W3C home > Mailing lists > Public > public-xg-lld@w3.org > January 2011

Re: YouTube video?

From: Thomas Baker <tbaker@tbaker.de>
Date: Wed, 19 Jan 2011 12:31:42 -0500
To: public-xg-lld <public-xg-lld@w3.org>
Cc: Joe Provenzano <provenzano@wis.edu>
Message-ID: <20110119173142.GA3552@octavius>
Dear all,

I have been remiss in following up on our discussion in
Pittsburgh of a YouTube video.  For those who missed it:

-- People like the idea of a YouTube video.

-- The video should be playful, even irreverent.

-- Because the topic is so deeply serious, it was felt
   that a focus on Anne Frank (see below) would not fit well
   with a playful or irreverent approach.

-- "Around the World in Eighty Days" was proposed.
   However, it was felt that the video should be grounded
   in a non-fiction topic (and not, for example, centered
   on a film or book).  For example, one possible starting 
   point could be Michael Palin's BBC documentary tracing
   the footsteps of Phileas Fogg.  The title "Triples around
   the world in eighty days" was proposed.

I wanted to hold off commenting on the Michael Palin approach
until I'd had a chance to view the videos -- something I
could finally do over the holidays.

As a first impression, I'm a bit foggy on how we
might go about extracting from this complicated and rambling
odyssey a simple story line, let alone one that creatively
uses links taken from sets of real Linked Data.  Even if we
could jump from dataset to dataset in a Fogg-like manner,
following a Fogg-like path, I have difficulty imagining
how we could do all this -- and establish the "Eighty Days"
framework -- all in five minutes.

It occurs to me that a completely opposite approach could
be to plunge into Linked Data at some familiar entry point,
such as Wikipedia, and follow our noses to wherever we detect
a promising scent, almost at random.  I'm thinking of some
Monty Python sketches in which they jumped around in quick
succession, with absurd segues and cut-aways, between people
in wildly different situations.  The Linked Data equivalent
could be to pick a triple and jump to something with a shared
subject, there to pick another triple and jump to something
with the same author, then to a shared birthplace, publication
date, and the like -- random-seeming flashes which guide the
viewer on a lightning tour of the data cloud.

With just four months left in our one-year LLD XG
roller-coaster ride, I do not see how we can spend
incubator-group time on this topic, and we should consider
removing it from the running agenda.  However, if others are
interested in brainstorming about the possibilities, please
post or drop me a line.  If we could find the right approach,
it could be fun...

