RE: [xml-dev] Formatting Processing Instructions

[Amelia A Lewis <amyzing@talsever.com>]

Disclaimer first: not a wizard, don't play one on television.  My 
colleagues think I know something about XML, but my SGML knowledge is 
mostly limited to Goldfarb's book (and adapting SGML to XML workflows, 
but that's always custom work).

Have some questions, though.

On Wed, 22 Feb 2012 18:19:00 -0600, Len Bullard wrote:
> So some require PIs to support
> 
> o  system ids

What?  Outside a doctype?  I doubt anyone could find an alternate 
solution without an example of how this is working; SystemID-s seem (to 
me) to be tightly coupled to doctype declarations.

> o  change tracking

I can see this one.  In fact, replacing this would be hard; change 
tracking could break the tree with (HTML-style) spans that break 
tag-start/tag-end parity.

You could do it with plain old markup, anyway <del timestamp="" 
author="">...</del> and corresponding <add /> and <repl />.  Doesn't 
map 1:1 where tag enclosure breaks, but otherwise easy.  Breaks DTDs 
(but could be made to work by DTD inclusion: define these bits in 
something that %pe-includes the real DTDs).

> o  short duration meta-descriptors

Don't know what this means.

> o  medium legacy code

Don't think I know what this means.  It could mean: there are 
processors that understand specific processing instructions.  Sort of 
the original use case for PIs, you know?  So, if there are alternate 
solutions of various sorts, replace the legacy processors with the more 
widespread ones, and replace their (pi-based) markup with (whatever).

> o  page fidelity

A great deal depends upon what this means.  If it means "page marker 
persistence" (so that someone can do the usual bibliographic cites, 
author, title (publisher: date), p. x), the ebook people seem to be 
solving this, though it's inevitably going to be an issue while display 
resolution varies.  If it means "inline presentational markup" (that 
doesn't compete with the semantic structure), then PIs are a solution, 
as is the hideous abortion that is the "style" attribute in HTML.  In 
general, the solution to inline presentational markup is "don't do 
that!"  If it means that presentational markup external to the semantic 
markup can be associated, you already know how to do that, I think 
(xsl:fo comes immediately to mind), so it must mean something 
different, right?

Amy!
-- 
Amelia A. Lewis                    amyzing {at} talsever.com
Love?
A joke, that.  Love was the problem, not the solution.  Being hit by a
car was better than love.
            -- Steven Brust, PJF, "Cowboy Feng's Space Bar and Grille"