AW: [EMOXG] timing spesification

Hi Marc and all
This is to answer your questions about my timing suggestion:

1.
* date
* timeRefUri
* timeRefAnchor
* offsetToStart
* start
* end
 Are all of these optional?

A: Yes, I see no reason not to make them optional with the exception of common dependencies, e.g. a timeRefAnchor or offsetToStart do not make sense without a timeRefURI.
 
2. Should there be any constraints about the combinations in which they 
may occur?

A: The only restriction I see is the dependency of timeRefAnchor/offsetToStart to a timeRefURI.

3. Is the distinction in meaning between "offsetToStart" and "start" clear?

A: I think "start" is obvious and "offsetToStart" was taken from EMMA which I think is good.

4. Where exactly can the attributes occur? Can all occur on <emotion> 
and on <link>? Or can "date", "timeRefUri", "timeRefAnchor" and 
"offsetToStart" occur only on <emotion>? (in this case, the time 
indication on a <link> would be only "start" and "end" -- maybe that is 
enough?)

A: Good hint, probably better to be restrictive here and less arbitrary, let's do it like that until proven better.

5. One of your examples uses an "id" attribute on a <link> tag. Can we 
try to simplify by putting "id" attributes only on <emotion> tags, or is 
that too restrictive?

A: There might be an issue if you want to distinguish 
  a) the (relative) time of an emotional display from 
  b) the (absolute) time of the emotion itself 
for a) you'd refer to a link element, for b) to an emotion element, 
but again I think we could start more restrictive and follow Marc's suggestion until prospective users protest.

6. If I read [2] correctly, absolute time in your example should be 
represented as "2001-11-23T14:36Z", where "T" is the date vs time 
delimiter, and "Z" means UTC, no?

A: yes. I wasn't sure how to denote date before christ though.

Talk to you later,
felix


-----Ursprüngliche Nachricht-----
Von: Marc Schroeder [mailto:schroed@dfki.de] 
Gesendet: Montag, 3. November 2008 09:19
An: Burkhardt, Felix
Cc: public-xg-emotion@w3.org
Betreff: Re: [EMOXG] timing spesification

Felix, all,

I think these are very reasonable suggestions.

If I may try to summarise, you are suggesting the following attributes:

* date
* timeRefUri
* timeRefAnchor
* offsetToStart
* start
* end

Here are a few questions:

1. Are all of these optional?

2. Should there be any constraints about the combinations in which they 
may occur? (e.g., when indicating start, do you have to indicate end? 
Can "timeRefAnchor" occur without "timeRefUri"? etc.

3. Is the distinction in meaning between "offsetToStart" and "start" clear?

4. Where exactly can the attributes occur? Can all occur on <emotion> 
and on <link>? Or can "date", "timeRefUri", "timeRefAnchor" and 
"offsetToStart" occur only on <emotion>? (in this case, the time 
indication on a <link> would be only "start" and "end" -- maybe that is 
enough?)

5. One of your examples uses an "id" attribute on a <link> tag. Can we 
try to simplify by putting "id" attributes only on <emotion> tags, or is 
that too restrictive?

6. If I read [2] correctly, absolute time in your example should be 
represented as "2001-11-23T14:36Z", where "T" is the date vs time 
delimiter, and "Z" means UTC, no?

