I have a bookmarked collection of ‘horrors’: Web pages which really should have been designed differently, many of which are mercifully no longer online. Backgrounds were one bête noire; I shudder when I recall the page promoting Scotland on a repeated backdrop of saltires, for example. A more recent example is the Ure Museum’s display of images of its entire collection (scroll down!) behind the search box. For the full effect, however, one needed to include animation. Spinning geometric patterns on a page about an exhibition of Islamic art (at the Bodleian, no less) or reduplicated Flash clips as can be seen here. The prize was taken by the guest-house in Wales which displayed six simultaneously rotating 360-degree views of each of its rooms, one above the other. It was impossible to look at this without feeling genuinely queasy.

There are of course many ways in which Web pages may be unsatisfactory. Investigating a local antique shop (or rather, the premises it used to occupy) I came across the page at https://www.antique-glawibbless.co.uk/. (To lessen potential embarrassment, I haven’t given it correctly; remove ‘wibble’). It appears to be a very simple page thanking past customers, and while one might think the mix of seriffed and sans-serif fonts a bit last millennium, uncovering the source code is the true revelation. It is 4,963 lines and 427,273 characters long (a mere 401,361 without spaces), that is, about as long as a typical novel. All to display a couple of images (one of which has had the top cropped off) and a message that is 31 words long!

How can this happen? I used the Wayback machine to look at earlier versions of the page, and while they had links to what was then a more extensive site, there seems no justification for the baroque confection of Javascript and stylesheet that lies behind it. The single largest component is a long list of instructions relating to a long string of fonts, including Cyrillic, Devanagari, Vietnamese and other scripts, linking to Monotype Imaging’s software. Wix.com seems to be the immediate source of the code. I think that pages like this survive because they are commissioned by people who don’t look ‘under the bonnet’ and ask awkward questions about why so much code is needed.