Digital Publishing Interest Group Teleconference

28 Sep 2015


See also: IRC log


Dave Cramer, Brady Duga, Heather Flanagan, Ivan Herman, Deborah Kaplan, Markus Gylling, Peter Krautzberger, Paul Belfanti, Bert Bos, Bill Kasdorf, Karen Myers, Julie Morris, Tim Cole
Tzviya, Leonard, Vladimir, Alan, Jeff, Luc


<trackbot> Date: 28 September 2015


<pkra> O:-)

<scribe> scribenick: dauwhe

mgylling: let's get going

<mgylling> http://www.w3.org/2015/09/21-dpub-minutes.html

mgylling: first order of business is to approve minutes
... any objections?


scribe: minutes are approved
... we'd spend one of these telecons on the glossary (foreboding music plays in background)
... we want to get to a FPWD

Glossary, FPWD for the new version of EPUB+WEB

<mgylling> http://www.w3.org/dpub/IG/wiki/Glossary

scribe: the intent is it will be an IG note eventually
... Ivan, what's your view of where we're at?

<ivan> Editor's draft of the new EPUB+WEB document

ivan: So, where should we start?
... there were two reasons

<HeatherF> He fades in and out for me, too

ivan: why this issue on glossary was important
... one is what Markus said; we had to produce a new version of EPUB+WEB document
... since it will be a public note of the IG
... but there was also another issue
... this is more coming from the web community rather than the community here
... from webdevs and browser folk
... why do we need anything at all?
... why do publishers need anything else other than what the web already provides?
... we have been publishing web pages for 20 years!
... we have to be very clear in what we are talking about
... we are not talking about gmail pages
... difference between publishing and web at large
... so this led to the discussion I started a while ago
... and that's what's now in the glossary page
... I put up some initial things
... I naively put up some definitions for document / publication
... that led to a thread
... with more than a hundred emails on a specific topic
... talking about portable web things
... we did reach equilibrium when talking about portable web documents
... do we want to get into details?

mgylling: yes, we should get into details

ivan: the goal was to define a class of documents
... web pages or collection of web pages that work on and offline
... we had start with more fundamental things
... started with concept of web resource
... a very general thing
... a thing on the web that can be identified by a URI
... and can be accessed by various protocols
... we also needed something called essential content
... dkaplan was great in finding out these terms were defined in a11y
... part of web pages are really important for the message
... some parts are "nice to have" but less important
... example:
... font statements. In most cases it's OK if the user agent picks one of the fonts
... but if book is on typography the font might be essential
... the next step is we define a web document
... we don't concentrate on one single file
... but a set of resources that together is an identifiable thing
... like a scholarly article or a book
... one or more html pages, svg, js, etc
... it's that set which is the central notion
... has it's own URI
... that's one of the central concepts
... there are a number of qualifiers of what web doc should look like
... if the tech around us changes
... so we get poorer environment or smaller screen
... we should get graceful degradation but keep essential content
... there are smaller discussions between Bill K. and I around web document vs web publication

mgylling: there's an issue in the tracker

ivan: there are some minor formulation issues
... final point is that we define what it means to be portable
... because a web doc can be a collection of all kinds of resources
... I may have a dependency on some CSS file on the other side of the globe
... so there's an issue of portability
... the user agent can render the essential content of the document in such situations
... then there's a portable web document
... this is where we are carving out the things that are on the web
... there is an equilibrium point on these terms
... there are some details to be discussed
... before we get into the questions
... Deborah was a great participant in this discussion
... Leonard has sent regrets

dkaplan3: that was a good summary
... and of the points that are still pending

brady_duga: how is our definition of web document different from other definitions

ivan: which other definition are you referring to?

