[
Lists Home |
Date Index |
Thread Index
]
(Resending this one, it didn't came through to the list)
J.Pietschmann wrote:
> m batsis wrote:
>
>> So a formal mapping between CSS and FO properties is the way to go by
>>
> You'll find it in the XSLFO spec.
My point exactly.
>> To return back to my question on inheritance, perhaps you are
talking about what I think of as groupings, for example the subset of
background properties.
>
>
>
> No, I was talking about how presentational properties are tacked to
> nodes in the XML tree. Usually nodes inherit certain properties from
> their ancestors. Two problems:
> - Which properties are inherited, and where is inheritance blocked.
I don't think there are differences since the FO property definitions
come directly from the CSS specs, unless you go beyond those and into
the actual tree of the XML document, mostly meaning containers (i.e. div
elements in XHTML) mapped to FO blocks.
> - The interaction of inheritance, shorthand evaluation and default
> values.
I fail to understand "interaction of inheritance". Shorthand properties
are usually evaluated after being broken up into more specific
properties. That's why I mentioned the importance of document order.
> - If the properties are declared separately and connected to the tree
> by a matching mechanism, it has to be defined how this is done. For
> example
> div.sect { margin-top: 3pt }
> div { margin: 12pt }
> Unfortunately, CSS mixes all aspects and thereby muddles what
> "inheritance" means.
CSS is pretty clear on it's view of cascade and inheritance. A div.sect
will have a {margin:3pt,12pt,12pt,12pt} in a sequence of
top,right,bottom,left.
But your example clearly points out an issue, which is the need for
algorithms that will efficiently harvest FO style properties and group
them under CSS selector statements, meaning it is easy to convert from
CSS to FO but not vice versa. I expect people to edit having the CSS
version as a reference (witch is where the benefit exists anyway) and
produce FO from that.
If you have two-way conversion in mind I'm sure many examples from FO to
CSS can demonstrate problems but I don't see conversions of FO to CSS
mandatory or even of any benefit at all. IMHO, all the benefits of CSS
is on the (easy and convinient) authoring side, where the document keeps
changing.
Manos
|