"xhtml" in Weblog

Roger Johansson of 456 Berea Street has just published a fantastic article on the pitfalls of serving XHTML as application/xhtml+xml, which is what it should be. I knew about a few of them - things like IE not supporting it and any errors making the page unrenderable, but there is so much more to it than that. Ye gods, what a nightmare.

It’s been a while since Dunstan made any technical posts but this Solving CSS Problems for Mozilla Europe serves to remind us of why he’s one of the most respected CSS folk in the business. Demo 2 won’t be a lot of use to me because it won’t work in IE but Demo 1 is a stroke of genius… A font resizable menu pulled up from the end of the source. Now that could be very useful indeed.

The Future of Accesskeys from Derek Featherstone of WATS.ca makes for some interesting reading… I decided a while ago that accesskeys were poorly implemented - I approve of the UK government’s accesskeys system as an attempt to standardise keyboard shortcuts but it’s quite limited in it’s scope and therefore not incredibly useful. Derek’s article highlights the replacement of accesskeys in the XHTML 2.0 spec with ‘access points’. This means that the user can define their keyboard shortcut for ‘contact’ and that combination will work on every site that implements that access point. It’s consistent, simple and very far off in the future! At least it’s there though… In the meantime, it would appear that the W3C has reached the same conclusions as a number of web developers and dumped access keys. That’s that then. Long live access points.

It’s been a while since I’ve been really surprised about a new hack but this one has amazed me… How To Clear Floats Without Structural Markup, documented on Holly and John’s Position Is Everything, is an incredible study in bending rules! I’m not sure whether I’ll actually use it yet but very very well done to everyone involved.

I’ve never had to monkey with HTML 4 Transitional doctypes before but the project I’ve just finished required it and it’s brought home exactly WHY I’ve never done it before.

Having the HTML 4 doctype without the URL declaration throws IE6 into quirks mode so it behaves like IE5.5 in terms of box model and hacks/filters… Not such a bad thing in the short term - at least it allowed me to keep my box model hacks simple (* html all the way), but it was only when I came to test in Opera that things got complicated. IE6 is in quirks without the URL and Opera is in quirks too. Fine. But Opera doesn’t fall for the * html hack so it doesn’t get fed the old box model dimensions. In fact I couldn’t find ANY sensible way of filtering Opera from Mozilla at all so Opera was stuck with W3C box model figures. That’s it. As far as I can tell, using the HTML 4 Transitional doctype makes it impossible to produce good cross-browser pages. This is hardly earth-shattering but I’d never thought about it before.

I think I’m a bit late on this one but style:phreak has put up a nice little demo of an XHTML form. Now that’s hardly groundbreaking but the code is very nicely structured and I particularly like his ‘hide optional fields’ option. As a user that’s something I’d really appreciate, especially for long forms, although I doubt you’d squeeze that one past many marketing departments…

There’s been a really good discussion of web standards and their uptake going on over at asterisk over the past few days. Keith called on beginners and pros alike to say what they thought the barriers were to the widespread adoption of web standards were and it appears to be something that everyone has an opinion on! It’s great to get so many different perspectives on what the issues are and how people intend to deal with them. The audience is obviously a bit biased - asterisk is heavily linked from the WS blog community, but there’s still quite a variety of views expressed…