Best,
Marc

Burkhardt, Felix schrieb:
> Hi all
> This is the answer to my action item from session 2 of the Cannes f2f
> for suggesting a revised timing specification as stated in the minutes
> [3].
> 
> Issues that shall be included:
> 1. to denote "start/end" 
> 2. to anchor timing relative to other events
> 3. to denote absolute time points, because this would be needed
> e.g. for the "emotional diary" application
> 
> Issues that will be disregarded in the first version
> 1. "duration" as it is redundant
> 2. sample values
> 3. onset/hold/decay, as they could be expressed by a trace of
> intensity or by using several emotion statements.
> 
> GENERAL REMARKS
> a) It's a question how to denote relative timevalues, one could use
> simple integers for milliseconds, I vote for the time values like
> specified in the SMIL specification [4] and will use them in these
> examples.
> b) We could either use preferably attributes or elements, I think
> attributes
> may be better as they give a more "compact" code.
> c) Attribute names could be written with a hyphen in them (time-ref-uri)
> or "java-style" (timeRefURI). I prefer the latter. 
> In any case this should be consistent throughout the markup.
> 
> ABSOLUTE TIMEVALUES
> I suggest to use the ISO-8601 standard to denote absolute timepoints
> as it is also approved by the W3C [2].
> Absolute timevalues are only needed directly under "emotion" element.
> The following example states "joy" on the 23rd of nov. 2001.
> 
> EXAMPLE 1:
> 
> <emotion date="2001-11-23 14:36">
> 	<cat name="joy"/>
> </emotion>
> 
> RELATIVE TIMEVALUES
> Can be used to denote time intervals in media changing over time.
> E.g. by attribute start and end. The following example states "joy"
> displayed in a video from the 3rd until the 9th second.
> 
> EXAMPLE 3:
> 
> <emotion>
> 	<cat name="joy"/>
> 	<link uri="videoData.avi" start="3s" end="9s"/>
> </emotion>
> 
> Of course this coud be combined with absoulte time, e.g. the following
> states joy on the given date displayed in the video at the given
> timeinterval:
> 
> EXAMPLE 4:
> 
> <emotion date="2001-11-23 14:36">
> 	<cat name="joy"/>
> 	<link uri="videoData.avi" start="3s" end="9s"/>
> </emotion>
> 
> ANCHOR TIME TO OTHER EVENTS
> It should be possible for Emotion statements to be anchored relative
> with respect to other events. I suggest to use similar attributes as
> the EMMA specification [5].
> Attributes
> timeRefUri : URI of an element that is the anchor
> timerefAnchot: either "start" or "end" indicating whether to measure
> 	       offset from start or end. optinal: Defaults to start 
> offsetToStart : time value indicating an offset: optional, defaults 0.
> 






> The following example anchors the display of Fred's sadness 12 seconds
> after the end of the display of Anna's joy in a video.
> 
> EXAMPLE 5:
> 
> <emotion id="annasJoy">
> 	<cat name="joy"/>
> 	<link id="someVideo" uri="videoData.avi" start="3s" end="9s"/>
> </emotion>
> 
> <emotion id="fredsSadness" timeRefUri="#someVideo" timeRefAnchor="end"
> offsetToStart="12s"/>
> 	<cat name="sad"/>
> </emotion>
> 
> Of course the sadness of Fred could also be anchored by Annay joy
> itself, so in this example Fred was sad on the 23rd of nov. 2001, 12
> seconds past 14:36. 
> 
> EXAMPLE 6:
> 
> <emotion id="annasJoy" date="2001-11-23 14:36">
> 	<cat name="joy"/>
> </emotion>
> 
> <emotion id="fredsSadness" timeRefUri="#annasJoy" offsetToStart="12s"/>
> 	<cat name="sad"/>
> </emotion>
> 
> Regards,
> felix
> 
> 
> [1] http://en.wikipedia.org/wiki/ISO_8601
> [2] http://www.w3.org/TR/NOTE-datetime
> [3]
> http://lists.w3.org/Archives/Public/public-xg-emotion/2008Oct/0020.html
> [4]
> http://www.w3.org/TR/SMIL3/smil-timing.html#Timing-ClockValueSyntax
> [5] http://www.w3.org/TR/2004/WD-emma-20040901/#s3.2.10
> 

-- 
Dr. Marc Schröder, Senior Researcher at DFKI GmbH
Coordinator EU FP7 Project SEMAINE http://www.semaine-project.eu
Chair W3C Emotion ML Incubator http://www.w3.org/2005/Incubator/emotion
Portal Editor http://emotion-research.net
Team Leader DFKI Speech Group http://mary.dfki.de
Project Leader DFG project PAVOQUE http://mary.dfki.de/pavoque

Homepage: http://www.dfki.de/~schroed
Email: schroed@dfki.de
Phone: +49-681-302-5303
Postal address: DFKI GmbH, Campus D3_2, Stuhlsatzenhausweg 3, D-66123 
Saarbrücken, Germany
--
Official DFKI coordinates:
Deutsches Forschungszentrum fuer Kuenstliche Intelligenz GmbH
Trippstadter Strasse 122, D-67663 Kaiserslautern, Germany
Geschaeftsfuehrung:
Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster (Vorsitzender)
Dr. Walter Olthoff
Vorsitzender des Aufsichtsrats: Prof. Dr. h.c. Hans A. Aukes
Amtsgericht Kaiserslautern, HRB 2313

Received on Thursday, 6 November 2008 12:22:39 UTC