- From: James Elmore <James.Elmore@cox.net>
- Date: Thu, 28 Jun 2007 10:46:33 -0700
- To: www-style@w3.org
David Woolley wrote: > > James Elmore wrote: > >> If the rules for margin collapse were consistent for every block >> object, they would be much simpler to understand. Right now, there are >> different sets of rules for blocks in text flow, blocks which float, >> and blocks inside tables. Make one set of rules for margin collapse >> and allow designers to control when and how to apply them. > > > The SS in CSS stands for style sheet. Style sheets are documents that > define the layout rules for a whole family of documents and typically > define a house style. In the original intended use of HTML, which > corresponds to white papers, user guides, and internal documents, not to > the glossy brochures that most web sites try to emulate,one wants to be > able to write simple and compact style sheets that will result in > correctly marked up text being rendered according to reasonably good > typographical practice. That means putting a lot of simple document > typography into the CSS language definition. I know this. When I first encountered CSS, about six or seven years ago, I thought to myself, 'what a GREAT idea!' I still think so. The original reason for CSS still exists and is just as valid as ever. But, (you knew there would be a 'but,' didn't you?) But, people are using HTML and CSS for other things as well. (Or maybe not as well, but they are being used, and abused, for other things.) What I suggested was not intended to allow designers to control margins on a per-pixel basis. (Although several people seemed to infer that from my proposal. Just guessing here, but I think there have been problems in CSS history from designers who want the type of layout features you mention below.) My thoughts were based on my experience as a tool maker and user. I see CSS as a powerful tool set. I also see that some of the wrenches and saw blades are missing from the set. That was the basis for my proposal -- the missing parts of the set bothered me, so I suggested adding a few of the missing parts to make the set more complete. > > The demand for sophisticated layouts comes from people writing the > equivalent of glossy brochures, a rather different problem area. Tools > for doing sophisticated but fixed layout, designs for these pre-date the > web; HTML made a very deliberate choice not to emulate them. I suppose that some of my proposals can be used to produce more sophisticated layouts. The margin-collapse feature seems to be one such hot-button. Without much consideration for the ultimate uses, I suggested that a missing, and apparently obvious, feature might be added. Now, I am not a designer and have no intentions (or ability) in that direction. I just recognize that HTML and CSS are powerful tools which have already far surpassed what their original designers intended in the ways they are used. My proposal, whatever finally happens to it, was merely intended to provide what I saw as missing parts of the tool set. With a more complete tool set, I would venture to guess that uses beyond either of our abilities to guess will appear. Perhaps designers will become more intelligent in their use of CSS tools, just as the general population became more aware of fonts and their intricacies when computers finally became able to handle fonts on screen and printers. Maybe history will work backwards and whitepaper writers will become more aware of design principles and produce more artistic work. I don't know, and I don't even begin to guess. I only spotted what I perceived as a lack in CSS and attempted to have that lack addressed. > > Personally I am not sure that many designers have the mathematical > skills to create rules for sophisticated fluid designs, and many work in > a very visual way. For fixed layout, there are tools designed for > electronic brochures, like PDF (which even allows the presentational > design to be annotated with the semantic structure, even if this is > rarely done; you can extract the HTML from a properly constructed tagged > PDF file. I don't know if designers have the requisite mathematical skills or not. But perhaps making the ability to better control margins will provide the user population with a better understanding of this sophisticated subject. I remember the difference in the general population when fonts became available generally on computers. Some people still haven't a clue about fonts and leading, about how characters collapse against each other, and whether serifs are better or not. But the understanding of fonts, in general, has improved significantly. I expect something similar with respect to styles, once the tools are available. On a side note -- before computers were able to handle fonts, font experts had to be both mathematicians and designers and font sets sold for tens of thousands of dollars. Only a very few people could design fonts. The number of people who understood and used fonts artistically (and in what was considered the 'proper' manner) was still a very small proportion of the population: probably near 1%. Back to your comments. I was not aware of that feature of PDF. As I said, I am not a designer and I only use PDF for portability, not design. It seems as though the designers of PDF (Adobe?) gave it some valuable thought. Could styles be extracted from PDF as well? I mean, if an author was told by his/her boss to match the company style, given a PDF sample, could some sense of the correct style be extracted automatically? (Yes, I know from what you say that many features of PDF do not exist in CSS, different purposes, different tool sets.) > > Although one might be able to use a model where the primary document is > semantics structured and there is a complex presentational overlay, I > think it will be difficult to do without presentational artifacts in the > semantics. The existence of semantic artifacts in tagged PDF is less of > an issue, because semantic boundaries are always potential > presentational ones. Sorry that I got up in my pulpit and started preaching. I know that the artifacts and complexities which exist in HTML / CSS will not go away. The tags with presentational capabilities will not magically disappear. The HTML/XTHML connections which are buried in CSS (because that was where it was first used) will be around a long time. Because of the number of times when I shot myself (metaphorically) in the foot by combining two different functions, I sometimes start preaching, trying to save the sinners from those sins I am already familiar with. What I would like, simply, is to take tiny steps away from what we all know is causing complexity -- overlapping presentation and content -- so our lives will be simpler. None of the complexity will ever disappear completely, but I can try and make my life, and the lives of other XML / CSS users easier with these steps. > > Note with regard to tables, I hope that you understand that > display:table is not restricted to table elements. Also, at least at > one time, CSS3 would allow phantom containing elements to be generated > and styled. > Yes, I know that any element can be declared as 'display: table' and the rules for adding unnamed elements should come into effect. What I am saying is that it should be possible to control some of the purely presentational features which exist only in tables using CSS in other elements. -- James Elmore 22162 Windward Way Lake Forest, CA 92630 Home (949) 830-9534 Email James.Elmore@cox.net
Received on Thursday, 28 June 2007 17:46:47 UTC