although it does look very nice, and the only place your gonna see something done that well is on a webdesign site.

Firstly, if we can’t justify doing it — as a web publisher/educator/direction setter — then probably no-one could justify it.

Secondly, your first ever CSS layout was probably a temporary backward step in terms of bang-for-client-buck over a traditional table layout.

But in the same way that you pay to attend courses and conferences without being able to directly attribute income to them (usually anyway), experimental stuff like this feeds back into how you approach a similar problem next time, and possibly how you approach more mundane daily problems. I learned really useful things about background-position percentages playing around with that ostensibly useless little limousine toy.

Of course, if you need to deliver a site yesterday afternoon, it’s maybe not a time to crack out the lab coat, but by the same token it is hard to expand your skill-set if you use the same solutions every time.

For the less ambitious, more pragmatic developer, I would have simply recommended providing a text alternative to the matrix, which could have been done with an image, SVG, or Flash.

Yep, that is exactly what I was thinking when writing my first comment. If this were a client project the choice may have been either quoting for ~1 hour to do an image + text alternative or 1 day (or more) to come up with the current solution. The cost difference could be substantial, and hard to justify.

As for the concern over the difficulty of implementing something like this, Cake may disagree with me on this, but I’d say what he’s achieved goes well beyond the level of semantic expressiveness that HTML was designed to provide. For the less ambitious, more pragmatic developer, I would have simply recommended providing a text alternative to the matrix, which could have been done with an image, SVG, or Flash.

Of course, pushing the boundaries of web technology is what we’re about.

On a positive note it is a very useful addition to the site!

The page itself is a technological wonder, making use of some hairy CSS in order to display information in a useful manner while retaining some form of semantic sense.

This is somewhat concerning, not because Brothercake has done a bad job, but because it means doing simple stuff on the web can be really complex. The matrix is good, certainly useful, but it doesn’t look like a technological wonder, or something that should require ultra-advanced CSS. If the degree of difficulty to produce this slightly rich visual representation while remaining semantic is so high then doing things the right way is going to be well out of reach for most web publishers.