SV_MEETING_TITLE -- 24 Jul 2017

<Avneesh> I am already on audio line. So, it does not look like Readium meeting

<toshiakikoike> presnt+

<tzviya> https://www.w3.org/2017/07/17-pwg-minutes.html

<mattg> Garth: Tzviya and I want to talk more about process - github issues versus the mailing list

Tone and Volume on GitHub and email

<mattg> ... we've had a number of people talking about the tone and volume on github issues - on one hand it looks exciting and invested with strong opinions but there have been comments of feeling intimidated or shouted down - we have some issues with civility and quantity of responses

<mattg> ... if we disagree with what people say we need to respond in a civil way - we don't need to respond to every post so that we encourage others to jump in

<tzviya> https://www.w3.org/Consortium/cepc/

<mattg> ... not sure exactly what the solution is but need for a well-functioning group - we want everyone to feel included and take these issues into consideration - link to w3c code of conduct is in the agenda - we have had some over-enthusiasm

<tzviya> https://www.w3.org/publishing/groups/publ-wg/WorkMode/#mailing-lists-policy-usage-etiquette-etc

<mattg> tzviya: if you take a look at the link there's a section on what we use github for and email for. github is for technical discussion - email is for overarching topics

<mattg> ... we've been having philosophical issues on github that belong on the email list - also need to be aware that consensus on github is not consensus of the WG

<mattg> garth: we want to move some of the issues that have become circular and deal with them on the call today

Technical work

<mattg> ... there is a link in the agenda to the active github issues - we've pulled out a subset to discuss

<tzviya> I also pulled together a google doc https://docs.google.com/document/d/1NzYLURq4zqkeGoCJbY-5rlIUsfPt9Dy5S10BLOA-tGg/edit?usp=sharing

<mattg> dauwhe: one of the sources of some of our discussions that haven't resolved is that there seems to be several different visions of what we're talking about when we talk about publishing a web publication

<mattg> ... one of the ways this has come out is does the person doing the publishing have control over the primary resources being published - if we get to a component can we discover it is part of a publication or part of more than one - example has been through a link

<MURATA> Yes.

<tzviya> s/murata/Garth

<laudrain> +1

<mattg> garth: is a publication a link to something that might be moving or more like a point in time - we decided on this in the interest group but not sure if we're still on the same page

<mattg> dave: the kind of content we publish has a time cycle measured in years - corrections may come in later but we hardly ever refresh more than once a year - it is very different from twitter or nyt that update constantly or on scale of minutes or hours

<mattg> ... is what we do related to anything that publishers do?

<mattg> hadrien: on the web you can update as often as you want and can link to whatever you want - basic principles of the web

<mattg> dauwhe: is there a difference between linking and publishing

<laudrain> yes

<bigbluehat> the W3C (for instance) prevents edits to published TR documents--groups must provide errata...even for trivial things like typos

<mattg> garth: but this is our mission to determine what publishing on the web means - this group is w3c's effort to make the web relevant to publishing - if it differs from the web that is relevant but not completely relevant

<tzviya> https://www.w3.org/TR/pwp/#whatisawebpublication

<mattg> hadrien: our goal should not be to add publishing but publications - identify the grouping of primary resources and metadata - should not be that we can't update or link

<mattg> tzviya: this is where the dpig use case discussions diverge - if we disagree with what dpig came up with we need to come to a new definition and work on that as a group

<mattg> hadrien: i don't believe it's that different

<clapierre> I see similarities here when companies do major web redesign to their website and the do this in a sandbox and then then "publish" their final website after they do all their final copy edits/ accessibility testing, link testing etc.

<mattg> tzviya: i don't have a url for anything until it is published at wiley

<MURATA> What is the current definition in the input PWP documents?

<bigbluehat> MURATA: see also https://www.w3.org/TR/pwp/#dfn-packaged-web-publication

<mattg> dauwhe: a web publication being just a collection of links what stops me from calling the top five ads from nyt a publication and putting ads in it - the act of publishing means that you had some responsibility in the creation of the content - more like playlists instead of albums

<MURATA> Got it. Thanks.

<cmaden2> Nothing prevents anyone from embedding others’ content for revenue now—except law (and suits) and technology.

<cmaden2> But if I license my content that way, you ought to be able to aggregate however you like.

