See also: IRC log
Priority for this morning is pagination
We will deal with logistical issues later to use this time most effectively.
(Quick round of self-introductions.)
We will not try to mimic CSS WG
scribe: instead we need to undestand the
roadmap of how to get there.
... Many different ways to approach this.
... DPIG needs to understand what is the _best_ way to accomplish what
we need to do.
Both WGs have a problem of resource scarcity.
We hope this pagination issue may help us recruit new people
scribe: especially new people who can contribute insights as to pagination requirements.
There is no canonical document regarding pagination.
This is a problem for the WGs because they need to be told more clearly what the issues are.
We will avoid low-level discussions about CSS syntax.
We want to have a clear roadmap at the end of this session.
Daniel: First step is to define what is
meant by pagination.
... Need a definition of what is a page.
... seems like an obvious question with an obvious answer but it isn't.
... How do you interact with a page, view a page, etc.?
... transition between pages is undefined and badly needed.
scribe: intersection between device industries and publishing is key.
Dave: We absolutely need to address all
users.
... presentation--unfortunately since the Web lacks native pagination he
can't just put this online. . . . ;-)
... this presentation is geared for people who may not have deep
knowledge of either the technology or the publishing requirements.
... Pagination addresses need for content that doesn't just fit on a
stone tablet: even God realized that. ;-)
Dave: How to deal with content that
"doesn't fit"?
... Most common response on the web is to scroll.
... In some contexts the best user experience is to paginate, move the
additional content to a new "page."
... This also involves the need to hyphenate (when a word doesn't fit)
which requires hyphenation rules and dictionaries.
... So what is the arrangement of the pages in virtual or physical
space?
... Horizontal? LtR? Vertical?
... Spreads?
... Useful to be able to view multiple pages that are not necessarily
contiguous in the linear flow.
... CSS has already done quite a bit of work on how to break up the
content: "Fragmentation."
... CSS Fragmentation Module Level 3 latest draft is Jan 2014.
... Overflow piece for how pagination is triggered and what the rules
are for chunking content into discrete pieces.
... We also need to define how one interacts with pages.
... Brady put together a bunch of use cases for page representation in
the DOM, or perhaps Boxtree.
... Mechanism for accessing pages, displaying multiple pages, various
arrangements, etc.
... Plus the need to _identify_ what page you are on, what page you want
to navigate to, etc.
... Page transitions (within and between documents).
... Content based on position (e.g. top or bottom of page).
... Retaining positoin when font size changes.
... User overrides.
... Fundamental principle of DPIG is user must have some control of the
presentation.
... [showed a simple structural diagram of the DOM]
... and diagram of how the html parser creates the DOM, the css parser
creates the CSS Object Model, and the content is rendered.
... and then a diagram showing the interactions between the DOM tree and
the CSS, and how CSS builds the rendering.
... The Path Forward: How do we bring pagination to the web natively?
Karen: Question about the spread slide, clarifying what is meant by spreadsd.
Bill: Some types of content (textbooks, magazines) have objects that extend across a spread.
<liam> [ http://4.bp.blogspot.com/-4jxNVJYcT40/T1lYMlD-QtI/AAAAAAAAdmA/Z690oSQDeu0/s1600/photo-745957.JPG ]
<bobtung> dpub agenda: https://www.w3.org/dpub/IG/wiki/TPAC2014-F2F
<tzviya_> fragmentanors
<tzviya_> glazou: automotive industry, want to have multiple documents with ability to "flip" through, similar to ebooks but differenrt
<Bert> (And LG's round watch, presented at a break-out ysterday)
<bradk> some related CSS drafts: http://www.w3.org/TR/css-overflow-3/ http://www.w3.org/TR/css3-break http://www.w3.org/TR/css-shapes-1/ http://www.w3.org/TR/css3-regions http://www.w3.org/TR/css3-multicol http://www.w3.org/TR/css3-page
<ivan> scribenick: tzviya_
alan: opera has implemented high-level
overflow:paged. IE and Safari have implemented regions to accomplish
some of this, such as knowledge of location ("page")
... hope other browsers will adopt
Deborah: we've been able to hack pagination models since practically the beginning of the web
Dave: The developments in IE and Safari reduce the amount of work you need to do to accomplish thisl
Deborah: this solves some but not all of the issues we're talking about.
Liam: We have closed the XSL-FO working
group; in fact I did that: nobody was showing up for the meetings.
... The other reason was that it was clear that the longterm future was
to make CSS do what can be done with XSL-FO
Liam: So I closed FO and joined the CSS WG and gave a presentation, afraid there would be a huge push-back re getting CSS to where FO was
In fact there was good support in the CSS WG because they want to be able to produce professional quality typography.
CSS is moving to replace ad hoc features with underlying mechanisms: big change.
E.g. we have flexbox and regions coming.
So we have an opportunity for redefining some of the mechanisms in CSS by which pagination happens
scribe: having an underlying consistent
mechanism.
... reinforced by needs of ebooks.
... early days but looking very promising.
Michael: AntennaHouse supports CSS for
pagination.
... At the same time, FO is very alive and strong in the user market:
several thousand FO implementations, 10,000+, 100,000+ licenses.
... Right now FO sales are stronger than ever.
... So despite our commitment to CSS we are committed to FO.
... Would like to see FO revitalized in the W3C.
... Nine years ago in Heidelberg AH advocated bringing FO and CSS
together and got crucified.
... We would like to make that suggestion again now.
