Frames are a great way for splitting a document into several distinct areas
and for providing a dynamic, resizable, easy-to-use head/navigation/content
view.
The current HTML specification describes a rendering scheme that is
insufficient in regard to control of frameset background, resp. frame gaps
and borders. (E.g., it is currently not possible to eliminate the gap
between frames in a frameset or to define a frame's border visualization.)
The assertions made in the current HTML5 specification result from
algorithms put in place before CSS became a wide-spread method of adding
presentation to content.
I'd like to suggest to amend the HTML/CSS specification on HTML <frameset>
and <frame> elements in order to provide sufficient control over frame
rendering to the web site editor.
My suggestions splits into following (independent and disjunct) suggestions:
___________
* "cols" and "rows" attributes should become deprecated
in favour of following new attribute:
flow (horizontal|vertical) #IMPLIED
Reason:
The number of rows/columns results from the number of frames contained
within a frameset. There is no need to duplicate this information by using a
separate attribute, which adds unnecessary ambiguity.
It is sufficient to define the direction of frame flow, either horizontally
or vertically. The suggested "flow" attribute provides for this.
If flow="horizontal", frames within a frameset are aligned from left to
right, equally distributing the available with amongst them.
If flow="vertical", frames within a frameset are aligned from top to bottom,
equally distributing the available height amongst them.
___________
* Leave presentation information to CSS.
I.e.: Following frame attributes should become deprecated:
- frameborder
- marginwidth
- marginheight
- scrolling
Reason:
The current specification leaves a gap defining the layout of frames.
Instead of trying to fill these gaps, HTML should rely on CSS regarding
frame layout.
The CSS properties for border and margin are self-explanatory. Scroll bar
visibility should be defined using the CSS "overflow" property applied to
the frame/frameset elements.
___________
* Changing the value of a frame's "noresize"
attribute should not affect layout/presentation of frames.
Reason: Visual feedback on the availability of a resizing option should be
the responsibility of CSS.
For functional specification, it is sufficient to specify that resizing is
allowed and an appropriate NS/EW cursor will be displayed only if:
a) The gap space is greater than 0
- and -
b) None of the affected frames' noresize attribute is
being set.
___________
I have added a PDF file to this suggestion report, trying to visualize the
concept's details. It can be found here:
http://www.w3.org/Bugs/Public/show_bug.cgi?id=9337
There is also a discussion running in the CSS mailing list. The archived
discussion can be followed here:
http://lists.w3.org/Archives/Public/www-style/2010Mar/0433.html
RFC,
Axel Dahmen
www.axeldahmen.de
