OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

docbook message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]

Subject: Re: DOCBOOK: Re: Image scaling

On Thu, Jan 24, 2002 at 09:16:11AM +0000, Matt G. wrote:
> I really don't like the idea of using PIs.  How is this any better than 
> adding this information directly in the content model?

- I'd see the use of PIs to be a transitionnal step towards having the
stylesheets completely handle the parpersize-specific issues

- PIs would make it easy to add specific infos for various processors,
and for various papersizes.  Currently a DocBook document making use
of those absolute-size semantics (including the use of ImageData
without any size-related argument) is specific to one paper format.
If it renders correctly on an A4 format, it will be suboptimal on a
Legal format (or vice versa).  That just means we did not succeed in
the separation of content vs. layout.
 OTOH, if the stylesheets really need a hint ATM, we can add one PI
per paper format, without any need to enhance the DTD.

> But, if you do this, you give up the ability to easily validate any
> of the formatting info, and you risk having it get stripped out, if
> there are any processing stages between the source file and the
> stylesheet processor.

Yes.  This has to be taken into account for each specific case, in the
current state of things, to choose between using the current
attributes or using PIs.  But if we change the ImageData semantics and
processing expectations (or maybe use a different element instead, to
avoid breaking current use of ImageData), so that the stylesheet is
responsible for the layout aspect of the thing, then PIs will become
unnecessary anyway.

> In general, however, I think that much of the time, people should use the 
> 'role' attribute of imagedata, and the stylesheets should apply image 
> formatting, based on that.  For the kinds of formatting control that 
> imagedata provides, it seems like various attribute combinations used on 
> the imagedata elements of a given document are likely to be correlated with 
> some particular aspect of the role of the images, in that document.

I do agree.

> However, omitting any formatting information, for imagedata instances, in 
> the source document, would probably be unacceptable to a significant number 
> of users, as it would force many into writing stylesheet customization 
> wrappers who otherwise would do fine without.

Except if the modular stylesheets implement this already I suppose.

> Furthermore, it seems like you'd also have some number of people
> writing document-specific stylesheet components, which essentially
> implement the out-of-band formatting I was describing.

Maybe.  Finally, I've no real opinion about favoring PIs or the
out-of-band formatting you describe :)

Yann Dirson <Yann.Dirson@fr.alcove.com>                 http://www.alcove.com/
Free-Software Engineer				      Ingénieur Logiciel-Libre
Free-Software time manager    	       Responsable du temps Informatique-Libre
Debian GNU/Linux developper <dirson@debian.org>

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]

Powered by eList eXpress LLC