Digital Publishing Interest Group Teleconference -- 10 Feb 2014

<trackbot> Date: 10 February 2014

<lizadaly> I pointed to the wrong URL in the agenda, my apologies

<lizadaly> We're discussing Pagination/DOM interaction under http://www.w3.org/dpub/IG/wiki/UseCase_Directory

<lizadaly> SCRIIIIIIIIBE

thanks Zakim

<scribe> scribe: azaroth

<ivan> scribenick: azaroth

Liza: Can we approve the previous minutes?
... Assuming no objections? Okay... approved.
... There were corrections that Karen missed in the previous set. Don't know if they've been updated or not

tmichel: They've not been updated, so we should leave

Liza: That's fine. Markus isn't back till next week

<tmichel> s/???/tmichel

<lizadaly> http://www.w3.org/dpub/IG/wiki/UseCase_Directory

Liza: Okay, there are use cases for pagination posted at the bottom of the use case directory. About 6 of them from Brady

<astearns> http://www.w3.org/dpub/IG/wiki/UseCase_Directory#Pagination.2FDOM_interaction

Liza: he isn't going to lead the task force so one goal is to find an owner to shepherd them
... Brady would you mind talking us through the goals for the TF?

brady: This came up in China. We talked a lot about how you style a page, Dave is doing latineq and lots of work being done in CSS
... but for reading system we need to know more than just how to break it into pages, but also how to reference a page
... so have to trick things into thinking that they're pages. RS needs to know how many pages are in the document, how to change the page size, display multiple pages at the same time
... some sort of object model for what a page is and how to refer to it
... that's the use cases. In China I was asked to write up a few, which are now on the wiki
... goes beyond style, into a model.

Liza: Assumption is it's about how a RS would behave, not necessarily how it would be supplied?

Brady: Wrote as a RS implementer, but could be use cases from the publisher side
... Publisher has lots of style issues, but also may need to know about where pages are and how they work. Esp if there's javascript
... a JS widget that needs to know how many pages there are, the publisher would be interested

dauwhe: A couple of things from CSS point of view, some discussion about this sort of thing in CSS, important to be able to know what a page is and how to address it
... work on fragmentation might help here. Publishers interested in styling things based on which page they're on and the position within the page
... so far css isn't really addressing that issue. working on selecting the master page in the InDesign sense, but that doesn't select the content of the page
... lots of pieces to this

Ivan: Ran into a problem that relates -- a book that also had illustrations and caption of illustration was quote from the text. Quote said this is on page 622. And then it wasn't, depending on reflowing.
... so publishers have to have some special way to encode it, so the RS can get it right. Can't completely forget about it

dauwhe: generated content for paged media module has that kind of page referencing property. Can retrieve the page number from the link and display it to have the right number

Ivan: Publishers have to know how to do that.

dauwhe: And it has to be implemented. Not sure of implementation outside of Antenna House

Bill_kasdorf: It needs to be embedded in teh content? or working interactively?

<lizadaly> Bill_Kasdorf was Bill_Kasdorf

dauwhe: You can say link to a heading somewhere ... there's an illustration far away. THen you write CSS to say get the ID value and find the page value, display that here

tzviya: If we're working on implementing JS, we can call it a viewport or page, we don't know where things are going to land on the page. Eg a touch event, we want to contain it to a particular screen

lizadaly: I wasn't really considering publishers writing interactive components inside a book, not common now but should be

<dauwhe> http://dev.w3.org/csswg/css-gcpm/#target-counter

lizadaly: they'll want a common way to reference pages

tzviya: page references in a cook book, you might say see recipe on page 72, but in academic work citations are more complicated
... working with references across works
... probably not in scope, but want to consider it

lizadaly: linking is important, but should probably confine it to pagination rather than the whole world of references
... sounds like someone other than a RS vendor should come up with publisher centric use cases before we can say it's complete

<dshkolnik> In digital magazine world we see a lot of custom JavaScript from publishers. The contract between Adobe Content Viewer and the publishers content is that we call onAppear()/onDisappear() when page is displayed.

ivan: what's the next steps? in China we wondered whether it means we need changes or extensions to the current DOM model, or not?
... that would be a very specific, heavy technical requirement
... need to be careful before issuing such a requirement
... is that where we're going, or can we get around it?

lizadaly: no way to reference a particular page in DOM other than with lots of JS hacks
... can't even say what page you're currently on.
... need a solution

brady_duga: Agree there isnt' a solution today. Two options -- modify the way DOM works or layer a solution on top of it
... that might be possible without changing the dom, but some other way to map between elements in the DOM and pages
... only want to show this portion of the DOM because that's the page we're on. DOn't know which is the way to go

dauwhe: Actually changing the DOM seems like an incredibly high hurdle. not an expert but lots of work on extensions to the DOM to talk about things not really in the dom like pseudo elements
... wonder if there's possibility in that space to be able to manipulate the objects we want to talk about

astearns: one possibility for building on top of the dom is to use css regions.
... when you fragment content with regions you have some scripting access to the region chain, you can find out which position the region is in the chain
... so you can find out what page content is on, so long as you map regions to pages
... so can build on top of that as it lets you do some of the things raised

lizadaly: Remember an early draft of the spec, excited that it gives that level of access.

