- From: Matthew Brealey <thelawnet@yahoo.com>
- Date: Fri, 5 Nov 1999 03:35:17 -0800 (PST)
- To: www-style@w3.org
--- Jan Roland Eriksson <rex@css.nu> wrote: > On Wed, 3 Nov 1999 19:39:08 -0500 (EST), you wrote: > >There are methods for creating hanging indents > using positive left > >margin and negative text-indent. > > There are simple reasons as to why you would not > want to use the margin > property for this... Not to mention the fact that it sets up a block for the element that is of the wrong size. (Imagine the havoc that such a method would cause in columns (the text-indent would overlap the previous column). > Just as I prefer to use RISC based CPU's for > whatever programming task I > may be assigned too at work, I would like to see a > well implemented > "mean and lean" version of CSS too 1. Use CSS 1. 2. Use the CSS 1 core. In addition, it is perfectly possible to make pages using only a subset of CSS that works. For example, I use a script that does browser detection to serve my style sheets, plus I have a "mean and lean" style sheet that works on any browser (containing color, font, margin-left, etc.). It is true, however, that the fact that CSS is so complicated hampers takeup. Authors see that although it would be nice to replace all those tables and FONT tags, there really is little point, because they'll probably find that the CSS they've replaced it with doesn't work, or, even if they go to the enormous trouble of checking it against all the CSS bugs, they may find that it turns out there is some previously undocumented bug. There is no doubt IMHO that the biggest reason that css is not used more heavily is the poor support for it. Unfortunately, this obscures the fact that 99% of common HTML formatting can be achieved in CSS with the gain of greater flexibility. And thus although things like color:, font: generally work well enough, authors are inhibited from using these because of the horrendous bugs elsewhere that act as a barrier to the whole of css. As I see it the best way to deal with this is to try and ensure better compliance with CSS. Rather than saying that because Browser X is hopeless at CSS, Word Processor Y shouldn't be given any more css, I believe the answer lies partly in modularisation . What I would like to see is a situation where a UA can claim CSS support if all it does is parse style sheets correctly and support (say) color and background-color. Ensuring that this is occurs is a somewhat hit and miss affair, so to assist in it I would suggest that a checklist for UA implementers is provided with each module - e.g., PARSING Does the UA recognise that an @rule is everything between the @ and the next semicolon or {}, which ever is the first? Does the UA ignore all property declarations that contain unknown values? Does the UA ignore all measurements that don't include a unit? etc. I am sure that this would help because at present what seems to happen is that UA implementers tend to say 'Oh look, here's background-attachment, let's implement that', and ignore the sections on parsing because they would rather have a product that supports loads of 'sexy' properties (badly) but one that interprets font: 800 12pt Arial as meaning 800 pixel Arial (as is the case with one UA) than one that actually works. Thus I would certainly like to see a situation where css support can be claimed with the support of only a few properties, but where support cannot be claimed with _more_ properties but where support is buggy. Unfortunately it is too late for existing UAs, so all that can be done is to attempt to hide buggy bits of CSS from them (either using a script, or @import "css.css" all; (which only catches Opera, which although buggy, does not have any page destroying bugs) and similar methods) > My interpretation of last years discussions around > CSS is that "the > advocates" wants it to go into a CISC state, where > authors are supposed > to rely on others to do their job correct first, > before they can start > to utilize the technology them self. > There is so much that can be done with CSS1 already, > together with a > fairly compliant ua, but I seriously doubt that > authors at large have > even tried to really learn about that part yet. So > when are they going > to have a chance to learn about the rest, if > proposals for new stuff > just keeps flooding in? But there is also so much that _can't_ be done. For example, you can't position elements and get other elements to flow around them. This is seen in almost any newspaper, and it is essential that CSS should be able to reproduce something as basic as a newspaper. In addition it shouldn't be forgotten that CSS is not just for the WWW. For example, many DTP programs now output HTML with css, but they find that many of the effects that have been common in DTP circles for many years cannot yet be achieved with CSS. ===== ---------------------------------------------------------- From Matthew Brealey (http://members.tripod.co.uk/lawnet (for law)or http://members.tripod.co.uk/lawnet/WEBFRAME.HTM (for CSS)) __________________________________________________ Do You Yahoo!? Bid and sell for free at http://auctions.yahoo.com
Received on Friday, 5 November 1999 06:35:19 UTC