Re: Escher could have drawn it (Re: XML Schema and Entities)

[Peter Flynn <peter@silmaril.ie>]

At Friday, 22 June 2001, Alaric Snell <alaric@alaric-snell.com> wrote:

>But... they're called Processing Instructions!

A historical curiosity. It refers to processing the document *after* 
SGML
has finished with it (eg typesetting).

>My only example of their real-world use is in PHP:
>
><foo>
><?php
>	... executable code! ...
>?>
></foo>
>
>....but PHP just happens to use that syntax for compatability, you 
can just as
>easily write:
>
><<?php echo $element_name?> foo="bar">
>  Content!
></<?php echo $element_name?>>

But then you have invalid markup, which is both useless and pointless.

>because it finds the <?php with a state machine that traverses the 
script.

No, because PHP neither knows nor cares about preserving
the non-PHP parts of the document. 

>> For example you can encode a forced  line-break in a title that
>> you know will be required when the document is typeset but which
>> you don't wish to hard-code wrt the document structure:
>> 
>> <title>How I Made <?LaTeX \\?>A Million Dollars</title>
>
>Eeeeek.

If you are worried by this then I think you have missed the point.

><title>How I Made <typesetting-hints:linebreak />A Million Dollars</title>
>
>....suits me more :-)

That is perfectly possible, but it requires adding to the DTD or 
Schema if
one is in use. The whole point about PIs is that they are NOT part 
of the
structure of the document. In TEI I often use <LB> for linebreaks 
because
it *IS* part of the document structure when you are encoding the 
typographic
appearance of a historical document. When the linebreak is merely 
incidental
to the appearance on a current occasion, a PI is more appropriate.

///Peter