• To: xml-dev@l...
• Subject: Re: Well-established uses of processing instructions?
• From: Rick Jelliffe <ricko@a...>
• Date: Mon, 09 May 2005 18:48:45 +1000
• In-reply-to: <200505090824.j498OakL003599@m...>
• References: <200505090824.j498OakL003599@m...>
• User-agent: Mozilla Thunderbird 0.6 (X11/20040502)

Michael Kay wrote:

>I sometimes use processing instructions in a source document as instructions
>to a stylesheet, for example
>
><?insert-current-date-here?>
>
>One could of course use an element just as well; but that means changing the
>schema/DTD, which one might have no control over.
>
>This is not really something I would recommend: it's very likely that other
>applications designed to process XML conforming to this schema will fall
>over when they encounter this undocumented PI. But it can be handy all the
>same!
>  
>
Why would they fall over?  Typically unless you specifically do something to
make your application act on PIs, they are stripped or ignored. 

Applications that fall over when they discover a PI they don't 
understand are
improperly written.

But PIs used to generate text are non-portable, and so appropriate for 
documents
that are being pre-processed. Probably what should be used is  entities:

<!ENTITY current-data SYSTEM "http://www.eg.com/dynamic?get-current-date">

because the information is data content.

Cheers
Rick

Cheers
Rick Jelliffe

• Follow-Ups:
  • RE: Well-established uses of processing instructions?
    • From: "Michael Kay" <mike@s...>

• Prev by Date: RE: Well-established uses of processing instructions?
• Next by Date: Re: attribute order
• Previous by thread: [off topic] Patent abuse: not just software
• Next by thread: RE: Well-established uses of processing instructions?
• Index(es):
  • Date
  • Thread