Markus: am I hearing that you are advocating moving feature set of FO to CSS or actually reviving FO?
Michael: actually both. Massive installed
base (e.g., almost entire aerospace industry) uses FO and is not going
to move off it.
... Thousands of existing implementations that aren't going to change.
Feature set of FO is what's needed in CSS.
<Zakim> liam, you wanted to note FO is also based on CSS
Liam: FO uses CSS properties and adds
region and page properties. CSS was not a recommendation when FO was
first published.
... So CSS properties are put in line with regions and flows.
... Antenna House has a lot of extensions that are also important.
Michael: In AH documentation we show
alignment between FO and CSS.
... Offers copy of Formatter to members for non-commercial use.
Ivan: remember that we don't want to get
too far down the technical path in this IG
... What can this IG do to help those in CSS who want to move forward on
this?
Bert: Writing up the requirements as Dave
has done is very valuable.
... The use cases are very good.
... Also, come and give demos of things you like, need to do, etc.
... They need to understand that pagination is not just breaking
documents into fragments. It's much more than that.
Dave: A lot of this is in LatinReq. What
else is being looked for that is not in LatinReq?
... In CSS, Fragmentation, Overflow, CSS Pages > Template, CSS Box
Model, and Page Transitions and Navigations are being worked on.
... DPIG can provide a lot of insight into how these things can best be
accomplished.
<bobtung> I think Latin Req is on i18n review list https://www.w3.org/International/wiki/Review_radar
Dave: what else is needed from the user community to move things along?
Michael: has there been any attention to digital rights?
Dave: no. That's out of scope for what we're working on.
Daniel: Remember that DPIG is an interest
group, not a working group.
... We have to make a strong use case for doing this on a broad scope.
<azaroth> +1 to Daniel's suggestion
Ivan: We have LatinReq that mostly addresses a corner of the broader use case.
Dave: we need a document called CarReq ;-)
Daniel: No, that's not what I'm saying. We
need to document number of tools, actors, size of market, number of
players, etc.
... To show that there are a whole bunch of new users that need these
things.
... Then sure, for each of these industries, list what the gaps are.
... Then, at the end of the document, have a table listing all the
features and all the industries that need each one.
Ivan: Huge amount of work, but I
understand.
... But to make it complete it should not be just from Digital
Publishing.
Daniel: Automotive group may not be addressing this but automotive _industry_ needs it.
Bill: Advocate a broader understanding of what is meant by publishing.
Dave: concerned that if it is so broad that it is everything people do on the Web it becomes meaninglessl
Karen: building on what Daniel said about
automotive, we are having conversations with a number of industries
(healthcare, aerospace, etc.) that all recognize that they have
publishing needs.
... How do we get the conversation to happen between these groups?
Tzviya: We don't want philosophical
conversations about what "publishing" is or what a "book" is. That
doesn't really matter.
... we need to understand who needs to work on this.
Ivan: we do have contact with people who
are more on the business side.
... could use LatinReq as a model and try to match that up with what
their needs are.
... Realistically we have to start with the industry we have around the
table.
... Once the document is done, then make an outreach to Boeing, etc.
... Detailed business case must start with the publishing industry as
now represented in the DPIG.
Daniel: Showed a "state of art" document,
not technical, addressing issues like "What is a Page Transition?"
... Included examples going back to IE9, PowerPoint, etc. which provide
a lot of insight into how the needs are remarkably coherent and could be
standardized.
... Some transitions are extremely complex, need very detailed
descriptions.
... We need the spec to be "extremely human-readable."
... Comparing Keynote to PowerPoint is instructive.
... Looking at electronic books is part of the picture.
... Widgets, connected appliances in industry is an issue.
... And he provided suggested action.
... What is missing is the business part, what is the size of these
industries, how many millions of users are waiting for these things?
Markus: We are looking for work items.
Bert made two suggestions. PageReq plus come and give demos.
... Need for people to see demos, say "oh, now I see what you mean."
Ivan: Is it possible to post Dave's demos?
... through our website? would that help?
Bert: yes, sure.
Ivan: Not having it only at a meeting, but
generally available.
... Let's do that. We can build this from what we've got.
Markus: We need to understand the overall
scope first.
... What else?
Daniel: In my opinion, the business case.
Markus: Do you need editors who understand these issues deeply?
Daniel: yes, badly.
bill_kasdorf: there is not really a
difference between what tzviya and daniel are tzviya are saying
... other industries may be a subset of publishing inudustry
... start with publishing, then go to other industies and ask what's
missing
Markus: We need to answer Dave's question about what is needed in the PageReq requirement.
Deborah: All in favor of getting others at the table, but if you define publishing as anything people do on the web, you're talking about the W3C
Markus: we have to work with what we have.
... We can't have a dependency on the broad scope at the outset or we
will never get it done.
Ivan: We need real action, things we can do now.
Markus: In order to have concrete marching orders
...we have identified the page req document
...to describe pagination requirements first for digital publishing
...automotive is not a dependency
...we'll reach out for review, but we will start with digpub as we know today
...Dave, you indicated some confusion or concerns
...to get started, would you verbalize again please
Dave: hard to talk about things that confuse me
...latin req
...is talking quite a bit about what factors
...we're thinking about and what rules we apply as we break content in pages
...constraints on page depth; want base lines to be the same on consecutive pages, especially on spread
...widows and orphans
scribe: breaking lines and a reading experience; issues of keep width
...not separate the heading from the content because it confuses the reader
...There is an initial discussion...I think of those things as requirements for pagination
...they are describing where it's appropriate to break content
...and what we think about as we make these breaks
...and paginate in a way that is as reasonable as possible
Bill: hyphenation included?
