Sam's infrequently-updated cabinet of curiosities
Friday, 06 January 2006

Network Ten Hates Everyone

It's most likely that they do not, in fact, hate everyone, even if they do routinely torture millions with Big Brother and Australian Idol. But if they aren't hell-bent on aggravating as many people as possible, their online TV guide is positively remarkable. It's common enough for website operators to hate the blind, which is still true here, of course -- the whole program listing is built from generated images with no alt text. (And, incidentally, it might be a violation of the Disability Discrimation Act 1992.) But they're not content to foil only Lynx users and the disabled, oh no. It's quite ingenious, actually, in an evil kind of way:

<td style="background:url(/pgutil/getbg.ashx?m=std&id=1227119);">
    <a href="#" onClick="window.open('episode.aspx?episodeID=1236691'); return false;">
        <img src="/pgutil/getfg.ashx?m=std&id=1227119" />
    </a>
</td>

There are two images, a PNG background and a GIF foreground, both dynamically generated. Here's one background:

background example

And its corresponding foreground:

foreground example

Each one alone looks like shaped static, but lay one over the other and presto! it's The Simpsons at 6 o'clock.

combined example

It's not only unreadable if images are disabled -- obviously, as there's nothing to read -- but the image overlay hack leaves it unreadable without CSS, because the background image isn't shown.

The link to the program details requires JavaScript, so that's another possible point of failure. Let's not forget the frames, either, or the requirement for cookies, or that the links only require JavaScript because they create popup windows. Or that the use of images creates other problems, like the lack of differentiation between visited and unvisited links and the inability of users to change the font size.

Why

Why?! I can think of no good reason. Not a single one.

The overlay does make for a ghetto fade-in effect when images are slow to load, but the only reason for slow loading is the inexplicable choice of images requiring literally a hundred times as much bandwidth as the text they represent.

Strangest of all is that the text is Verdana. Using generated images to embed elaborate fonts is understandable, but Verdana? On the off-chance that a user doesn't it, it'll be substituted by something equally dull; and no-one but a typography nerd is going to notice, let alone care. It's like embedding a HTML renderer in a Java applet embedded in a HTML page: what's the point?

It could be to foil screen-scrapers -- admittedly, I only noticed because that's what I was trying to do. Not that I see the utility in preventing free publicity, mind you, but it's the best I can come up with. There's still no obvious reason for using the two images, since combining them is trivial compared to OCR.

There may be a better reason, but like the saying goes: never attribute to incompetence that which can be explained by malice. So I'm think I'm safe in my assertion that the network hates anyone with a text-mode browser, and the blind, and anyone else with poor eyesight, and anyone with a WAP phone or WebTV, and anyone with JavaScript disabled, and anyone who likes to deep-link, and anyone using Netscape 4. (I'm with them on the last one, actually.)