Let's have a look at presentations. Who hasn't seen a presentation with a diagram about a bunch of computers and their interrelations? These days, it seems to be the norm to go through one's clip-art collection and paste images of computers in these diagrams. Unfortunately, the fact that a node in the diagram represents a computer is rarely what the diagram is supposed to say (most of the time that's perfectly clear from context), we're usually only interested in which computers play which roles, i.e. which are clients, which are servers (and serving what) and so on. So now people take different pictures for different computer roles. Too bad that there's no widely accepted norm on what a picture of a client or server computer is supposed to look like to make it instantly recognizable as such; typically we'll end up with the server machines looking somewhat more angular, but there's still no way we can clearly tell their roles from the pictures. Most presenters have at least enough sense left to add the role as a legend, so we'll end up with some generic pictures with legends. Brilliant, now you've got your money's worth out of your clip-art collection and wasted tons of space on the page (which is always in short supply, in particular in presentations) and since none of this conveys any meaning at all you'll need additional space for the legends; much better than just using the legends and leaving the clip-art collection for the kids to play with, isn't it? Of course, this sort of thing doesn't just apply to pictures of computers, but practically everything that can be found in clip-art collections. If you want a rough estimate of how much a picture says, evaluate how long it'll take you until the picture is in your presentation: it's highly unlikely that the picture will say more than the amount of words you could have written in that period. So if you just copy something in from your clip-art library, you can definitely count yourself lucky if the picture says as much as a single word, let alone a thousand. Clip-art pictures are by definition generic and therefore convey no information other than that the presenter tried to cut corners. A picture that says more than a thousand words is one that manages to present a complex process in well-structured graphical form and takes serious effort to make, typically much more than just writing the whole process down. Its purpose is not to save time making the presentation, but to help the audience to quickly absorb the process without having to read pages of fine print. If you're using pictures to cut corners, you'll end up with what's called "Making presentations with gimmicks just for the heck of it", or in short "Industry presentations", a format where style is everything and content is nothing. If you absolutely, positively want to go for an honorary MBA, you should also include a garish background (better yet: a high-contrast background image) and pictures of children or generic office staff grinning stupidly in the camera (in particular pictures of attractive women who, in real life, wouldn't want to be seen dead in your line of business). If you prefer to make an actual point, you're better off sticking to the kind of thing that's often debased as "old-fashioned" or "dull", because what that really means is "readable" and "doesn't get in the way of content". If you actually have content, that is...
One of the other big catastrophies resulting from the abuse of the thousand-word-picture theory is current GUI design. I don't know who came up with the theory that putting hundreds of tiny, identically-looking icons on toolbars is good GUI design (probably Microsoft, they got everything else dead wrong too, after all), but apparently it was someone who didn't have the slightest idea what he was talking about. A tool button is only a good thing when it doesn't take you more time to find it and see what it does than it'd take you to read its name; in contrast, a button that looks almost exactly like 20 other buttons in the same toolbar, the meaning of which can only be guessed by putting the mouse pointer over it and waiting for the tool-tip to pop up, is nothing but a waste of space and user time. In particular if the tool-tip is so large that it covers the tool button and 5 of its neighbours and always pops up 2 milliseconds before you try to click the button, intercepting the mouse event. That kind of thing has nothing to do with good GUI design, which is all about ergonomics, this is just mindless repetition (and abuse) of the old thousand-word-picture mantra. If you think there's no alternative because you have so much functionality in your application, you obviously haven't understood what context menus are for. But even plain menus haven't escaped the retarded need for tons of nondescript icons, so good old-fashioned checkmarks that were instantly recognizable at a glance had to go to make room for icon sets representing the same information (on/off) in an orders-of-magnitude less readable way. More often than not the difference between the two icons is a tiny frame or two pixels that are a 10% darker shade of grey, where you can only tell the difference if you've got a magnifying glass and the complete icon legend for the application handy, and even then you've only succeeded in telling the two states apart but still don't know which one's "on". Or they're clearly different from each other, but each is similar to another tool button that does something different. Utter chaos.
Is it really so hard to just cut the crap and leave the fucking pictures out of it unless they actually manage to represent content better than text can? Is anybody really still impressed by stupid "my clip-art collection is bigger than your clip-art collection" and "my Powerpoint has more effects than your Powerpoint" presentations which further nothing but the "art" of saying absolutely nothing with lots of colours? Are contemporary users really so retarded that they prefer spending 5 minutes hovering the mouse pointer over 60 tool buttons, reading tool-tips, rather than opening a well-structured context menu and actually finding and starting the operation they were looking for within 3 seconds? Whether you're writing presentations or software, cut down on the graphics for crying out loud. Because most of the time a picture doesn't say more than a thousand words; most of the time a picture says absolutely nothing.
Back to the manifesto index.