Dave: to some extent; where can I go from one line box to another
...concept of @ hyphen
...I have to turn page before reading a single word
Bill: even within Latin languages, in German
...when you hyphenate decken, that changes to a k
...should we exclude that from scope?
Brady: This is a great document
...point of page req is to take it in another direction
...action was on me and that was a mistake
...Latin req looks at layout and styling, how to determine breaks
...pagination is about reading system; page 27 and how to get reader to read it
Dave: Closer to the page DOM requirements
Brady: right; write from the implementer side
...put a URI on top of it
Dave: you are describing a spec?
Brady: not necessarily; lots of ways to approach
...end of day, I have to show user page 27 after page 26
Ivan: Is this the kind of document that you are looking for Daniel and Peter?
...apart from the business case
...what is the most efficient weapon you need?
Daniel: Anything that shows the vendors
...I know it's a balance
...that we need to extend the platform to meet these feature sets
Alan: describing use cases from reading systems
...that cannot be done in web platform
...will be useful to browser vendors
Peter: I can bring a demo, but I have to bring a PDF instead of CSS
...if you can bring a demo in CSS, you can do it
...bring us things you cannot do in CSS yet
...that will upset the browser mevendors
...that's what we need to do; get people angry that they are not using their stuff; needs are not being met
Ivan: Brady is talking about
...one step further
<bkardell_> Is there a link for page stuff just discussed?
...you have a group that wants to get away from PDF
...and they hit these brick walls
...The two together
...the good examples in PFD and why you cannot implement yet
s/PDF
Tzviya: I think the doc we have now is use cases from publisher's perspective
...I want to define widows and orphans that funciton well
...and Brady is saying I need to tell reader when she transitions from page 26 to 27
...and Peter we need to know what you need to tell the reader
...both of those?
Tim: will they work together well is two separate documents
Ivan: Latin req is separate
...PDF examples
...and the page req
...whether this is one or two docs
...but as a block, it's an adjacent document to Latin req for me
Tim: if we satisfy page req and not Latin req too bad?
Ivan: no, it's not a page related problem
Tim: header and footer are more page oriented
Ivan: Latin req is not related to paging
Dave: Even paging stuff related to Latin req is to implement in OWP
...is far beyond the basic pagination we are talking about
...would involve going back up the tree
...to fix widows and orphans
...I don't expect that any time soon from OWP
...wondering what more we need to describe fragments
Markus: In Latin req?
...so there are two questions
...How do we do the page req, with a different perspective
scribe: second are there things from publishing perspective missing from Latin req still
...Should we take second question first
...So nothing about spreads, or very limited
scribe: I think that is the top-level abstraction to describe
...Chapter five goes into widows and orphans
...but the top most abstraction of page, or collection of pages
...that has a logical arrangement
Dave: A definition of what spread is
...and some of the constraints of how complex content interacts
...there will be replaced elements that span the spread
...like a picture that spans across the spread
Markus: yes, like a couple of screen shots
Liam: and a table that runs across a spread
Markus: So spreads is one thing
...transitions
...should we go into that?
...Comics industry is hypnotized by need for page transitions
...Is it in scope for Latin req; yes or no?
Dave: Daniel was talking about these transitions that allow media and transition effects
...transition ideas in Latin req
scribe: more about borders and fragmentation of boxes
...what happens to backgrounds when you have side bars from page to page
...omit top quarter on second fragment to indicate continuity of object as it is fragmented
...At least Latin req documenting a transition
...no craft document to refer to
Markus: I don't think we should go there right now
Ivan: And it's not
...I don't see why this would be Latin language specific
Dave; Right; I'm not going to describe fire consuming one page
Markus: Dave, are you willing to take the action about spreads behavior and spread arrangements progression
<scribe> ACTION: Dave to add new section to Latin req that describes spread behaviors [recorded in http://www.w3.org/2014/10/30-dpub-minutes.html#action01]
<trackbot> Created ACTION-28 - Add new section to latin req that describes spread behaviors [on Dave Cramer - due 2014-11-06].
Dave: The behavior of elements that cross the gutter
...use the terminology
Tim: only spreads from the classic print book?
...print analogs or also the scroll examples, spanning three pages for example?
Dave: In some sense Latin req is documented print
...given the scope is already nearly infinite
...sending scope to @
....is probably not something...
...fold-outs and documenting idea
...that page orientation may change
Liam: fold-outs give more than a single spread
Dave: spread may be two or more pages
Markus: Like widows and orphans observation; need to focus on the fundamentals
...we can do the funky things some other year
Dave: Page orientation is not fixed is good to know at this stage
...so you don't assume something that is no longer going to be true in other situations
...not hard code it into your thinking
Peter: especially items like that
...implementations are easier when you assume page size won't change
...variations and spreads need to be taken into account early on or it will be harder to do later
Dave: Diversity in print
...is hint of what will happen
...address digital use cases now
...even if from a source with less flexibility
Markus: Another use case I heard from a poetry publisher
...they wanted overflow for poem
Dave: In length or width?
Tzviya: Can go either way
Brady: depends upon language
Dave: could happen in both directions
Markus: publisher perspective in Latin req; anything else?
Peter: Did you cover bleeds?
