now over. Thanks to all who took part in creating the online coverage, which will remain archived on this site.



Official WWW2004 site
Developer's Day site
WWW2004 on the ESW wiki
W3 Photo and WWW2004 photos
Semantic Web Interest Group scratchpad
Planet RDF

Maintained by Edd Dumbill and Dave Beckett. Chump hosted by XMLhack. Logs hosted by Institute for Learning and Research Technology, University of Bristol. IRC was provided on Freenode, an interactive chat service of the Peer-Directed Projects Center.

Created by the Daily Chump bot.