alternatives to the concept of a page, Gutenberg press vs LCD screen

[Peter Flynn]

On 29/08/2019 10:12, Mike Marchywka wrote:
> All the dvi language and generally latex and typesetting seems stuck 
> on the idea of "setting" fixed pages that may as well be images like
> jpeg. HTML and I guess PDF can be more flexible in how they are viewed
I don't think many people use DVI any more: PDF has been the default for
many years. TeX *IS* a typesetting program, so it's not unreasonable to
expect it to set type in a fixed way — but PDF can always be read and
indexed and converted to plain text (so can DVI, but rarely done).

LaTeX can be converted to HTML (eg Pandoc) but that's going backwards:
the optimal workflow is to create XML or HTML, and then generate LaTeX
from that, which I *much* more reliable. But HTML does not have the
formatting scope of LaTeX, so it depends on your application.

> (this is for ignoring things like data entry by user, just cases 
> where a publisher wants to publish a specific document or work for
> a reader and options for the viewer to nagivate). Someone here 
> indicated that companies like this approach as they know what the
> document will look like when the viewer gets it

Yes, most publishers treat appearance as the single most important
thing, because they believe their reputation is based on appearance
rather than content. They are convinced that they need to control
appearance first, before anything else. They are wrong, of course, but
changing their minds is VERY DIFFICULT.

> although in the past anyway that may not have been the case with pdf
> IIRC.

I'm not sure I understand that. It has always been the case, back as far
as Gutenberg, all through the metal-type days, and through the
Postscript era into the days of PDF.

> Computer generated documents can be structured or at least
> repetitive and lengthy compared to attention span of audience. But
> mashing the logical structure to fit a fixed page layout seems less
> important when you don't need to literally etch the result in stone.

Who is mashing the logical structure to fit the page layout? I have
only seen this for ephemeral documents (eg 1–2 page brochures etc).
> Scrolling and paging sound like easy natural ways to navigate but if
> you have played with collapsible sections in html documents it
> sometimes seems easier to just remove a logical section rather than
> scroll or flip pages.

You mean to hide it from sight (not physically remove it completely)?
Yes, it's attractive to start with. But when you are reading a complex
document, having to reveal and hide sections is just as annoying as
flipping pages or (worse) scrolling). The lack of pagemarks is a serious
problem: on paper you can use the little coloured sticky tabs for all
the places you want to refer to — it's technically possible in a PDF but
much less usable.

> I guess my question is how to "typeset" logical structure into the
> compiled latex output- dvi or pdf or even html.

I'm not sure I understand this either: LaTeX already has logical
structure markup (eg \section, \subsection etc), and this can be exposed
as links in a PDF using the hyperref package. This can be extended to
provide all kinds of other links.

> I was playing with my own "special" macro's and modifications to the 
> xdvik viewer to indicate logical sections and it seems like a nice 
> thing to provide in the output. Dealing with things like BOP/EOP was
> a bit of a problem I bypassed for now by using scroll size paper ( I
> think it was 8 x 100 inches lol ).

Nice idea. But I think there are lots of developments like this,
including the current vogue for 'endless' HTML pages (not really
endless, of course, just very very long) with smooth, accelerated
Javascript scrolling, and LOTS of cross-reference links

> Has anyone had a reason to typset a scroll or other unnatural page 
> formats?

[Be careful using 'scroll' because it means something else to developers
of reading systems for pre-book documents (scrolls).]

Not to typeset (my work is mainly with books and journals) but on the
web there is an experimental development for the CELT project
(http://research.ucc.ie/celt/) where documents are no longer served in
section-based chunks but as very long 'pages' with links and scrolling.
It's not finished yet, so not everything is working, and the metadata
still needs cleaning, but it does reflect the nature of early documents
which often had no visible logical structure because it was left to the
reader to work it out (even though the CELT transcriptions are taken
from Victorian printed editions). This is unnatural to us, but it was
very natural to the reader in the 6th–14th centuries.

Peter