...crop marks/
...when you circulate
Dave: CSS page
Peter: has been in CSS for a while
Dave: It got discussed
Markus: but fact that it's implemented does not mean we should not document it
Dave: right, but I don't want to go into the details for crop and registration marks and bars
Peter: a general note that there are printer marks that are not considered overflow
...we don't want to trigger overflow mechanisms
Dave: right
...elements that should not trigger overflow
<scribe> ACTION: Cover bleeds [recorded in http://www.w3.org/2014/10/30-dpub-minutes.html#action02]
<trackbot> Error finding 'Cover'. You can review and register nicknames at <http://www.w3.org/dpub/IG/track/users>.
Tzviya: STEM content in general
...a line in equations
scribe: we don't go into detail for styling or math
Markus: pagination or line formatting
Dave: Would like to have in Latin req in future
...not in pagination right now
Tzviya: breaking reflow?
Liam: when you set mathematics, you align all the equal signs vertically
...some publishers align
Tzviya: So we do talk about alignment
Markus: But not an immediate priority
...specific to pagination
Dave: right
...spread is specific
@: Do we need a border line in the page function?
...to express this is the right or left hand page?
...now it's empty
...some publishers ask us to put the border in
Markus: What do we call this?
Dave: page gap, gutter
Brady: gutter
Markus: that cam up in epub fixed layout
Brady: there should be a gap
Dave: often we have an infinite array of margins and boxes
Markus: Any more on Latin req from publisher perspective?
...David has his marching orders
...Would anyone like to sign up to help as an editor or co-editor
Dave: Fame and fortune awaits
Brady: what are you looking for?
...co-editor of Latin req
Markus: Let's move on to page req
...Brady, you were assigned the action item?
Brady: yes, I was assigned it
...I said I would do it
...I have not yet done it
...I can attempt to do it
...unless someone else wants to write it
...I have had a busy two months due to time
...might be able to get to it soonish
Markus: Anybody who would like to sign up
as a co-editor for page req?
... I'll sign up
Liam: I cannot be primary editor but will help with process and mechanical reviews
...I have experience
Bill: you'll be a contributor
Liam: yes
<scribe> ACTION: Brady to start authoring page req doc with assistance from Markus and Liam [recorded in http://www.w3.org/2014/10/30-dpub-minutes.html#action03]
<trackbot> Created ACTION-29 - Start authoring page req doc with assistance from markus and liam [on Brady Duga - due 2014-11-06].
Markus: Is scope of page req clear to you?
...should we document in the minutes?
...would you like to give a try?
Brady: It's the requirements of a reading system implementation for access to certain page and page level elements
...so you can actually create a reading system
...and get everything styled in Latin req
...elements on pages, initiate transitions
...in some sense turning the use cases into a requirements document
Markus: According to the schedule we have four minutes left
bkardell: First time I have heard of page req
...is there an existing page?
Markus: Latin req exist, but not page req
bkardell: ok, just curious
...I would like to see notes about it
Dave: What I am describing now
...is some use cases
Brian Kardell: Could you go through with me?
Dave: yes, I can do that
Ivan: for minutes
Brian Kardell, on CSS WG and JQuery Foundation
Ivan: Thanks, everyone
...while we are on action list
...we also said
...not sure if Brady will include it
<tzviya_> use cases for Page Req: https://www.w3.org/dpub/IG/wiki/UseCase_Directory#Pagination.2FDOM_interaction
...is to have some striking examples in PDF that we cannot do
...would be a good idea to have
Brady: what exactly you are looking at
Ivan: Something Tzviya said, some cases that she has that she can only do in PDF as a publisher
Brady: That sounds like Latin req rather than page req
Dave: we're talking about a book store
Liam: Something that speaks to CSS WG well
...I have noticed is places where you can almost get there from CSS
...so you give an incomplete example
Dave: Like drop caps; could make bad ones but not good ones
Liam: so then CSS usually says we should fix that
Tzviya: like books in graphic design
Bill: focus on main cases not fringe cases
Liam: Group could say it out of scope; give them something simple
...we cannot paginate our own anniversary document
<scribe> ACTION: Tzviya to provide PDF examples of paginated content that cannot be done online [recorded in http://www.w3.org/2014/10/30-dpub-minutes.html#action04]
<trackbot> Created ACTION-30 - Provide pdf examples of paginated content that cannot be done online [on Tzviya Siegman - due 2014-11-06].
Ivan: And what should we do with the business case kind of document
...The three of us plus Karen will have to talk to people
<scribe> ACTION: Karen to initiate bus use case [recorded in http://www.w3.org/2014/10/30-dpub-minutes.html#action05]
<trackbot> Created ACTION-31 - Initiate bus use case [on Karen Myers - due 2014-11-06].
Markus: Logistics
...Tzviya, you have arranged additional meetings
Tzviya: I have updated the wiki
<tzviya_> https://www.w3.org/dpub/IG/wiki/TPAC2014-F2F#Mid-day_break_with_ad-hoc_meetings_.28plus_lunch.29
...we have a meeting with Accessible Authoring Group tomorrow at 1pm
...we'll re-arrange schedule slightly tomorrow
...1:00pm Accessible Authoring tools on 31 October
scribe: everyone is welcome
...lunch will be 12-1
Tzviya: Today we convene at 2pm to meet with Protocols and Formats WG
Deborah: So nothing scheduled during break?
Tzviya: Correct
Markus: Any final comments on pagination topic?