astearns: the spec is moving along. Mozilla is not terribly interested
... Chrome decided to take it out for now as mozilla aren't doing it
... implementation in safari, and older version in IE. Need to wait for other browsers

ivan: good example. Don't know why mozilla isn't interested, but tells me that if it's a solution or main component for the publishing industry, then the IG should make it clear to the parties
... that as far as we're concerned, this should have a high priority
... as it solves a big problem. Alan you're on the CSS WG, what can this IG do to try to increase the priority of implementation?

lizadaly: Helpful to say that it's very important

astearns: I think it would be good to talk about use cases rather than specific technology
... knowing the page number, what content is on the page, etc. need to get that information from the reading system to act on
... noting regions has the capability built in. A better way to put pressure.

dauwhe: wondering if regions can address these kinds of issues? Can it do it behind the scenes without publishers writing in all the region machinery in CSS?
... can it interpret pagination or something a RS could use to implement pagination and then expose hooks without publishers having to assign everything individualy

astearns: Good question. Goes to promoting use cases not technology
... believe it will be that there'll be JS libraries built on top of regions that gives the functionality
... or we look at how regions and js interact, and find a lower level way in CSS to do the same thing
... so regions+ js to me is the right way to address a lot of the use cases, but I could be wrong

lizadaly: when i first saw some region support in chrome, pagination and seeking to a page, can't say it would solve all the use cases here, but was very promising
... better than previous solutions

astearns: Some limitations, when you get to very long documents
... regions + script may not be performant enough. might inform where we need to work.
... Brady and Rosen from MS were talking about the next steps and how to get a long pagination scenario working okay

lizadaly: agree that's been a problem with all the solutions I've seen so far -- they fall apart in long streams

ivan: not like trying to push anyone but this tells me it would be relatively urgent to get these use cases documented and published
... could be the time when the pressure could work
... if we are too late with the UCs then could be too late and will be pushed back to CSS4

brady: Curious what the next steps are? Not sure I even proof read them. Need more people to add use cases?

ivan: At some point we need to publish working draft -- a more formal doc from the IG -- that we can put on the desks of appropriate people

dauwhe: another use case is the desire to style things on page position, eg recto/verso
... do those use cases help?

+1 to r/v from medieval manuscript scholar

lizadaly: changing properties of the text based on pagination

dauwhe: a better foundation for pagination than currently

lizadaly: would like to see some publisher focused use cases listed
... styling or js apis that publishers could make use of that understand pagination seems relevant
... might want to be explicit that performance is important

<tzviya> I was thinking of Random and Simon

tzviya: Simon & Schuster, also Random House

dauwhe: wanting to do things particularly with a spread

ivan: who should reach out to Lisa ?Cachin?

<lizadaly> Liisa

dauwhe: I will
... action on Ivan to get them to join :)

ivan: reaching out a good idea.

tzviya: S&S work on children's books a lot. but if Liisa has good examples that's probably sufficient
... RH is a great start

ivan: what about STEM? mathematical text books with cross refs in pages, formulae etc was incredibly important
... can also ask Jean

tzviya: can talk about over lunch at EduPub

lizadaly: Another pass through the use cases would be good

brady_duga: I'll take another pass through them

lizadaly: publisher focused ones will take a little longer

ivan: brady are you in a position to lead the TF?

brady_duga: I think we should find someone else, if I take the leadership it won't get done
... can't really take on anything much at this point and expect anything to happen with it
... can help out

ivan: so next question is who?

tzviya: I vote for Liisa :)

ivan: There was a meeting last week but I don't think it would be appropriate at this moment

<lizadaly> but +1 :)

karen: Don't want to discourage reaching out, but they don't see a compelling reason to participate at the moment

ivan: We'll reach out anyway for use cases regardless

karen: I think she'd be happy for use cases

ivan: Getting her as an IE probably not doable at the moment
... so need someone else

dauwhe: would love to but am wildly overly committed

tzviya: have you heard from elsevier?

ivan: contact goes back 5 or 6 years.
... oxford university press ... some contact but not the right person. OUP is a member as part of Oxford.
... matter of finding the right person

lizadaly: we build their impelemntations so can talk to team and find the right person

ivan: I can reach out but several attempts would be good

lizadaly: we normally work with humanities and reference group

ivan: a problem we have to solve, but not necessarily right now

lizadaly: This was only thing on the agenda

<Luc> Sorry, I have to leave

<tzviya> azaroth: we discussed approving annotation use cases and added accessibility to use cases

<tzviya> ...should i begin work on note?

ivan: starting to work on note will help to clean up use cases

<lizadaly> (Thanks tzviya)

ivan: so I would start turning them into a real document. While doing that some issues will come up to discuss
... you have the infrastructure?

azaroth: Yup

ivan: If you need help you know where to find it :)

lizadaly: AOB?
... Okay, thanks every one

- DRAFT -

Digital Publishing Interest Group Teleconference

10 Feb 2014

Attendees

Contents

Summary of Action Items

Scribe.perl diagnostic output