[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: XML-appropriate editing data structures

• To: Dave Pawson <dpawson@n...>
• Subject: Re: XML-appropriate editing data structures
• From: Ari Nordström <mayfair@t...>
• Date: Mon, 12 Apr 2004 10:53:26 +0200
• Cc: xml-dev@l...
• In-reply-to: <6.0.1.1.2.20040412084411.026c2e40@p...>
• References: <000601c41cf4$0e7fa1e0$6800a8c0@bagpeddlerjlake> <200404101916.26068.mayfair@t...> <6.0.1.1.2.20040412084411.026c2e40@p...>
• Reply-to: mayfair@t...
• User-agent: KMail/1.6.1

On Monday 12 April 2004 09:49, Dave Pawson wrote:
> At 18:16 10/04/2004, Ari Nordström wrote:

<snip>

> >The DTD that is used will have to be right for the purpose--writing that
> >particular document type in that particular context--but it doesn't have
> > to be the same that is used when publishing the complete document.
>
> I'm curious about the rationale behind this statement.
> I can think of subsetting say docbook, to reduce the list of valid
> tags at any one point, perhaps remove a set that I'm unlikely to use,
> is that what you mean?

Subsetting is certainly one way to go. I've done this on a number of 
occasions, removing structures that aren't appropriate for the context in 
which the information is edited but which could pop up when the complete 
document is published/processed. 

An example that comes to mind is the J2008 DTD often used by the automotive 
industry that I made more author-friendly for a number of contexts. The 
original DTD was simply too big and used constructs that were too permissive. 
There was no way authors could make easy-to-follow, no-nonsense service 
procedures with that beast.

Or, to pick another example from J2008, that DTD uses some ugly linking 
constructs that I don't want authors to use (having replaced them with models 
more appropriate for the document management system in use) but that I need 
to convert to when exchanging information with other manufacturers or 
organizations.

Another way is to be more liberal for the authoring DTD, in controlled ways, 
and rely on processing before publishing to handle the places where the 
author's been "liberal". This can be useful when publishing in different 
media--for example, online and paper will sometimes present very different 
requirements. If you use a DTD that simply allows everything when publishing, 
that DTD cannot ensure that you've used the correct structures in an online 
context.

An example that comes to mind is, again, J2008; its meta-data and linking 
models aren't enough for a modern document management system but that are the 
ones you need when exchanging information. When authoring, you want to allow 
for your in-house structures since those are the ones that make the magic 
happen when publishing, but which you simply cannot allow when exchanging 
information, because the exchange structures aren't rich enough.

Best,

/Ari

• References:
  • Converting xml to csv with xslt
    • From: "Jerry Lake" <jerry@b...>
  • Re: XML-appropriate editing data structures
    • From: Ari Nordström <mayfair@t...>
  • Re: XML-appropriate editing data structures
    • From: Dave Pawson <dpawson@n...>

• Prev by Date: Re: xml editors
• Next by Date: RE: XML Binary Characterization WG public list available e
• Previous by thread: Re: XML-appropriate editing data structures
• Next by thread: Re: XML-appropriate editing data structures
• Index(es):
  • Date
  • Thread