...Enjoy lunch and see you at 2pm
<ivan> Chair: Tzviya
<glazou> ROFL
[people introduce one another]
<mhakkinen> Mark Hakkinen - ETS
<marcjohlic> Marc Johlic - IBM / WCAG
tzviya: [gives brief intro to concept of more detailed semantics in HTML markup with role attributes]
[projecting: http://w3c.github.io/dpub-content-and-markup/ ]
Janina: [gives example of knowing where a footnote begins and ends, and how to go back to where yu came from even if a footnote is referenced from multiple places in content]
tzviya: thanks, associated behaviours is important to mention
Rob: are the roles associated only with section element?
tzviya: it's up to the author to associate inflection with an element, sections are often nested, so maybe we do need to define an endpoint
[questions about the TF charter]
Rob: when do we get started?
Janina: today is good!
Mark: QTI is question/test
interoperability language
... used by Pearson and others have thousands and thousands of questions
and answers in this XML format.
We want to be able to take some of the semantics we have in QTA into HTML
scribe: to help in understanding the test tructure, as roles
There may be some specific ARIA properties we'd like to look at bringing over too. We have the Accessible Portable Item Profile task force,
APIP is a layer on top of QTA to define additional a11y.
scribe: Our TF is looking at long term goal of harmonizing QTA with HTML5 so yuo can author directly into your test item, makes transform more straight forward.
We have brought ARIA into QTI.
scribe: Also we have included speech
pronunciation SSML, CSS speech, PLS lexicon spec from EPUB
... But there's not yet uptake from user agents or assistive
technologies.
Janina: so let's just bring it up. Longsdesc is just about done for HTML 5 but it's not enough
Mark: When you have complex images, e.g.
the water cycle, in text books, we need a way to author the alternate
versions of that image,
... e.g. a simplified image, an embossed tactile file, a 3d image, need
to capture this with the image, as well as long image description.
... E.g. the diagram content model.
<azaroth> Not to derail the conversation, but accessibility contributions could be made via Annotations
<tzviya> +1
Janina: If you (HTML) opposed longdesc, help us build a better mousetrap.
Many people have not seen good image descriptions, only bad ones.
Mark: delivering a test with image descriptions, we may restrict which versions the student can get at test time, but more flexible for studying.
We have WGBH-trained writers of image descriptions,
scribe: but we're stuck not having mechanisms for delivering them.
Cynthia: Our last meeting was with SVG WG; water cycle good e.g. of an image you'd want to have explorable, ave you considered that?
Mark: yes.
tzviya: someone else noted annotations may
be a good way forward too.
... Our immediate conversation was to have a module for a vocabulary.
<azaroth> +1 to focusing on the topic at hand
<azaroth> ack +1
<Zakim> +1, you wanted to focusing on the topic at
Rich: is there anyone objecting to opening up role/value beyond what was specified beyond accessibility?
open up the validator.
Janina: you mean a limited vocabulary?
Markus: it could be unbounded
[discussion, how restricted should the list be?]
Zen: a list, or could I jjust use chapter as a role?
Rich: yuo could use chapter, and if the browser didn't know what it was, it'd fall back to its default.
Mark: who would object - what about with the AT companies?
They're going to have to do work from their perspective.
Rich: I had a call with one AT vendor, need something I can share.
Mark: when we apply a role, the role isn't necessarily human-readable, might not be announced by the AT, so how to attach a label?
Rich: we're talking about a localized role for ARIA 1.1, still having those discussions
Mark: lots of whiteboard discussions
E.g. might be chapter but in the book Canto
Jon: but what will a validator say?
Rich: we've got an aria reader we've expanded to read epub, but for books it'd be helpful to know, there should bea real definition
tzviya: some things are documenting definitions and mapping for AT. If there's a typo it's not valid.
So we don't plan just to throw this out in the wild.
<Zakim> ShaneM, you wanted to ask about providing role descriptions
Shane: On the validator.... can do
whatever it wants, and HTML 5 community alerady has certain things that
have a community-expandable
... list of values, and the validator still deals with.
But, role descriptions, being able to assign a description to an arbitrary role, not per-element but across a document, that's SEXY
Markus: instead of repeating aria-label=canto throughout the document you put it once in the head.
Mark: in NISO we used to have an external resource file.
<tzviya> liam: supports idea od extensible list, but open list could be dangerous because people may use terms in unintended ways
<tzviya> ...chapter can mean a chapter of chapter of a club
<cyns> +1
Rich: if we had an extensible list would that work for publishing?
<azaroth> +1 to modeling after link rel registry
Markus: like rel/rev registry? yes.
... Could be decentralised, different communities could have their own
and register as needed.
[discussion of vocabulary spaces, prefixes in epub not being used]
Ivan: I'd like to understand where we are.
What we say is it's a possibility to use the role attribute so that this
group or IDPF or wheover defines the terms for role for digital
publishing.
... Shane [tells us] this is no problem with the HTML validator, as
anyone can come and plug in these terms,
... or do we still have to negotiate? What are the roadblocks?
Shane: we'd ned to persuade whomever that a role registry needs to be established similar to the rel registry, for which there is already a precedent.
JohnF: it used to be a wiki page with HTML 5; there was no gatekeeper but not sure on status.
<azaroth> Are we not talking about http://www.iana.org/assignments/link-relations/link-relations.xhtml ?
tzviya: I think we want a gatekeeper as this has to be functional ARIA
Ivan: why does HTML have to change? There are a limited number of values.
