There are many clever things that web designers can do with Cascading Style Sheets (CSS). However, when evaluating possibilities, I think it is worth checking that they actually solve a real problem without creating significant new issues. Today I have looked at a couple of examples on one of the sources of inspiration I regularly check via Feedly, Hongkiat.

One article looks at creating equal height columns without falling back on the established faux-column trick. It seems to involve wrapping elements in an invisible container and getting them to display as table cells. The implementation of pure CSS columns is still thwarted by cross-browser incompatibilities but, if so much effort needs to be put into making a page display like a table, I wonder if a simple table wouldn’t be better? We have come a long way from the bad old days when tables were the trick used to create a visually pleasing juxtaposition of layout elements. This approach for equal height columns seems to rely on using a different syntax to create… tables.

Another suggests that the number of HTTP requests the browser has to make to create a page can be reduced by encoding the image as a binary block within the CSS file. Surely that is counter-productive? As well as not working reliably on all browsers, it means that the client has few connections to download larger resources. Furthermore, those resources are harder to manage because they now contain blocks that must be byte perfect and are not (safely) hand-editable. I can envisage some situations where the technique might be usable (for example, using client side javascript to perform on the fly recalculations to give, for example a manner of sliding between colours or perhaps making a texture increasingly rough) but, as a vocabulary for standard sites, I don’t think I will take the time to learn it.

If I invest the time in learning a new approach, I generally want to use it to get past a challenge, not to merely go all the way round it an end up back where I started.