Gabriele Romanato asked me some questions about CSS and the web, and I obligingly answered.
Quite late actually. We got a dialup connection at home in the spring of 1999 and I spent my first hours/days online in that summer. Afterwards I went to university and had broadband web access there ever since.
What is interesting is that from the outset I was interested in alernative browsers, and particularly Opera. I knew about Internet Explorer and Netscape and had them both installed, but I was immediately sold when I tried Opera 3.60 which had just been released and sported very good CSS support for its time. I have tried other browsers since, but never really turned away from my first choice!
I played a bit with simple HTML, obviously with a table-based layout (remember this was 1999). I still distinctly remember where my interested in CSS originated: seeing Rijk van Geijtenbeekís page on PowerWare, which I believe was an entry for a webdesign contest issued by Opera Software. I was awed by the design (only Opera could render it correctly at the time) and it definitely sparked my interest in CSS, although it would take a while before I came proficient in it.
As far as I can remember, my first real thorough exploration of CSS started when I got my hands on the early Opera 7 beta's in 2002 which had a totally rewritten rendering engine. That is also when I stopped with my haphazard trial-and-error approach to learning CSS and started doing tutorials and reading specifications. I found I could often immediately skip to the 'Advanced' section of the tutorials, which made it all the more fun to continue. By this time I had also registered my domain, so I had a place to practice my skills.
Obviously I would say Eric Meyer with his Complex Spiral Demo, Tantek for his polygons, and I learned a lot by discussion with The Literary Moose (now known as Lofotenmoose, or simply Moose) and his never-ending drive for pushing CSS to its limits.
My CSS experiments were attempts to show off new (at least, to myself) and exciting possibilities of CSS. I created my first experiment when I learned of Opera's support for CSS generated content and counters. I thought it was fascinating and took it to the extreme with my No Content, Only Style experiment. That is how it all started.
Most of my CSS experimentations never saw the light of a computer screen, other than my own. Like most people starting with in CSS I did my own version of of a two-column, three-column, two-column-with-header, three-column-with-..., etc. layout. I never stored those, or similar, experiments. I only stored the ones I thought were interesting or fun enough. Especially the latter criterion was important!
The most fun I had creating was my scaled background image experiment. I was travelling in the train, bored, and started thinking about problems in CSS. I wondered why you couldn't scale a background image, and then decided to try and tackle the problem. Shortly later I had thought up the trick to do it, I wrote down the code on a piece of paper, tried it at home in Opera and it worked!
Many of the experiments also were a result of an itch I needed to scratch. A good example is how on my site the id's of headers appear when you hover over them, making linking to specific sections of a page much easier; in my case I frequently had to link to specific portions of my M2 tutorial, and this bit of code made my life easier.
Many people would give a common answer to this question: browser incompatibilites. It can be excruciatingly annoying. In all fairness I never really ran into that problem, as I never had to design complex webpages that run on all currently available browsers. I was able to design for the more modern browsers and was satisfied if the page looked reasonable in older versions.
I ran into difficulties of another kind when I got to the point that I really wanted to understand what I was doing, and wanted to understand why things went wrong in browsers: I started to dig through the CSS specs. It took me quite a while before I was able to read the CSS specs with ease. The readability has definitely improved over the years, but it was also my increasing experience. Two summers ago I even sat down to read every paragraph of the spec, and although I learned things I'd never known before, I would advice people against doing the same thing: buy a good book on the topic instead!
As I mentioned before, I started using Opera quite early on, in 1999. My reasons for choosing it then are still equally valid today: first and foremost the multiple document interface (MDI) which allowed you to have multiple pages open at the same time. The tabs didnt' appear in Opera until, I believe, Opera 4, but the multiple pages have been there from the outset. Other reasons for choosing it were speed and keyboard shortcuts.
In the years after that, the innovations added to the User Interface such as the integrated Google search, mouse gestures, and FastForward, maintained Opera as my browser of choice. It also became my e-mailer of choice with the advent of M2 in Opera 7. This focus on being an internet suite, with a lot of integrated features (mail, newsfeeds, notes) and the fact that the interface can be precisely tailored to my needs, makes Opera my browser of choice.
I believe 'accessibility' was an overhyped word a couple of years back. In my opinion it should simply mean that the content of a page is easily available to all users. That means structured, semantic markup and clear styling. CSS is helpful in both respects, even more so because it allows users to override the page author's settings and apply their own preferred layout. So yes, accessibility and CSS are a good combination.
I have only followed IE7's progress from the sideline, as I cannot install the beta's on my Windows 2000 machine. From what I've read they have fixed numerous pressing rendering issues, and I'm very glad indeed that this improved CSS support will be available to the IE users. However, people should realize that a new IE version is nog a magic fix. Not only will it take time before people upgrade (and that only includes XP users), but the problem of compatibility with older IE versions still remains. What is more, on the short term it will even result in extra work for webdesigners as another browser has to be taken into account when creating sites. Still, by all accounts it is very good news that IE is being developed again, and it might even give Firefox and Opera a run for their money.
Iím a very poor designer! In the past I have sometimes said that I could create a webpage of almost any design a good designer could conjure up — provided I need not bother with older browsers. Unfortunately I have never really taken myself up on the challenge, and my designs have remained simple, bordering on boring throughout the years. It serves to illustrate that although CSS is a powerful design language, not everyone proficient in CSS is a good designer! Currently I have a very basic design, which brings the content across clearly, but I intend to update it something a little more visually attractive in the not too distant future.
From a practical perspective I believe it could take a lot of time and effort to switch, mostly because it requires a very different way of thinking. CSS has its own way of doing things, which is admittedly not always immediately intuitive or obvious, and designers have to learn how to rephrase their designs in CSS. I often see designers who treat webpages as a blank sheet of paper, and create their design how they deem fit, without recognizing the possibilities and limitations of CSS. First and foremost they do not realize that webpages are a continuous medium, as opposed to a paged one such as a magazine, and that a lot of existing ideas simply don't apply.
Coming back to the question about switching from table-based designs to CSS-based: I think it is inevitable considering the tremendous benefits. And aside from the benefits, we are living in the 21st century! Any self-respecting webdesigner should have made the switch to CSS by now.