<mattg> hadrien: I don't think that's true - the first paragraph has no impact - it's only when we get to pwp that it matters - at wp level it's not existent - a mix tape is a kind of creation

<tzviya> scribenick: tzviya

tim: The IG was trying to take the idea of existing publishing and make sure it can work on the web
... not look at what on the Web can be called publishing
... We do have to look at both perspectives

<mattg> george: i'm viewing this from a semantics perspective - when i'm in a web publication that means something to me and my expectations are different from a web page

<scribe> scribenick: mattg

UNKNOWN_SPEAKER: want to be able to read in a particular order, search the publication instead of the whole web

<MURATA> +1

<clapierre> +1 to George's comments

garth: if we look at epub today it is a collection that will be the same tomorrow - the basic content is not changing

<cmaden2> Also, the New York Times for Mon 24 Jul 2017 is nothing other than an ordered collection of New York Times articles. Might they want to be able to use PWP to publish that?

garth: that is key to what a web publication is - except for certain dynamic content things in the spine won't change
... if we do something that is not relevant to our publishing community we have failed

tzviya: the first question we have to answer is where we differ from the dpig document
... publishers are not updating on a regular basis like an atom feed

<Rachel> +1 tzviya

<MURATA> Suppose that a music magazine has a list of future concerts. Will that list change if some concerts are cancelled?

tzviya: first we need to define publishing on the web - we don't need to define a new version of atom or rss

<garth> +1 Tzviya

tzviya: also need to solve cross-origin - there are reasons to address and not to address - we can't solve every issue for fpwd

laurentlemeur: it's not about capacity to search or update but about control of the documents that are part of the publication
... do we want publications that contain documents that we don't control - for some of us publishing involves control of the documents - others aren't concerned about control but about ability to create the overarching publication

garth: not sure how we decide on this to move ahead - finding the thing that's being pointed out requires stability

Hadrien: it's not even an issue of control - content can be spread across multiple domains - an article could be included in multiple publications, for example - remixing is useful even when you have complete control

<garth> Publishing: defining the parts that won’t change over time.

dauwhe: based on what hadrien just said, there are some significant trade-offs - one thing we see as desireable is if you find yourself in a web primary resource you have a way of finding you're in a publication but that closes off some of these possibilities

<Hadrien> +1 to what Brady said

duga: being able to identify that a primary resource is a part of a publication is powerful and so is including other primary resources - can handle this without links in the document - can be done with magic formatting of the link or http headers - we could find some mechanism to get back to the parent without modifying the html

bigbluehat: one of the things i ran into was whether portable web publications provide offline-ability - the boundary line is increasingly a boundary of what can be published - pwp allows distribution but a wp can be loaded in firefox and have the rest of the book

Hadrien: from a technical point of view I don't think we're making that big a distinction - discovery is the main thing - providing a link is only generally a "should" - if you have ten links it might even be messy - if we don't require links then you can have both views

duga: discoverability shouldn't be a requirement but the ability to make something discoverable should - we have to allow the publisher to make the publication discoverable from a shared primary resource

Hadrien: agree. the impact on a technical basis is small so not sure why we're making such a big deal out of it

dauwhe: sounding to me like we're in a place where a little prototyping and experimenting would be helpful - especially around brady's ideas - interested in how nesting will work - some setches of these things would help

Hadrien: all of the things we're talking about we're implementing in readium2 - not just conceptual and intending to use in production

garth: the fact that an implementation exists does not mean it is the thing this group will adopt - readium moving these forward is relevant but not a determination

<tzviya> https://www.w3.org/TR/pwp/#whatisawebpublication

<MURATA> Agree with Garth.

tzviya: we're still working with a fuzzy definition of web publication - dpig put together a definition in the link - I'll send around an email vote later today on whether to use this or refine it as we may have differing perspectives

garth: we need to agree on a definition

<MURATA> Hadrien, why?

laurentlemeur: looking at the definition of web publication in that document - what does it mean about obtaining primary resources

tzviya: we'll welcome suggestions to improve

garth: we need to try and get to consensus on the definition

dauwhe: we've already been talking about whether there are solutions that don't require an explicit link to a manifest
... when I go to a link finding a link that indicates that this belongs to a bigger whole is important

<garth> Can a UA with the first resource of a WP figure out that it's a WP (and which one)? YES

<garth> Can a UA with any resource of a WP figure out that it's a WP (and which one)? YES

murata: I would like to replace resource by primary resource - some resources are not primary - primary are like spine items

<dauwhe> +1

<tzviya> +1

<garth> Can a UA with the first primary resource of a WP figure out that it's a WP (and which one)? YES

<garth> Can a UA with any primary resource of a WP figure out that it's a WP (and which one)? YES

<bigbluehat> +1

<tzviya> +1

<laudrain> +1

garth: sounds like these two statements are non-controversial

<Hadrien> it can discover it's part of a WP if it has a link to it

<Hadrien> so that's not a +1, more like a +0.5

- DRAFT -

SV_MEETING_TITLE

24 Jul 2017

Attendees

Contents

Tone and Volume on GitHub and email

Technical work

Summary of Action Items

Summary of Resolutions

Scribe.perl diagnostic output