RE: [Glossary] Portable Digital Document's states

We seem to be crossing concepts again.

Are we perhaps getting too far into the weeds for a glossary?

These are issues we need to define as we work on POM (or whatever we choose to call it), but we are not writing a spec now. We are talking about what an “offline state” is, for example. We have already concluded that this is needed, even in a world of cloud readers (e.g. for archiving purposes).

Tzviya Siegman
Digital Book Standards & Capabilities Lead
Wiley
201-748-6884
tsiegman@wiley.com<mailto:tsiegman@wiley.com>

From: Leonard Rosenthol [mailto:lrosenth@adobe.com]
Sent: Thursday, September 24, 2015 1:31 PM
To: Ivan Herman
Cc: Brady Duga; W3C Digital Publishing IG; Ralph Swick; Bill Kasdorf
Subject: Re: [Glossary] Portable Digital Document's states

Why does it have to be the same UA/RS?     And is the UA only considered to be a “client-side” component or can it also incorporate “server-side” pieces as well?

In the current packaged publication world – the UA/RS (IMO) is well defined.  It’s the client-side piece of software that provides the UX to a reader.  But as we move towards the idea of “online” publications, the UA and/or RS live is blurring.

For example, consider Helicon’s recent announcement about their Cloud Reader (<http://www.prnewswire.com/news-releases/helicon-books-launches---epub-cloud-reading-system-528475951.html>).  Is the UA/RS the browser?  The server component that prepares the EPUB for processing? Something else?  Does it change if the user is using their mobile EPUB reader?

As I said before, I think the distinctions are relevant in context and not in generalities…

Leonard

From: Ivan Herman
Date: Thursday, September 24, 2015 at 1:13 PM
To: Leonard Rosenthol
Cc: "Brady com>", W3C Digital Publishing IG, Ralph Swick, Bill Kasdorf
Subject: Re: [Glossary] Portable Digital Document's states

We seem to be in a complete disconnect and, honestly, I am lost by this thread.

Can you guys give me an alternative how we want to describe the behavior of a user agent faced to a PWD being packaged vs when it is not? Or when it is online or not? Surely a user agent has to take these situations into account and we should probably find what features are invariant or not.

I would welcome a coherent set of definitions...

Ivan

---
Ivan Herman
Tel:+31 641044153
http://www.ivan-herman.net


(Written on mobile, sorry for brevity and misspellings...)



On 24 Sep 2015, at 18:02, Leonard Rosenthol <lrosenth@adobe.com<mailto:lrosenth@adobe.com>> wrote:
I agree with you, Brady.

My understanding of the previous discussions was around whether the UA/RS needed to get “essential content” from somewhere other than the current document…See the definition of Portable.

As such, it shouldn’t matter if the content is online of offline – only whether it is Portable or not.   You either have a Portable document (as per our definition) or you have a non-Portable document (aka a Web Document, as per the definitions).   Whether they are connected to a network or not, accessed via standard web protocols or not, should NOT matter.  Only whether it is Portable (as per our definition).

Leonard

From:  "Brady com>"
Date: Thursday, September 24, 2015 at 11:18 AM
To: Ivan Herman
Cc: Leonard Rosenthol, W3C Digital Publishing IG, Ralph Swick, Bill Kasdorf
Subject: Re: [Glossary] Portable Digital Document's states

I am still a little bothered by the definitions. So, if I have a document that references some resources via http, but I have a reader implemented in javascript that uses service workers to fulfill requests for certain resources and has downloaded an entire publication, making it available for offline use, is that document considered 'online' by these definitions? Is loading a resource whose scheme is http the same as using the http protocol? And if I use localhost (and hence the loopback network adapter) when I have no network available really online? Is the distinction that an authority/server is processing the requests? Is the service worker considered an authority/server?

On Thu, Sep 24, 2015 at 12:02 AM, Ivan Herman <ivan@w3.org<mailto:ivan@w3.org>> wrote:
I have put a set of definitions on the Glossary[1] trying to close this particular issue. As for the offline/online vs. protocol/file, I propose to use the offline/online differentiation, because that is the common way differentiating in practice, noting (also in the glossary text) that a resource access through http on the local disc may be considered as "online", too. Again, nothing is frozen, but by choosing an equilibrium point we can move on.

Ivan

[1] https://www.w3.org/dpub/IG/wiki/Glossary#States_of_a_Portable_Web_Document


On 21 Sep 2015, at 15:53 , Leonard Rosenthol <lrosenth@adobe.com<mailto:lrosenth@adobe.com>> wrote:

Because I believe the differentiation between online/offline is more important for some criteria (such as identification) while others are more relevant for file/protocol (such as signatures).

I agree that a UA that is working entirely from protocols doesn’t know local vs. remote. However, a UA (or more likely the RS on top of the UA) may wish to implement file-based processing.   So the method of access has a clear impact on identification and other things.

Leonard

From: Ivan Herman
Date: Monday, September 21, 2015 at 9:45 AM
To: Leonard Rosenthol
Cc: W3C Digital Publishing IG, "Brady com>", Ralph Swick, Bill Kasdorf
Subject: Re: [Glossary] Portable Digital Document's states


On 21 Sep 2015, at 15:30 , Leonard Rosenthol <lrosenth@adobe.com<mailto:lrosenth@adobe.com>> wrote:

I think it depends on what you are trying to “profile”…

Is it identification?  Security?  Signatures?   What??

I am not sure how this refers to the the choice between online/offline vs. file/protocol.



I also think that a UA should be free to choose which method it uses for working with local content.  It may choose to process it as a file OR it may have a local server and process it indirectly that way.  Both would/should be valid.


