xml-dev - Re: Why not PI's?

[tpassin@home.com]

Paul W. Abrahams writes -
>
> I see PIs as something you wish to pass on to a particular processor or
class
> of processors.   My sense of the matter (admittedly, unsupported by any
> documents I know of) is that they are, or ought to be, inessential to the
> general interpretation of a document.   In other words, a well-formed and
> valid document ought to remain well-formed and valid if all PIs are
omitted.
>
> Examples of PIs would be an indication that the following space character
> should be non-breaking or that a page break would be desirable at this
> point.   These would be "tweaks" in the form of typographical hints, quite
> possibly particular to a specific version of a document typeset with one
> specific typesetting program, e.g., TeX.

With these kinds of usages in mind, you could consider PIs to be a different
kind of markup superimposed on a marked-up text.  For example, an XML
document that contained RTF would have two more-or-less orthogonal markup
systems imposed on the underlying text.  PIs are orthogonal to the rest of
the XML markup.  So in a way, PIs represent a little bit of some undefined
orthogonal markup system.  The XML provides the structural markup.
Furthermore, the PIs do not produce a self-delineating structure, which
again is quite different from (the rest of) XML.

Mabe this is the basic  reason why so many people are uncomfortable with
PIs -they appear be be from a different, undefined system.

Cheers,

Tom Passin