Emmanuel Pietriga wrote:
> Ryan Lee wrote:
>
>> Emmanuel Pietriga wrote:
>>
>>> Ryan Lee wrote:
>>>
>>>>>> 3. Ordering / conflict:
>>>>>>
>>>>>> - Consider formats dealing with children to be a frame outside
>>>>>> of the box, not interfering with the actual box. E.g., an
>>>>>> allPropertyFormat on a resource will place its content
>>>>>> 'outside'
>>>>>> of every property, and if any property has further specific
>>>>>> directives from a Format, those contents go inside the
>>>>>> allPropertyFormat frame. The two do not influence one
>>>>>> another.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> I'm not sure I like this idea of having one more box. In my mind,
>>>>> all*Format and group-level declarations are just a convenient way
>>>>> of declaring formatting properties that apply to many things once
>>>>> and for all.
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> Without calling it a box, I suppose my proposal is that these things be
>>>> additive instead of conflicting directives. The order of addition
>>>> relies on the scope of the directive. And in practice, I think the
>>>> content* strings will appear on the output as one element, not a set
>>>> of nested boxes.
>>>
>>>
>>>
>>>
>>> My concern is not really about boxes. It's about the scope of these
>>> declarations. I consider them to do the same thing. The only
>>> difference is that those declared at the group level apply to all
>>> formats that belong to the group, even if a format by itself does not
>>> declare them. Group-level declarations are thus just convenience
>>> methods and should not have a different semantics.
>>
>>
>>
>> Group-level styles are added to any other applicable styles as it is
>> possible in XHTML to declare multiple classes on one element. This
>> allows you to e.g. say something about how every rendered resource
>> should look.
>>
>> The same approach seems to make sense to me here as well. I think
>> they all do do the same thing, in a cumulative and ordered sense.
>
>
> If you do:
>
> <html xmlns="http://www.w3.org/1999/xhtml">
> <head>
> <meta http-equiv="Content-Type" content="text/html;
> charset=iso-8859-15" />
> <style>
> p.foo1:before {
> content: "foo1 ";
> }
>
> p.foo2:before {
> content: "foo2 ";
> }
> </style>
> </head>
>
> <body>
> <p class="foo1 foo2">p's content</p>
> </body>
> </html>
>
> only foo2 appears as content before p's content.
>
> Not both foo1 and foo2.
I don't fully agree that different scopes should be interpreted as
conflict and thus require resolution, but I'll go along with it. I
suppose someone who wants to do formatting regarldess of type can put it
in themselves when transforming the Fresnel tree.
Revising the original proposal:
2. Stricken, unnecessary ("Mint new terms...").
3. Conflict:
- Consider formats as specified in fresnel:Format to take
precedence over those in fresnel:Group.
- instanceFormatDomain directives take precedence over
classFormatDomain.
- Only allow one instance of any *Format predicate on a given
FormatDescription.
I've amended the vocabulary to reflect any changes. It should be frozen
at this point; I'll add these changes to the manual a bit later.
--
Ryan Lee ryanlee_at_w3.org
W3C Research Engineer +1.617.253.5327
http://simile.mit.edu/
Received on Thu Jul 14 2005 - 22:30:17 EDT