Shane: HTML defers to ARIA, the list isn't in the HTML spec.
Markus: they have a table of restrictions on where ARIA attribtes are allowed
Rich: You put a checkbox and it says true but aria says false, that's an example HTML restriction.
Ivan: is this a roadblock for us?
Rich: no.
Ivan: what changes in ARIA?
Markus: has to point to registry
Rich: you're going to create a taxonomy and when it's ready you add it to the registry.
Shane: you register your terms, say where the description is, and if that were done in how the roles relate to the existin ARIA taxomony the AT devs would know what to do
Markus: if your plan is to stay with OWL then makes sense to plug in to that framework.
Rich: we'll probably not publish OWL in the ARIA spec, but it's helpful to us as desighers of the taxonomy
MikeCooper: want to offer possibility of
ARIA modules to extend the taxonomy,
... so a community like epub would do an extension. Reason to submit to
PF WG is to prevent collisions.
<Zakim> MichaelC_, you wanted to mention aria module module
Ivan: what out of this has to go through a Rec track?
Rich: we're going to make a module for the book semantics
Ivan: the beauty of a registry is that if
there's some sort of a community which controls the content, e.g. IDPF
as an organization,
... then why should that go through the painful exercise of W3C
Recommendation?
Janina: if it's an IDPF validator I suppose not.
Rich: you wanted this to work with AT vendors. Each platform has its own services layer. Have you looked at any of that?
Ivan: publishers I think need these terms
yesterday. SO publish list, go through the registry and validator,
... and a 2nd step could be a Rec. This IG has a charter for only one
more year.
Rich: how do we know yu're not going to
have collisions? Two roles that mean the same thing.
... We're suggesting the subteam would work to prevent duplication.
Janina: it's a clash between their list
and a list we have elsewhere [that's the problem]
... how do you keep it from clashing over tme.
Deborah: I think Ivan's concern is W3C's
process is slow and top-heavy.
... You don't necessarily need to go through the W3C approval process.
tzviya: we're not suggesting giving you our existing terms, we'd like help
Ivan: can't do a Rec in 12 months
Liam: fastest I know of is 2 years.
Shane: no need for a rec track.
Markus: suggestion is that aria role vocab changes from being a table of stuff inside a rec-track doc to being an external online resource that can be appended to
PF is the caretaker of this registry. Others come along and make submissions.
You have a set of requirements to make a submission, e.g OWL ontology.
scribe: IDPF, ETS etc would come to you with this more agile framework.
Rich: I don't have an issue with this.
Shane: that's what I propose.
<mhakkinen> +1 to Markus
Ivan: +1
Janina: [ +1 ]
Markus: to ARIA would change to take out the table?
Rich: If you start going into states and properties, that gets really messy
MichaelCooper: if this change is accepted it's not a hard editorial change for us.
Liam: ARIA could leave values in the spec and say that the registry is additional, and pre-populate the reistry.
<ShaneM> See the current "registry" at http://www.w3.org/1999/xhtml/vocab
Ryladog: so this is a recommendation to make a non-normative resource taht could be part of ARIA? Sort of like techniques doc for WCAG:
Rich: it'd be a registry, a glossary of values with definitions, and that'd be a normative definition in terms of the role value
Ryladog: could be continuallyupdated or improved
but PF would be responsible for non-collisions.
tzviya: [shows IDPF glossary work]
<tzviya> http://www.idpf.org/epub/vocab/structure/
http://www.idpf.org/epub/vocab/structure/#h_glossaries
[ tzviya projects http://www.w3.org/1999/xhtml/vocab ]
Markus: intent with glossary from idpf is that if it sits on a dl element the scope of the glossary is the dl element only.
[ questions about APIs ]
Ryladog: governments would need something normative for API stuff
Ivan: does an IDPF document serve?
Ryladog: if it's a recognised standards body, depends on the government.
Rich: I tell people this about standards: often the technology is the easy part; you have tooling, policy,...
[ PF WG, quality control questions ]
<scribe> ACTION: ShaneM to create a registry for ARIA roles in HTML 5 [recorded in http://www.w3.org/2014/10/30-dpub-minutes.html#action06]
<trackbot> Error finding 'ShaneM'. You can review and register nicknames at <http://www.w3.org/dpub/IG/track/users>.
tzviya: this needs to be backwards compatible.
Can we change these descriptions?
Shane: [yes, can change]
Rick: So the PF WG is going to work on processes for the registry.
Charles: There might be a dpub chapter vs. an svg chapter definition
[others: this is why we need a gatekeeper ]
Bill: are there prefixes in this context?
[discussion of why prefixes/namespaces are too annoying]
Mark: we'd participate to align.
tzviya: we'd do this in the joint dpub pf tf.
Deborah: i'll help
Rich: who will lead that tf?
Markus: I've been fishing for some people
Janina: we don't have to decide this today. We do have the W3C infrastructure to use.
Markus: tzviya or i will lead if necessary, we can't wait any longer.
Janina: get people signed up.
Ivan: can be done on mailing list
Janina: we're not going to get any kickback from dong this
Shane: thing to say @ HTML is we really like the precedent you set with the rel attribute!
tzviya: Markus, Charles, Tzviya, Mark, Deborah will be in the TF initially
Rich: we're looking at epub books at IBM.
Is there anything you'd like to see adopted when people write books; you
might want to have test tools,
... I should see the following semantics in a book.
