W3C home > Mailing lists > Public > public-digipub@w3.org > August 2015

Re: CSS Pages and Pagination (was: Prioritisation)

From: Johannes Wilm <johanneswilm@vivliostyle.com>
Date: Wed, 5 Aug 2015 23:24:50 +0200
Message-ID: <CABkgm-Rcx+1_Sx1fU232Wz_gw88OO4VrL24h0ZrCHy5Zgsy3Og@mail.gmail.com>
To: Daniel Glazman <daniel.glazman@disruptive-innovations.com>
Cc: W3C Digital Publishing Discussion list <public-digipub@w3.org>
On Wed, Aug 5, 2015 at 8:40 PM, Daniel Glazman <
daniel.glazman@disruptive-innovations.com> wrote:

> On 05/08/2015 18:13, Deborah Kaplan wrote:
>
> > This discussion is all quite interesting, and a good one to have, but I
> > just want to clarify -- the way pages and references are being discussed
> > here are more relevant to the meaning of "pages" which is not part of
> > the CSS prioritization document under the original discussion. Unless I
> > am misunderstanding again! In which case maybe we should change the
> > subject line. :-)
>
> Yes. Done. And that discussion would be better held in www-style than
> here, IMHO...
>
> While I'm here, let me give my PoV on Pages and CSS, please.
>
> Whatever is a page, CSS are going to be used to count them. So
> the insertion of a reference to a given page - or to a page
> containing a given text or a given node - will be done at least
> partly through CSS.


Yes and no. In an ideal world, and hopefully the world as it is in 5 years,
where CSS has been defined for everything page related, this may be the
case.

In a non-ideal world, where one has to draw pages manually using a
combination of JavaScript and CSS in a browser to give the impression of
pages to the end user (and even to produce PDFs of those pages), the
meaning of CSS is of course still very central. If the CSS can be specific
enough to ensure that a page of height X pixels and width Y pixels with
text of font Z will ensure that the exact same number of letters will go on
each page in every one of the implementations, that would be an important
way to ensure page number consistency.

However, AFAIK there is still the issue that counters defined in CSS are
unreadable to JavaScript. So basically if one puts the page numbers on each
virtual "page" using CSS, but one needs to use JavaScript for some other
part of the entire package (for example for a cross referencing mechanism
or a word index mechanism) because these parts haven't been defined through
CSS entirely yet, one has the problem that one cannot read the page numbers
using JavaScript. That is why page numbers in most such solutions are still
added using JavaScript.



> Since the whole thing smells, looks and
> feels like generated content, the natural host for this is
> probably CSS, even if I agree we're slightly leaving the field of
> pure styling to reach higher-level features.
>
> I'm all in favor of it. We have been sitting on these features
> for far too long and I prefer having them in CSS even they not
> entirely belong there rather than arguing forever again about where it
> should land.
>

I agree that eventually they should end up in CSS there and of course if
one wants to get there eventually, one needs to work on this for a while.


> FYI, Dave Cramer is working on all the generated content's
> side. The page notion itself is contained in a CSS module that is
> far too old. There are many reasons for that, often quite
> bad. And I'm not even sure our notion of page contained there is
> not obsolete, deriving from concepts that are almost twenty years
> old. I repeat: obsolete. The model in CSS Paged Media is in my
> opinion something that is to what we really need what was CSS 1
> to the current Web pages we browse on a daily basis; a good beginning
> but a too limited one for 2015. I choke every time I see the 16
> margin boxes specified in CSS Paged Media and the associated magic...
>
> There is some resistance to changes there, again often for quite
> bad reasons. With other concepts like Templates, Regions, Shapes
> and Exclusions we could certainly specify much more powerful - and even
> simpler - page models. We need implementors for that, and that's where
> you all can help. The CSS WG is where it should happen.
>

While it would be great to eventually have pages both specified in an
adequate manner and have implementations follow such a specification, it's
probably also important to see what other CSS features that are more
popular than pages can be reused as pages. CSS Regions worked great as
something to base virtual pages on. Columns can also be reworked in
combination with JavaScript to create something that appears like pages to
the end user. There are likely others features that could be reused in a
similar manner that we haven't thought of.

But this should be no news to most/all on this list, as I am guessing that
is the precise reason why CSS Regions are of interest to people here.
Received on Wednesday, 5 August 2015 21:25:26 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 August 2015 21:25:27 UTC