brady_duga: OK
... if you just google web document wikipedia points to an old w3c spec
... and it sounds a lot of what we already did

ivan: I think I remember dkaplan3 went through those

dkaplan3: the problem is that it's not the same
... we're dealing with 2 industries
... with two different meanings for documents
... publishing is diff. than web at large
... the terms we came up with are not the same as what w3c has
... document as a resource
... a way of defining a particular package of bits
... document has intention, some intellectual thing attached unlike w3c definition
... we are doing this because we are publishers who use English in this way
... the web uses the same words in different ways
... so our definition should be different than w3c definition

ivan: there is no widely accepted w3c definition
... they are not fundamental to w3c

mgylling: two questions
... web documents
... it strikes me that if you read it quickly it's not a very specific difference
... we should highlight identifiability
... you point it out, but only implicitly
... maybe one of the bullets should call that out
... "one URI" used to reference all the resources within the document
... that's quite explosive stuff
... there are 200 web pages that should be addressable by one URI

Bill_Kasdorf: so pleased to hear Brady and Deborah say those things
... I'm coming from publishing side
... from the web side
... whether or not there's an authoritative definition from w3c
... the word is used a lot
... when we talk about hierarchy of headings
... the top level section is an h1
... the next is h2
... and that's in your document
... when you have a book with 20 chapters
... and each chapter is a document
... or a magazine with 20 independent articles
... the h1 has to be the title of the book
... I say no, my document is my chapter or article
... now we have a problem
... i think the web is saying
... that s the document but not the publication

ivan: just to make it clear
... what you would prefer
... is that we talk about web publications as a collection of resources

Bill_Kasdorf: documents are parts of publications
... but a web pub could be a single document

ivan: I am ok with both
... we need order in how we make such a decision
... I haven't seen any major change
... after a hundred emails I don't just make a change on the fly

mgylling: we do have a quorum here today; we can barge ahead

<pkra> dauwhe: tend to think of these things as: we make publications and split them up for historic reasons (limitations)

<pkra> ... mostly technical limitations.

<pkra> ... e.g., html5 files

Bill_Kasdorf: I know lots of people who make an entire book with a single content document

<pkra> (took over scribing a bit)

Bill_Kasdorf: it's clearer with magazines, where they are independent

TimCole: with regards to bills last comment
... it's not just the way the content provider thing
... it's what needs to have identity
... like http URIs
... there may be a subsequent edition with a URI
... and there will people who will think of the two things together as a thing

<Bill_Kasdorf> +1 to the identity concept as critical

TimCole: we have to be careful

dkaplan3: searching for the perfect word is not going to work
... just make it clear we define our terms in our namespace
... we'll be dealing with people who use terms different
... web folk will read "webbily"
... publisher who does lots of anthologies will look at collections
... the author of one essay will define their contribution as a document
... we can't find a word that will work across all the horizontals
... any product of DPUB uses the terms as definedd in this glossary
... this is a namespace problem

Bill_Kasdorf: I agree with dkaplan3
... but if you agree we use it one way we can't use it the other way
... you invite confusion
... the word document creates so much confusion
... use publication as the aggregation

mgylling: I agree with you dkaplan3
... in my world document is very loaded already
... a very specific thing in html
... we are maybe inviting more confusion than necessary
... but we should be very clear it's in our namespace

ivan: if there is anyone who will mildly criticize me if I change to publication?

<TimCole> +1

<dkaplan3> +1

<pkra> +1

<ivan> +1

<HeatherF> +1

<mgylling> +1

<Bill_Kasdorf> +1 to publication

<brady_duga> +1

mgylling: IRC poll--plus one if you want to change to publication

<Julie_Morris> +1

mgylling: why did we argue for ten minutes if we all agree? ;)

<pkra> because Ivan is awesome.

ivan: can i move on?
... what I also did is that I took the epubweb document

<mgylling> http://w3c.github.io/dpub-pwd/

ivan: and I tried to use the new terminology
... we do want a FPWD
... it's no longer a white paper
... it would be helpful if you all could review this document and provide feedback and/or pull request
... having a stake in the ground is important
... the next term is not yet closed
... what is identifier
... do we mean URI
... do we mean something else
... how do we use these terms
... Bill_Kasdorf will love this topic
... let's not start now
... there is also the issue of semantics
... my background is in semantics
... maybe we want to have a clear idea of what we mean by semantics, metadata...
... even in world of metadata

<Bill_Kasdorf> Re "semantics," I think we should clearly make the distinction between "structural semantics" and "content semantics."

