Semantic Web Interest Group IRC Scratchpad

Welcome to the Semantic Web Interest Group scratchpad generated automatically from discussions on IRC at Freenode channel #swig 2001-2018 approx by the chump bot.

Nearby: IRC logs | semantic-web list | W3C Wiki (Recent changes) | delicious swigbot

last updated at 2004-07-15 16:20
crschmidt: didn't mean to chump the chump, heh ;)
 
em: by Stefano Mazzocch
 
chaalsCPH: Bengt demonstrates it. The demo shhows how 54 concepts are encoded, with representations in english and 3 different symbol sets
chaalsCPH: The CCF Pad allows you to create a message from the available terms, and change the representations.
 
 
chaalsCPH: A paper by Joseph Reagle, on synax and semantics, and different places that interoperability can break
 
Hosein and Liddy - looking for meaning across languages.
chaalsCPH: What is important for finding the meaning of a new word (e.g. in a new language) is its context
chaalsCPH: We want to have some more relations to describe abstract structures
chaalsCPH: So we can use these to work out ways of putting new words into contexts that map to ours. How?
 
Bengt and Matt - making translations
chaalsCPH: Problem is that many concept databases don't contain everything you need. Similar problems with cultural context.
chaalsCPH: How does a person learn when doing translation? How do we track what they have learned? People learn stuff that can be a local dictionary, and that maps to a larger dictionary...
chaalsCPH: If you have a dictionary that you are using, and you find a term that is outside that dictionary, can you find an explanation in terms of the local dictionary? Or can you decide that the user has acquired a new term that can be added to their dictionary.
chaalsCPH: You may have a child who has never encoutered a term before, but will learn it when they meet it, or a person who has to have the same term explained every time they encounter it.
 
Dom and Helle - glossaries for translation
chaalsCPH: When you describe a concept, and give a definition, you need to adapt the definition to the audience (rough idea for beginner, detailed for a professional...)
chaalsCPH: If you have a set of concepts and describe them in english you expect someone to map those to other languages. But in some cases they don't map neatly because of the structure and cultural context. So should there be a single authorised translation, should you authorise several translations, ...?
chaalsCPH: Can we use these glossaries to help automating translation? Hand our glossary to an automatic translator?
chaalsCPH: Can we look for a translation of a document?
chaalsCPH: r12a - if you want to translate "OK" on a dialogue box in a system, there are contextual differences. In spanish, windows uses "si", Mac uses something else, Unix something else again.
 
Catalogues and libraries
chaalsCPH: When librarians make Marc records they record lots of stuff they don't show in the public catalog. A lot of what they do show is nearly the same as Dublin Core
chaalsCPH: Likewise, the most commonly used little bit of ISO19115 is pretty much what is in Dublin Core.
chaalsCPH: and so on for other metadata standards, where the information "for the public" is a small part of what the professional users want, and the small part is basically Dublin Core.
chaalsCPH: You cant' translate everything in the "professional's part" between Marc and ISO - which is OK because they are not the same user groups.
chaalsCPH: I don't think I can see a user requirement for being able to map the complex stuff across.
chaalsCPH: But there seems to be agreement on the "element names" for the common (dublin core) bit, but there isn't so much agreement on the semantics, and so there isn't always interoperability. Language differences are one of the barriers across which the interoperability can be limited.
mortenf: chaals: sometimes it's necessary to make mappings at the lowest level - where possible - could be critical
 
What Eva and Leif are looking for
chaalsCPH: A person knows some stuff about themselves, a particular piece of information has metadata.
chaalsCPH: How can we put these things together, so the person finds what they are looking for?
chaalsCPH: The problem is that you cannot rely on the people creating the site to provide the kind of metadata that you need.
 
combining some stuff from chaals, Morten, R12A (Richard Ishida)
chaalsCPH: Two main things we wanted to do - describing people
chaalsCPH: and reacting to those descriptions. The descriptions had a number of aspects - things that were permanent, some things changed a lot. This is a continuum.
chaalsCPH: There are descriptions that are contextual. For example chaals calls his Mother "Mum" but other people call her Liddy.
chaalsCPH: And things change according to language and location - Spanish names and japanese names are structurally different. We want to be able to map them into systems that are not aware of the difference, but to be able to keep the different structures for doing things correctly.
chaalsCPH: We also want to be able to decide how we share the information - what we give to whom (will I tell Richard's website about me? Will I share my phone number with my Mum?)
chaalsCPH: And we don't have one piece of information that can be used to identify each entry - there is no unique.
chaalsCPH: ... key.
chaalsCPH: We want to pass the information around, labelled according to what is useful for the person we are passing it to (ideally, we know because we store that information and use it in various local applications)
 
 
libby: David's Picture
 
I would like my team page to automatically present information in the time zone of the viewer. To enable that, I would like a standard way for people to describe their preferred languages, location, time zone, form of address, preferred format (html/xml),
r12a-CPH: current platforms, etc. so that information can be presented to them in an appropriate fashion.
 
I would llike to have somothing like a multilingual knowledge base to help people on translations... multilingual dictionaries aren't enough in a global e-Society
 
Hosein's use case
chaalsCPH: Don't want to collect different cross-language metadata in the same place - difficult to maintain, people get confused
chaalsCPH: Make a mechanism that can map information from one domain to another, without using compex deductions. A metadata structure for handling mappings.
 