<tzviya> ACTION: tzviya and company to review EPUB 3 SSV and pare down list for submission to PF [recorded in http://www.w3.org/2014/10/30-dpub-minutes.html#action07]
<trackbot> Created ACTION-32 - And company to review epub 3 ssv and pare down list for submission to pf [on Tzviya Siegman - due 2014-11-06].
Rich: We (IBM) store all our docs in DITA format, so we'll need [validation]
<tzviya> ACTION ivan help create OWL for list of terms created by TF
<trackbot> Created ACTION-33 - Help create owl for list of terms created by tf [on Ivan Herman - due 2014-11-06].
Janina: all these definitions we're going to, don't think you want to restrict them to AT, you can take advantage of them in general.
Ivan: they may already be used e.g. for educational books for reacting on educational level.
Janina: some resistance about ARIA is that it only helps people using AT, but there's nothing saying others can't use it.
s/martus/Charles/
[break, joint part of the meeting adjourned]
<ivan_> scribenick: bjdmeest
mgylling: at Bib, ivan and I had a talk
mgylling: which was a first public outing
about standards and digital publishing
... ivan and I are working on a (private) paper, still tweaking
... this paper is a elongated version of the BiB talk
... it is a vision about DPUB on OWP
... the notion is: reduce distinction between publishing and consumption
... portable document (pdf, epub, etc) today is a second class citizen
on the web, we want this to come to an end
... should become first-=class citizens
... the paper [which is being projected] talks about this
... seperation offline and online should become zero (i.e., a dynamic
state)
... so you can choose which state to publish in
... user can choose how to consume
... no compromise on feature set
... (work items list follows)
... we're not talking about one big cache
... (but it could be a partial state)
... purposes: archiving, publishing in oppresive regime, US marines
(selling complete offline ereaders)
... not about caching, about web content in a package that doesn't need
network
... chapter 3 is about work areas
... a.o., generic archive format
... there is a deliverable about packaging happening atm
ivan: difference of EPUB: based on multipart mime instead of zip
brady: been there, done that
ivan: there is a group working on that, we cannot reasonably expect a browser to implement multiple archiving formats, so we should talk about this now, before EPUB dies
brady: historically, there exists something like this, not adopted, as it was to complicated (multipart mime), however, as a standard, it is better (zip is a mess)
mgylling: 1: archive format
2: capture overall publication structure
3: document and fragment identification (select specific document, specific part of a document
ivan: final solution might not come from
IDPF of W3C
... i.e., what is the URL of a book, involves many many partners
... W3C and/or IDPF might not be solaly ideal
mgylling: 4: metadata for discovery
... harmonize metadata expressions
... inkling is doing stuff about RDFa and discoverability
... having a specific format, is completely out of question, needs to be
(plain) HTML
... 5: pagination
... is a dealbreaker if it doesn't exist (for some parts of the
publishing community)
6: security: origin is often unspecified
brady: for a local file, origin is a little strange
mgylling: 7: presentation: users that can
change the presentation is important
... epub is not ideal for this, very difficult for HTML and CSS
... 8: models for specific restrictions and extensions
... there is a need to restrain how documents are structured (i.e.,
comics, extreme amount of restrictions)
(btw, this list is not exhaustive)
scribe: we see this with EDUPUB in epub
... that are the 8 work items
... should be a living document
... idea: publish this paper, get input from diverse audience
... there are massive sectors that aren't here now
... could have wide repercussions
ivan: enhanced publisher: IBM going to
EPUB is an example of this of advantage of offline and online
... partners are e.g. IBM, they have massive problems
AH_Miller: people are using DITA now (oracle, ...
TimCole: worried about consistency, not features?
mgylling: yes, minimal required, what you put in there, may vary a lot
TimCole: there is a value in having a base, even if you don't use all
mgylling: yes, rudimentary stuff
AH_Miller: where would you put government?
ivan: inhouse publisher includes FAO, they
have a lot of publications
... they might advantage of this
Bill_kasdorf: The EU just joined, they
have a lot of publications, in a lot of languages
... it is really important that an offline publication is static, unless
it's online
... the content is fixed
... a website is fundamentally not
ivan: i can imagine that there a websites
that do not fit in this scheme (e.g., gmail)
... website authors might want to flag non-archivable content
Rob Cromwell: merging contexts: solve the tailoring for distribution ways?
mgylling: yes, part of the problem
... the overall vision: publishers use OWP directly, no second-class
citizenship
... we try to hit the points that we need to fix in OWP to let
publishing become a first-class citizen
... so we need an archiving format, online or offline is chosen by the
user
ivan: currently online publications cannot
be easily viewed offline
... e.g., view html, download pdf, no way back to html
... so, unified way, circle, could be easier
... same for archiving (how to archive major websites?)
... epub could be used for archiving
tzviya: some stuff is published in one
channel
... epub publications: word to XML to HTML5 (with prefix tagging) to PDF
... in reprint cycle, all files need to be corrected
... proper citation is a problem (page number?)
Rob Cromwell: so, there is a workflow, in this new world, you only have one master?
tzviya: well, we will still need to print
books, but there could be a citation system
... atm, that does not exist
... a correction mechanism can be easier (one source)
... there are a lot of things to do, (e.g., html to print), but it could
be solved by this
clapierre: is this a new standard, to
converge to?
