Meeting minutes
Annotations - model to add audiovisual comments
<wendyreid> w3c/
LaurentLM: the question is about audio/images annotations
… ... Ivan tried to modernize the body format property annotation which go beyond text e.g. audio
… yet another problem: where to start?
… after segmenting and sending to the system, the epub export make problems to emerge
ivan: is it realistic to transfer an ebook model for export/import?
LaurentLM: the audio annotations storage solution will not be in the epub itself
Ivan: when exporting annotations, the zip is exported with json+audio, which works
LaurentLM: which implies an internal structure inside zip
… the use case i'm worried about the annotations exposrted as part of the file
Hadrien: open domain content will not arise issues and the file sharing is ok
… authors writing their own annotations is a possibility hence making this solution a standalone is not the best, having both is the best
duga: adding audio etc is not new and feasable
… relative urls are already present, zipping the annotation will change the input format requirement, and might be confusing
… zip content should be transparent, if json is there it must be clearly stated
LaurentLM: the issue is that the section 6 of current draft state a singe file usage
Hadrien: OPF shouldn't be touched, by principle
... it will be wierd if multiple things are able to be exported/imported, zip at least open up an extension possibility, which is nice
wendyreid: packaging up and embedding into epub, beyond OPF, do we run up into valdiation issues?
… otherwise, outside of meta, should we target another new location?
LaurentLM: not sure if another folder can be put into meta
Hadrien: audio and other files in the back of epub is more part of an export action process
LaurentLM: the wording would be extended then
ivan: nothing in there forbids creation of a subdir for annotation
… with nothing inside, a flat tree structure
… on the other hand, the complications are overruled by user cases
CharlesL: the security is not to be overlooked
… the injection disguised as annotation is a posibility
<sueneu> +1 CharlesL
LaurentLM: it's true
ivan: the scope of accepted base media files should stay the same
duga: anything appropriate for epub must be definition be acceptable fo epub
wendyreid: there are reading system out there that support images not matching the dimensions showing screenshot of annotation in case of problem
Hadrien: I'm more worried about images than audio, sure mic needed etc
… but for example the text layout can be changed completely, the view too, the interoeprable fashion setting might not work as a result
GeorgeK: the annotation access for future use must be exportable too
wendyreid: export of annotation only or export of the whole book annotated co-exist
ivan: from specing POV, meta and data handling is not of our problem
LaurentLM: even if only text annotations exist, we export a zip, with json structure
ajellinek: zip works with export and import as it's widely used
wendyreid: there's a consensus, strcutre is to be defined
LaurentLM: once data is stored inside epub, the encryption poses a problem
Hadrien: in case of authored annotations, they must be encrypted
seuneu: +1 for authored annotations encrypted
… namely for copyright reasons, if though it might not be our problem
CharlesL: what about the superposition problem, i.e. annotation on annotation problem?
ivan: it's for the future given the can of worms this discussion opens up
… the publishers allows for export of annotations importable by the peers, which just gives the authorization
… a single file without a book containing annotations is ok
duga: DRM annotation: exportable or not exportable? that is the question
CharlesL: the other point: multiple students in front of the same book, does this situation produces a set of subdirs?
Hadrien: the attribution becomes a tricky issue now, should we open the Pandora's box?
wendyreid: the difference between new annotations and responding to existing ones are two different problems
… the recursivity of annotations is not a topic for now for the current model
LaurentLM: to be clear, each annotation has its own author
… all the annotations, whatever the author, are thrown into one universal bag
… i have a bigger cocnern with authored annotation in a DRM context because the encryption key makes the annotations intrinsic to the book itself
seuneu: authored annotations must be processed differently from non authored ones?
LaurentLM: Yes
wendyreid: annotated editions encapsualte annotations (i.e. they are part of the book by definition, by essence), thus are treating the annotations as integral to the book as a feature?
<ivan> +1 to what wendyreid said
duga: annotation mechanism should not be used in a DRM context
<gpellegrino> +1 to duga
<GeorgeK> +1 to what Brady is saying
<ivan> +1 to duga
<sueneu> +1 to duga
duga: annotations that are part of the content is not a valid use case here
… if we ignore it, the DRM problem goes away
Hadrien: mixed feelings here
… footnotes are horrible
… media overlay via authored audio is possible too
… people's creativity is without boundaries, thus we must be careful in our spec
wendyreid: i don't disagree, but the line must be drawn somewhere
… and it's : Authorized (created by author feat. publisher) vs Unauthorized (the other use cases)
seuneu: agree with Authorized vs Unauthorized idea
… but with improvements of already existing features
shiestyle: the old footnnote feature and this annotation feature are two distinct things
LaurentLM: improving footnotes with the annotation methodology is complex
… reading systems capabilities are dictated by the market
… and annotations can be DRM protected just in case
wendyreid: the thing to figure out: where to store these annotations, inside or outside?
ivan: my feeling -- this issue is PR-ready
… that's the next step