Teaching CSS: there’s a long way to go
This email to the css-discuss mailing list does a great job of describing the confusion and frustration that still confronts traditional web developers who are only just starting out on the road to mastering CSS. When you’ve “got it”, it’s easy to forget how much of a paradigm shift it is away from old school table methods. Here’s an extract:
Step Eight.
Just when you think you’re settling down into a slow and steady learning curve, this is about when you start getting emails from everyone who uses your site describing all kinds of variations on your layout as it has been interpreted by their varying browsers and platforms. This stage is the most important of all, the one where you realize that CSS support is far, far more random than any HTML workarounds that you’ve been dealing with for the (insert personal experience here) years you’ve been making web pages.
(Excerpt from an email from a user of one of my sites: “the new color and stuff on the homepage looks good, except on my computor [sic] some of the pages are cut off at the bottom and have big gaps in them”)
Maybe a good analogy to make here is one with Linux: both are great in principle, but if you aren’t comfortable with what you are doing you can run in to a whole bunch of problems. I wouldn’t recommend anyone who is still on the CSS learning curve to move a big commercial project to pure CSS, just as I wouldn’t suggest a Linux newbie start hosting their own internet facing server.
At any rate, it’s obvious that we as a community still have a long way to go in creating useful resources for people who want to make the switch to CSS.
ben - 19th November 2003 06:15 - #
Simon Willison - 19th November 2003 06:20 - #
Plus, I think, you have to make actual business decisions (or, if you don't care about money, at least actual customer experience decisions) about which client environments get the "full experience," which get the "graceful degrade" and which you won't worry about at all unless you get lots of complaints.
The one everyone seems to struggle with is the Netscape 4 series. You still have product managers out there who want to provide the full experience to that browser. And to do that, you're talking about either CSS hacks or HTML hacks.
My CSS life got so much easier as soon as I decided I would relegate the N4 series to the graceful degrade. And as it becomes easier to simplify markup (toward, I think, but never quite all the way to fully semantic HTML), I figure I might be "accidentally" supporting client environments I didn't even contemplate. No harm there.
Jay Small - 19th November 2003 12:55 - #
W3 Schools' has a page on browser stats. Browser stats for Netscape 4 seem now to be running at around 1%. Once a browser is first, no longer under development, and secondly, being used by less than 1% of visitors, I'd have thought gracefully degrading its experience - or even hiding all the CSS from it - was pretty reasonable.
Of course, that is a worldwide average and there will be different figures for certain sites. There are a number of caveats, too. Opera usage must be higher, as it can lie about what it is. I wonder if Opera usage is actually higher in Scandinavia, or if there's no difference.
Michael - 19th November 2003 14:14 - #
owen - 19th November 2003 14:40 - #
Clint - 19th November 2003 16:18 - #
Lou - 19th November 2003 16:44 - #
Michael Wexler - 19th November 2003 18:23 - #
The answer to Michael's problem? Don't send HTML-formatted emails!
Andrew - 19th November 2003 18:59 - #
While I'm nowhere near as good at CSS as most people here, I have to say I didn't find it terribly difficult to learn. Then again, I'd been using LaTeX for about a year at that point and the whole separating content from presentation was very natural to me. The fact that browsers don't always interpret things the same way was just another thing to tweak/keep in mind/occasionally get really frustrated by when something was really braindamaged.
I second Andrew in HTML e-mails BTW.
Mili - 21st November 2003 08:30 - #