Re: Formatting_the_result

[Ron Ross <ronross@xxxxxxxxx>]

On Mon, 4 Oct 1999, David Carlisle told this:
> 
>> I don't mean to suggest that we should put a
>> full-blown pretty-printer in there, but at least avoid splitting the
>> tags up like it does now.
> 
> Is it clear that there is a prettier way that is valid for an
> arbitrary DTD? Adding the newlines within the tag is always safe, but
> once you decide to add newlines within the content of elements you
> have to fight with the whitespace handling rules of sgml, which is
> never much fun, to check that you are not changing the content of the
> document in the course of pretty printing.

Oh god, not the whitespace connundrum! ... but actually, could OpenJade
be made to distinguish between data or mixed content elements and block
elements in outputing its linefeeds? Any element with #pcdata (and all
its children) gets output to a single line (AFAIK, linebreaks are only
significant as pcdata -- or am I wrong?). Of course, a DTD like HTML,
which provides for pcdata anywhere at all, poses a problem; but then
whitespace is never significant in HTML except in the PRE element (and
some scripting elements), so HTML, specifically, could get newlines
anywhere except in pre.

Sorry for the neophyte's 2¢ interference, but this characteristic of
jade output renders it just a little less cool than it could be.

Also, in the same vein (in terms of output usage - or should I start a
new thread?), I would like to see RFT output become more "structured".
It's paradoxical that, at a time when word processors are finally
bringing some intelligence to document structure, we, engaged in editing
the current model and epitome of structured content, are outputting a
single stream of micro-formatting codes. I've started using Jade's
RTF-specific heading-level extension, which is great to have, but it
would be nice a) to be able to create any named style, such as for body
paragraphs, special blocks, etc. and b) have all (or as much as
possible) of the formatting resulting from the FO characteristics be
encapsulated in the resulting rtf style (when this is used for that FO).

I'm not the one to make heads or tails of the C++ code. I'd simply like
to know what comments others have on the feasibility

Cheers,

Ron.

-- 
Traductions Ron Ross Translations
ronross@xxxxxxxxx
"They were magical delusions, fireworks." -I. Murdoch


 DSSSList info and archive:  http://www.mulberrytech.com/dsssl/dssslist