I do not think we said, at any time, that any of those are valid/invalid. But a UA, I presume, does not know whether a content accessed via HTTP(S) is local or not: it is all HTTP(S). Again, I am not sure how this affects the choice of the various states we want to consider

Ivan


Leonard

From: Ivan Herman
Date: Monday, September 21, 2015 at 8:15 AM
To: Leonard Rosenthol
Cc: W3C Digital Publishing IG, "Brady com>", Ralph Swick, Bill Kasdorf
Subject: Re: [Glossary] Portable Digital Document's states


On 21 Sep 2015, at 14:02 , Leonard Rosenthol <lrosenth@adobe.com<mailto:lrosenth@adobe.com>> wrote:

As you note, Ivan – there are subtleties involved in actually trying to spell out what each of the quadrants means.  Your example of an unpacked PWD accessed locally via web protocols is a good one, but it could just as easily have been a packaged PWD accessed that same way – in fact, I built one of those back in 2011 as part some work we were doing.

Oh yes, absolutely. I just did not want to overload the figure too much.

The question remains, though: should we stay by the offline/online differentiation or by the protocol/file access one. I am tempted to go for the second for very pragmatic reasons: any user agent would have to make a difference between file vs. protocol, versus they do not care (actually, they do not know!) versus an HTTP access stays local within a machine or not.



But more importantly, the two sets of dimensions are indeed key and for the reasons you state – one for the defining the packaging technology to be used for a PWD and the other for influencing the identification mechanism.

Also, now that we have defined a term – Portable Web Document (PWD) - how about using that in all instances where you might previously have written EPUB?   It took us a lot of time/effort to get to that term – let’s use it!


:-) Right.

Ivan


Leonard

From: Ivan Herman
Date: Monday, September 21, 2015 at 7:18 AM
To: W3C Digital Publishing IG
Cc: "Brady com>", Ralph Swick, Bill Kasdorf, Leonard Rosenthol
Subject: Re: [Glossary] Portable Digital Document's states

So… it seems that we do have two dimensions for 'states. Because Ascii art can go very bad with different clients, I created images as attachments. The first attached image shows is the offline/online vs. packed/unpacked situation with some typical usage scenarios.

However, offline/online may be an elusive notion. I would think (although I am not sure) that it may be more precise to differentiate along access methods: in practice, if we are in a Web setting, the difference may be whether the content is accessed via an HTTP(S) protocol, through Web protocols, or whether the content is accessed through a file system. So I created a second image doing that.  Note the difference between the two: if I have an unpacked set of files in a folder, the same content can be accessed via the file system or via a server running on my machine with the same content served through HTTP. Although, in both cases, the content is accessed offline, the access method is different.

Whichever four states we choose (and their definitions should be properly pinned down) the real question for this Interest Group is which of these states are of real interest (sic!) for the group and for the digital publishing community at large. And, in fact, I believe all four are. Indeed,

- the packed/unpacked dimension is (obviously) of interest for the details and requirements on packaging; ie, it influences the details on an architectural view for some sort of a unified approach for readers' core (our service worker based scenario)

- the offline/online or, alternatively, file or protocol access dimension raises questions on identifiers. Do we use generic identifiers overall, regardless of locations, how should the references within a Portable Web Document be organized to ensure a unified identification scheme. (I cc-d BillK explicitly, because I know these questions are of a real interest to him:-)

If we agree with these two dimensions, we can go forward and define them more thoroughly

Ivan

On 18 Sep 2015, at 18:21 , Leonard Rosenthol <lrosenth@adobe.com<mailto:lrosenth@adobe.com>> wrote:

Yes, I believe that caching is more an implementation issue than a format or design issue.

From: Ivan Herman
Date: Friday, September 18, 2015 at 9:07 AM
To: Leonard Rosenthol
Cc: "Brady com>", W3C Digital Publishing IG, Ralph Swick
Subject: Re: [Glossary] Portable Digital Document's states


On 18 Sep 2015, at 14:35 , Leonard Rosenthol <lrosenth@adobe.com<mailto:lrosenth@adobe.com>> wrote:

[Sorry for coming in late]

I agree with Brady – there is a huge difference between package/unpackaged and online/offline.  The original terms made more sense to me as they focused on whether the content was usable without standard web protocols (aka offline) or not.

Indeed. As I say below, we seem to have two different dimension for 'states' and not one dimension only. I can be unpacked and offline, and packed and online. So the question is whether defining two types of states (offline/online and packed/unpacked) is enough. Ie, whether treating cached separately on any dimension is necessary or not (my feeling is that it is not, it is more of an implementation issue).

Ivan



----
Ivan Herman, W3C
Digital Publishing Lead
Home: http://www.w3.org/People/Ivan/

mobile: +31-641044153<tel:%2B31-641044153>
ORCID ID: http://orcid.org/0000-0003-0782-2704



<online offline.jpg><file access protocol access.jpg>


----
Ivan Herman, W3C
Digital Publishing Lead
Home: http://www.w3.org/People/Ivan/

mobile: +31-641044153<tel:%2B31-641044153>
ORCID ID: http://orcid.org/0000-0003-0782-2704






----
Ivan Herman, W3C
Digital Publishing Lead
Home: http://www.w3.org/People/Ivan/

mobile: +31-641044153<tel:%2B31-641044153>
ORCID ID: http://orcid.org/0000-0003-0782-2704






----
Ivan Herman, W3C
Digital Publishing Lead
Home: http://www.w3.org/People/Ivan/

mobile: +31-641044153<tel:%2B31-641044153>
ORCID ID: http://orcid.org/0000-0003-0782-2704