FO. Renderx DTD. 'Overloaded properties should not be inheritable.

Subject: FO. Renderx DTD. 'Overloaded properties should not be inheritable.
From: "Paul Tchistopolskii" <paul@xxxxxxx>
Date: Sat, 18 Sep 1999 14:29:01 -0700
Sep. 18.

We have significantly updated our DTD with 
some tables-specific information and comments.

At the end of this letter there is one of 
our ( typical ) comments from FO DTD 
published  at  www.renderx.com

Hope some day somebody from XSL WG would 
read our comments for 10 minutes and then 
spend 10 minutes answering us with brief: 
"you are wrong in assumption N". If days of our 
hard work are not worth spending 10 minutes 
for writing us just one letter, it means that WG 
is working very hard for last moths - and  it will 
result in something realy cool quite soon. 

If this is true, we apologize for bothering you - 
maybe it is realy a problem to find 20 minutes.

If our work is not just the waste of our time, it 
would be also great to get a clue what is the 
scheduling for  the next version of  XSL WD. 
It would allow us, at  RenderX,  to syncronize our 
movement with the movement of  XSL WG 
( if there is any. It is hard to tell, because there is 
zero information available from XSL WG and in 
may look like there is no movement with FOs for 
months.)

It may be also reasonable to assign somebody who is
participating in the XSL WG to work with us. I apologize
if it goes against  one of  W3C policies  e t.c.

Rgds.Paul.

<comment>

B. All OVERLOADED properties are not inherited.

By "overloaded" we mean properties that may occur in elements of 
different types, assuming different meaning. For instance, 'height'
may be used to specify the height of a container area or the height
of an image; specifying this property too high in the document
tree may give rise to confusion: you should always bear in mind
*all* the elements to which the property may be applied. That's 
why we have preferred to keep all such properties not inheritable,
awaiting for further revelations from next versions of WD XSL.

On the other hand, there exist a group of properties related to
a single type of elements (or to a group of closely related types,
like fo:inline-rule / fo:display-rule or fo:inline-graphic / 
fo:display-graphic). These properties are "safe to inherit" -  if you 
specify 'rule-thickness', you are sure that only rules will get it. 
All such features are permitted at almost every element in the tree 
(see below about limitations).

Certainly, the above criteria are not enough to determine the 
inheritability of all properties, and in many cases, we have made
arbitrary choices based upon esthetic considerations rather than
strict principles. One important deviation from the criteria
is our handling of 'vertical-align' in the tables. 'Vertical-align'
cannot be made universally inheritable, since there are at least
5 element types that consume it, sometimes with different semantics;
on the other hand, specifying 'vertical-align' on every cell might
be boring. At this point, we have neglected our own principles for 
conciseness' sake: fo:table and all its fo:table-* children 
are permitted to bear 'vertical-align', passed then to their 
fo:table-cell children.

(We confess that we don't like this solution. We'd rather prefer
to split the <vertical-align> into several heteronymous properties
and make something like 'table-cell-vertical-align' universally
inheritable).

</comment>

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 paul@xxxxxxxxx   www.renderx.com   www.pault.com
 XMLTube * Perl/JavaConnector * PerlApplicationServer
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=



 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list


Current Thread