- From: Wingnut <wingnut@winternet.com>
- Date: Wed, 04 Feb 2004 11:27:58 -0600
- To: www-html@w3.org
Hi Gang! The noisy nun here again. Um, after sending an email to a fine style tech, I realized that I needed to take this proposal a bit further... even if some of you DO yell "spam!". Regarding the proposal to move border, along with underline, linethru, and overline... out to becoming CONTENT, possibly made with OBJECT tags... ...this new "floating, bound, re-useable, object-tag-made border" does one OTHER drastic thing to the CSS spec. It opens the possibility of removing all MARGIN settings from all containers and box models. Let me try to explain how. This hypothetical new object-tag-made-border, heretofore "OTMB", is something that is "binded" to one or more selector-found box models in the document. When its "elastics" parameter is set to {50%, 50%, 50%, 50%} the border exactly straddles the intersection of a box model's content edge, and its container's padding edge. It borrows HALF of its width from the content's padding, and HALF of its width from the container object's padding. I am leaving "margin" OUT of the talk at this time, for a reason. Pretend the margin is set to ZERO on the content box model. Now, let's set the OTMB's elastics={50%, 40%, 50%, 40%}. This setting indicates that the border is willing to stretch "outward" to a 50%/50% straddle of the content/padding intersection ON THE TOP AND BOTTOM, but it is only willing to stretch-out to a 40%/60% content/padding straddle ON THE SIDES! Ok, now apply this "stretch-to-almost-fit" border to a line of adjacent inline box models, and what do you get? Bordered "cells" with a small space in-between the side-borders of adjacent cells. The adjacent cells' CONTENT actually butts-up against one another... their innerTEXT's distance set by each box model's PADDING setting. BUT, the new floating border allows the side borders on each cell... to be shunted short-of the content edge... giving the APPEARANCE of margin-separated adjacent inline cells. But only their (floating, stretch-to-almost-fit) borders make them appear separated by space. And yes, this creates a brand new problem for BACKGROUND COLOR, because we might see two background colors intersect between each cell of the inlines. That MIGHT be a show-stopper to this idea. Box model background colors and images will need to have a "visible outside the border? true/false" on them, I suppose. All in all, though, I wanted to try to show that these new floating, non-padding-interfering borders (OTMB's) could eliminate the MARGIN parameter of CSS... and thus solve yet another "little tug-o-war" that seems to happen between a container's padding and a contained-element's margin. This "OTMB binding" is another story. I envision a certain type of OTMB to be reuseable, and "stretch-to-fit" for many onscreen box models, much like a style is reuseable for all elements of a given selector. We would NEVER want to do an absolute positioning of an OTMB. It IS absolutely positioned in a certain place "atop" all box models its bound-to, though. Its sizing params ALWAYS uses "percent of stretch across content-container/content-padding intersection" and avoids/disallows hard-set absolute positions/sizings. The upper left corner of these floating borders is "bound" (in stretch-related, slightly-variable positions) to the upper left corner of various box models onscreen. Therefore, it is my opinion that even though these borders are floating, they are "bound" to that (and maybe many) element, and will therefore not suffer from bad scaling during window resizings. Page reflows should be fine as well. The border (or its "type") will travel with the element just like its style does. Now, though, its out of the way of the true stylers. It floats, and therefore can no longer cause content intersection problems. Then, if margin leaves, content/container intersection certainly gets a fresh dose of "all-box-model tables are indeed a possibility again" vitamins. Now, lets get going on building that "Visible outside of bound borders?" setting for box model background colors and images, huh? :) Best wishes! Wingnut
Received on Wednesday, 4 February 2004 12:31:08 UTC