ivan: the pub world using the term reading system
... web world talks about browsers
... we should settle this for the future
... state of a portable web publication
... we talked about offline and online

<Bill_Kasdorf> Re reading system vs. browser, how about "renderer" or "rendering agent"?

ivan: we realize that there are two dimensions
... one is whether we talk about a set of web resources spread over filesystem or URI
... or whether they are packaged into one format/package
... for millions of reasons the notion of packageing is important
... for distribution etc
... the portable web is packaged or unpackaged... that's one dimension

<Bill_Kasdorf> e.g. a Braille rendering agent renders the publication in Braille (i.e., rendering doesn't need to mean visual, and we could define it as such in the glossary)

ivan: the other area is online vs offline and what it really means
... its not clear what we mean by offline and online
... could be http online vs no http offline but that's not true
... I could use localhost
... how do these things relate?
... if we use on/offline we run into issue
... I say what's important we HTTP-based protocols or filesystem protocols
... whether HTTP is local or not is less interesting

<Bill_Kasdorf> +1 to the distinction between http vs. file system access

ivan: I would ask brady_duga
... in both cases we had Leonard

brady_duga: why is it important to determine which protocol accesses things
... as a user the question is whether the pub is completely available
... how I access that doesn't matter
... you could have a remote publication
... I don't know what this distinction is important

ivan: my feeling is
... when it comes to user agent side
... the architecture of the user agent
... things like service workers
... when we think about identifiers
... that's where the difference comes in
... i agree the distinction is unimportant for reader
... but important in how things are realized

brady_duga: yes, for a browser that is true
... for a reading system
... there will be access issues
... they will have to do different things

ivan: that's what I concentrate on

mgylling: OK
... what are pain threshold for FPWD?
... how much do we want to settle before FPWD

ivan: two issues
... publication vs document, we solved that
... in the document I have added some additional wording on the kind of issue dkaplan3 alluded to, about context
... there is another aspect
... orthogonal to what we already did
... I tried to make the whole concept independent of EPUB
... the criticism was that it was describing EPUB4
... so I tried to make this weaker, although the dependance on EPUB is still there
... we don't want to provide a bad message to the industry
... there is a fine line there
... we don't know what Leonard wants to do
... that will happen this week
... that's probably the biggest remaining issue
... once this is done we can publish a FPWD

<Bill_Kasdorf> fyi in discussions on the magazine side and separately on the journals side there is a strong desire to include file types in a publication that EPUB currently doesn't allow

mgylling: for the other IG members, what kind of inputs would you want between now and FPWD?

ivan: for FPWD, general review is good
... nothing specific

mgylling: do we want to look at time here, given pub moratoriums around TPAC?

ivan: I would love if there was an agreement on the content at the meeting next week
... because it's a fpwd the IG must resolve to publish
... we'll see if it can be done by next Monday
... still OK if it goes to the next Monday

<pkra> non-benevolent?

mgylling: since we have a few minutes left
... I'm looking at wiki
... next-to-last bullet confuses me

<mgylling> “A Web Resource in a Web Document is Portable if an OWP compliant user agent can render its essential content by relying essentially on the Web Resources within the same Web Document”

mgylling: this line confuses me
... the trailing end... what does it mean

ivan: this means that when I produce a PWP
... all the things that are essential like CSS and JS are part of that portable publication
... an extreme case would be everything should be in the PWP, but things like fonts are perhaps not as essential

mgylling: ok
... can a web resource be "without" a web document rather than "within"

ivan: the web doc might refer to fonts
... are the font files themselves part of the web doc, or elsewhere on the web?
... it should be portable even if the fonts aren't with it
... but in some cases the font file must be part of the doc
... an epub example
... for some epubs
... a scientific book
... the data might have to be part of the book because I interact with it
... in other cases it's ok if the data is on the web

mgylling: of course
... so you have a portable web doc exploded online
... so what does it mean then

ivan: is it part of the set or not?
... everything that's necessary for that web document must be part of the set

Bill_Kasdorf: can some be offline or online?

ivan: then it's not portable

mgylling: I'll try to clarify in an email

All: thanks and bye


Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015/09/28 16:11:22 $