... who would get these tools to convert current standards?
mgylling: that cannot be answered now, we
want to publish this white paper, toget feedback
... to kick off a real effort
... atm, just discussion, no timeline, no agreement to start working on
this
... atm, we don't know, depending on the threshold of the feature-list
of necessary features, it will take time
... multi-year effort
ivan: we hope that we can finalize the
document in november, and get it public
... then, a long cycle, with a lot of internal and external comments
... also, what can be done?
... (technically)
... hopefully: at the end of winter, we have a better idea of the
specific types of work of what needs to be done
... we may find some areas where there is existing work about e.g.,
archiving, where we need to participate
... other areas, we might need a specific group
... all conditional on the feedback
... sort of a very vague timeline
fettes: seems: archive of a very discrete
resource?
... else, you are fragmenting the web
... most websites are living, eg NY times, annotating, notes are gone
immediatly
... can be usefull for more static resources
mgylling: agreed
... generic does not mean: archive the web, it means: archive dig
publications
... would not be for front page of NY times
ivan: other direction is: currently,
cannot directly read an EPUB
... I should be able to do that automatically, without plugins, as if it
was a web page
... I should be able to adapt it
... I should be able to display it with pagination on
... web page is not a separate world from EPUB for digital publications
fettes: you might want pagination depending on e.g. viewport, landscape vs portrait
ivan: keep user control, there might be default setup
TimCole: we might need better terminology
than archive and offline
... detached vs attached
... detached from publisher, but online
... I may annotate partly online, partly offline
... NY Times: annotations to yesterday front page should not disappear
... maybe using archived webpage
... it is about packaging, so I can do something with it (online, you
can follow links, etc)
deborah: concerning citation: there is no
such thing as a fully static document, everything has versioning
... citations in paper world is very specific way of coping with it
... would this view have some kind of inherent versioning?
mgylling: yes, epub has that inherently in
metadata
... so, very rudimentary metadata, and maybe extra versioning
ivan: another possiblity: best practice is
that there are 2 ids: ID for document, ID for version
... also done with w3c docs
... another approach: the memento protocol
... URI identifies time in the life of that document
... specs exist
mgylling: it is a solvable problem
Bill_Kasdorf: we are already down this
path
... from day one with EPUB3, there is a commitment with staying inline
with HTML5
... EDUPUB as well
<Zakim> liam, you wanted to note that there's no reason in principle not to be annotate a back-issue/frozen issue of NYT
Bill_Kasdorf: this can be viewed as the long term vision of this kind of stuff
liam: W3C can change the web going forward
... so we can do annotations to NY Times 3 weeks ago
... eg using redirects (same for ebooks), here is the archived version
deborah: you can make a comment on a line, and historic comments are available on lines that no longer exist
ivan: issues: for publishers to move from
EPUB3 to EPUBX is a challenge
... in fact, move from 2 to 3 has required a very big step, by keeping
in line with HTML5
... so that step has been done
... whether it is zipped or multipart mime, doesn't really matter
... [we're over the hump]
Bill_Kasdorf: this [scheme] will make things easier
ivan: for now, DPUBIG is the most sounding
board to move on
... we need a wide reach out
... so updates on this document will flow back often to DPUBIG
TimCole: perception is that HTML5 is insufficient, not all things are possible (eg, math)
mgylling: eg unable to handle STEM
... we should add an extra bullet point to render STEM (basic and
advanced)
iBooks: ordering of the content (stream read): if you want to be able to read things before downloading everything, is important (when you the zip-path)
ivan: reason for multipart-mime is exactly for streaming content
brady: I'll give you the old spec, a lot of work has been proposed
ivan: do that!
... publishing industry should be part of the move, _now_
<astearns> +1
liam: [something about TAG, XCite?] could
do random acces with stream content
... it is not a done deal, other technical solutions are possible
<liam> s/EXI (efficient XML interchange format)/EXI (efficient XML interchange format)
Alistair: if I want to store a PhD, i need digital signature
mgylling: EPUB allows that, but it is
certainly a use case
... eg use case for governments: they can make sure it is static
TimCole: some publishers timestamp and IPaddress stamp publications, to get provenance
tzviya: we call that social watermarking
mgylling: we are not talk about DRM
Bill_Kasdorf: refer to it as Rights
Management, not Digital Rights Managment
... DRM has become access preventing, so let's not talk about it
mgylling: there is ongoing stuff, we need to act on that (eg CSS, webapps) for this archiving format to work
takeshi: on discoverability, the context is not preserved, how is the metadata used for discoverability?
mgylling: packaging is just a state,
publisher can provide unpackaged metadata
... packaging is just a state
Alistair: what would preclude a search engine to index this packaged format? they do this already for eg PDF
mgylling: it could be done by EPUB, but the zip archive needs to be searched, etc, ...
Alistair: same for PDF, Word, so it can happen
TimCole: so looking online for an epub, a link could be used for you local copy of that publication
mgylling: yes, a digital publication
search could just be an HTTP request
... digital publication of this paper is upcoming
... slides of BiB are public
<ivan> BiB Slides: http://www.w3.org/2014/Talks/1023-SF-IH/
ivan: video will also become online
<tzviya> https://www.w3.org/dpub/IG/wiki/TPAC2014-F2F#Friday_31st_of_October
mgylling: first, STEM update by Peter
... next, metadata session, will there be resolutions? (@Bill_Kasdorf)
Bill_Kasdorf: yes, we will see which
things to focus on
... three directions
updates are in the wiki
Ann Basetti: there is a tool Talky.io, to use a webcam (using web RTC) to 'skype in'
<ivan> ADJURNED