Tom

On Thu, Oct 14, 2010 at 01:57:22PM -0400, Thomas Baker wrote:
> Some of you may have noticed a new item on the Pittsburgh
> agenda -- a YouTube video on library linked data [1 and below].
> I am in touch with instructors of art and Web design at a
> secondary school in Washington DC that is interested in doing
> the production side.
> 
> The idea is that the video would show visually how statements
> about things correspond to links in graphs and how different
> graphs can be linked with each other.  Visually, we're
> thinking along the lines of stop-action video using either
> Tinkertoys or the sorts of model kits one uses in organic
> chemistry courses to construct molecules.  Some animation
> could be used.  Kids at the school would provide voice-overs
> but no talking heads would appear in the video.  Since it is
> an international school, they are especially interested in
> the idea of voice-overs (with sub-titles) in languages they
> emphasize, such as French, Chinese, and (unusually) Dutch.
> 
> The role of LLD XG (and DCMI) would be to provide a script for
> the video -- real examples of linked data, and to specify a
> sequence by which different graphs would be constructed and
> linked with each other.
> 
> When asked for a book that could provide a focal point for
> the video, the school librarians suggested The Diary of Anne
> Frank -- a book that all the kids know, and that kids in other
> countries are also likely to know and relate to.  It looks to
> me like a perfect focus because it has been translated into
> every major language, adapted in films, and is linked to so
> many topics (and also happens to be Dutch in the original).
> 
> The task of LLD XG, as I see it, would be to distill out
> of the endless possibilities a simple story line, starting
> with a description of the book in English and linking out to
> a translation or two, film adaptation, further information
> about Anne Frank, Anne Frank House, and historical context.
> Things like FRBR, authority control, and Wikipedia could be
> worked into the narrative.  The possibilities are so vast that
> the biggest challenge, it seems to me, will be to narrow the
> focus enough to fit into a short video.
> 
> The examples should be real, the connections understandable,
> and I'm thinking the film should end up showing a visually
> compelling cluster of information.  If we give them a good
> story line, the kids can work with their instructors to make
> the film graphically engaging.
> 
> Everyone I have spoken to about this idea is very enthusiastic
> and motivated, especially at the school.  
> 
> Though it is on the agenda for the face-to-face in Pittsburgh,
> we may not have much time for it during the meeting itself
> because we have so much else to discuss.  However, this is
> a great topic to brainstorm over beer in the evening.
> 
> If you all like this idea, it would be great if you could
> have a look in your favorite linked data sources over the
> coming week to see what you find -- particularly in French,
> Chinese, and Dutch.
> 
> Tom
> 
> [1] http://www.w3.org/2005/Incubator/lld/wiki/YouTube_Video
> 
> ----------------------------------------------------------------------
> Project: YouTube Video on Triples in Linked Data
> 
> Version: 2010-10-12
> 
> Coordinators
>     Tom Baker <tbaker@tbaker.de>
>     Joe Provenzano <provenzano@wis.edu>, Washington 
>         International School, http://www.wis.edu/
> 
> Concept
> 
> -- Video showing triples being constructed into 
>    graphs using tinkertoys or biochemical model kits.
> 
> -- Scripted by the W3C Library Linked Data Incubator Group
>    (at any rate the part about example triples used).
> 
> -- Narrated by WIS students.
> 
> -- Stop-action video, with hands coming in to connect
>    new triples to the growing graph.
> 
> -- Other visually interesting elements, such as animation,
>    perhaps scripted by the students, for example to convey
>    the notion of mashing-up data from different sources.
> 
> -- Produced as a co-production of:
>    -- Washington International School (WIS)
>       http://www.wis.edu/
>    -- W3C Library Linked Data Incubator Group (LLD XG)
>       http://www.w3.org/2005/Incubator/lld/
>    -- Dublin Core Metadata Initiative (DCMI)
>       http://dublincore.org/
> 
> -- LLD XG and DCMI to provide a basic script with voice-over.
> 
> -- Tom to work with Joe and his students by explaining the
>    content, brainstorming with them about presentation, and 
>    providing input and feedback on the visual results.
> 
> -- Published on YouTube.
> 
> Brainstorming...
> 
> -- Tom: Video could start slow, connecting up a few triples,
>    then once the main idea has been presented, the action could
>    accelerate, with hands flying in from right and left until
>    a complex graph of linked data has resulted.
> 
> -- Tom: Script could include merging data in French, Dutch, or
>    other WIS languages, with the voice-over spoken by
>    native-speaker students, with subtitles in English, tying
>    in with the spirit and mission of WIS.
> 
> -- Antoine: Using FRBR, show how the more intuitive "work"
>    notion can allow to provide access to all these URIs
>    of book-related E/M/Is (or any mixing of them) in a
>    multilingual domain. Starting with one language-specific
>    E/M/I worked out in RDF, then have hundreds of balls thrown
>    at a poor guy with as many language-specific titles voiced
>    in the background. But FRBR comes to the rescue, bringing
>    structure with one magic ball that connects them all.
> 
> 
> -- 
> Tom Baker <tbaker@tbaker.de>

-- 
Tom Baker <tbaker@tbaker.de>
Received on Wednesday, 19 January 2011 17:32:20 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 19 January 2011 17:32:22 GMT