Trying to use a website, stuck because of a problem
chaalsCPH: (Leif's use case). E.g. not enough rights, accessibility problem, etc. Looking for a way for websites to declare what they require, that is machine readable.
 
formalizing mapping between words across languages
domCPH: I'd like to be able to describe translations in a machine-processable way, so that a computer can assist me and others in picking a term when translating a document; so that a computer can show me a list of translations available for one document
domCPH: I want to model the relationships between the various connotations of a word and its various translations in a given language; I want to model that a concept is defined in a given cultural/linguistic context (a referent context). I don't want to re-invent linguistic :)
domCPH: (related questions: Mapping between languages are hard to model, even using informal languages; can they be modeled in a useful way in formal languages? is there an existing basis on which this modeling could be built? how fine-grained should this be? how much a theoretical concept can be shared across languages and cultures? is the notion of concept even acceptable out of a well defined area?)
 
A multilingual interface for entering and editing
mortenf: An advanced version of something like foaf-a-matic
mortenf: ... driven (in part) by schemata
mortenf: ... but most importantly being accessible to all
mortenf: Example: Maintaing a group website for a community, with events, people and places etc. being decribed "correctly"
mortenf: ... in the prefered language(s) for each
mortenf: Also, presentation of e.g. search results with the right labels, including grammatical differences accross languages
mortenf: basically giving users power to choose their preferred language
 
A concept translation system
bengtfCPH: to send messages between some individuals that can be represented in various forms depending on their needs
bengtfCPH: and take into consideration differences between conceptual understanding
 
I would like to be able to manage my contacts and calendar so they could talk to each other.
chaalsCPH: I use a mobile phone/PDA, a laptop, and the web. So I want to be able to get my data from all these places. Web as in "from a public kiosk"
chaalsCPH: I want to be able to share some of the information with other people.
chaalsCPH: I want my calendar to understand where I am going to be, so it can look up people who are there.
chaalsCPH: I would like it to be able to talk to other people's calendars, so if I am going to be in the same place as someone I am trying to meet, the calendars can let us know.
chaalsCPH: I need to be able to extend the information in arbitrary ways
chaalsCPH: I want the name and address information to correctly encode people's names and addresses, which my current software doesn't manage nicely for different countries.
chaalsCPH: I want different ways of indexing people, since some people I know by a nickname, others by one or other part of their real name.
chaalsCPH: I want my calendar to know that journeys start in one place and end in another, and that there may be a time-zone shift.
chaalsCPH: I want nicer ways of managing the time that things happen in my calendar - knowing what is the "home timezone" of an event as well as what time it occurs wherever I happen to be.
chaalsCPH: I want to be able to put in things from non-Gregorian calendars without having to know when they are (for example, I find it handy to know when Greek and Anglican Easter and Lent, and Ramadan are, because there are events that happen "every second anglican easter".
 
 
chaalsCPH: There is an infrastructure for having translation, which is not switched on. Social management questions...
chaalsCPH: Based on SKOS - concepts rather than terms are the basic unit, so a translation is not direct.
chaalsCPH: Dom demonstrates the glossary. Developed by an intern at W3C.
chaalsCPH: The idea is to take the glossaries out of all the W3C specifications, and combine them. You have lots of terms entered into the system, extracted from the published specification, and turned into RDF.
chaalsCPH: Each term is assumed to refer to a concept, and is a mapping into english. There is also a definition (in english).
chaalsCPH: Select a specification from the list, and it gives the terms used.
chaalsCPH: You can search for all instances of a term.
chaalsCPH: e.g. user agent is defined 11 times.
chaalsCPH: So you can get 11 definitions for a term, to help you decide how to transate it.
chaalsCPH: Uses the SKOS schema from SWAD-E
chaalsCPH: You can add translations - terms and definitions in different languages, but this is turned off for social reasons.
chaalsCPH: You don't assert that you can translate the term the same way all the time, you say that two or three terms (maybe in different languages) are labels for a particular concept.
chaalsCPH: You can see where a term is used in definitions. It is not very clever - it just does text matching.
chaalsCPH: We would like to add functionality for linking concepts - for example "browser" is a label for a concept which is narrower than the concept labelled "user agent"
chaalsCPH: Uses XHTML embedded in RDF (because you can do that in RDF syntax).
 
Problem one - understsanding a menu
chaalsCPH: How do we know what size things are?
mortenf: lunch will be at Nyhavns Færgekro
chaalsCPH: What is a "typical" meal?
 
mortenf: in Copenhagen, by SWAD-Europe and CEN/ISS MMI-DC
chaalsCPH: Present: Charles McCathieNevile - SWAD-E project
mortenf: also present Morten Frederiksen
chaalsCPH: Dominique Hazaël-Massieux - W3C
chaalsCPH: Eva Mendez - Universidad Carlos III, Madrid
chaalsCPH: Leif Andresen, Danish National Library Authority
chaalsCPH: Bengt Farre - work on conept coding framework
chaalsCPH: Helle Bjarnø
chaalsCPH: Richard Ishida - W3C Internationalisation Activity
chaalsCPH: Liddy Nevile
chaalsCPH: Hosein Askari - working in the National IT and Telecom agency in Denmark
chaalsCPH: Matt May - W3C Web Accessibility Initiative (WAI)
domCPH: Charles MacCathie-Neville - W3C SWAD EU
 
Created by the Daily Chump bot. Hosted by PlanetRDF.