Neil Hopcroft

A digital misfit

I’ve just been catching up on Creating Passionate Users – Is /your/ work remarkable? I’m not sure mine is, but I would like for it to be. That article describes something I’ve been thinking for a while, it asks you to ask yourself how you can make the life of those who use your products better. They limit their thoughts to books, technical manuals. But why? Theres no reason why your camera, your phone, your microwave shouldn’t be making you a better photographer, listener, cook.

Think about how power-users will benefit from what you are creating, then give everyone who uses it a chance to be a power user. This isn’t about a paperclip that tells you stupid things, its about making intuitive interfaces that make it easier to use the product as you use it more – what if you had predictive text that learned that you said ‘pub’ more often than you said ‘rub’, or better, what if it could figure out from the context what you meant? And without dropping you in at the deep end with the power of VI behind its arcane sequence of keystrokes.


8 comments

  1. Predictive text that learns like that is a fairly simple machine learning problem really, isn’t it? I am frequently surprised that it’s not been done (if indeed it hasn’t). And although guessing the right word from the context is a lot harder I think currently Natural Language processing technology is up to getting it right significantly more than chance…

    • I think the difficulty is one of remembering the users past behaviour – the dictionary can quite efficiently be stored as a variant of RLE in ROM, with any words the user adds in a hash table checked after the main table – hence words you add are the last it shows when changing words – but how do you index the usage counts to the RLE table in ROM and the hash table in persisent storage?

      The problem itself isn’t difficult, but the power and storage contraints of mobile devices make it impractical. Though I would expect that to change over the next few months.

      In fact context guessing is probably an easier problem to solve – the general case can be covered easily with a dictionary of a couple of hundred rules, plus a few exceptions for the weird things in the language. Of course it does lead to the potential amusement of completely incorrect sentences occasionally being sent rather than just individual words, thats gonna be a lot more difficult for the recipient to parse.

  2. But not everybody wants to be a power user. There are (or so I believe) still some people out there who just want to use their mobile phone to make/receive phone calls. I think steveandabilgail’s suggestion of hiding half (or maybe that should be 90%?) of a product’s available features by default has a lot to recommend it. Not so sure about charging for an upgrade key – after all, the additional features are already there, and the real extra cost to the user is their investment of time in getting to grips with the features (even if they’re intuitive to use, you’ve still got to familiarise yourself with the possibilities).

    Creating multiple interface modes for a device should be quite easy if you’re talking about electronic interfaces, not so straightforward for mechanical ones. Call me old-fashioned, but I much prefer the clockwork timer on my ageing microwave to those nasty modern contraptions where you have to work out how to program in the number of minutes required. I’d argue that the former is a much more intuitive interface, and meets the needs of the vast majority of users. Also, give me a rotary volume control over fiddly up and down buttons any day [1].

    “Theres no reason why your camera, your phone, your microwave shouldn’t be making you a better photographer, listener, cook.”

    Can a piece of equipment actually make you a better x, y or z? Well, possibly yes, but not just by compensating for your shortcomings. A hi-tech camera can certainly help you take better photographs, but that doesn’t make you a better photographer. You could, I guess, have a camera that acted as a tutor – for example, rather than automatically adjusting the exposure, it could alert you to the problem and guide you towards the solution. Of course, a lot of people wouldn’t be interested in this, as their goal would be to “take better photographs” rather than “become a better photographer”. But I think the distinction matters – I don’t want to see skill and craftsmanship devalued.

    So, yes to intuitive interfaces (and a big yes to clear, comprehensible, well-laid-out technical manuals), but if a tool is going to try and second guess my intentions, it better make a damn good job of it, or it’s just going to end up irritating me. Which is not the route to a satisfying user experience.

    [1] That would be on the stereo of course, not the microwave ;-)

    • hiding half of a product’s available features

      Agreed, though I suspect the interesting thing is how the hiding, or probably actually the unhiding, happens.

      I much prefer the clockwork timer on my ageing microwave

      Theres no reason why you couldn’t have that interface onto a much more accurate timing system. Though there is an argument about whether that is necessary – I always got annoyed by the dial on one microwave that wouldn’t reliably do ten seconds, you could get a minute or more but any less and ping, it just thought you were turning it off.

      “take better photographs” rather than “become a better photographer”

      Improving skill is not devolving responsibility to the machinery, the trick is to get the machine to be /able/ to teach you if you want to learn. Firstly if you take better pictures with one brand of camera than another you’re more likely to buy that brand again (or chose that brand in the shop), if you feel your camera helped you do the things you wanted to do better then you’re going to be happy with it. If you’re happy with it you’ll use it more.

      clear, comprehensible, well-laid-out technical manuals

      …and fun, they’ve got to be fun…oddly this is something that some of the cheap Korean brands achieve accidentally with their poor translations, but if its deliberately built in to the pictures or the layout then thats better, I seem to remember Panasonic being quite good for this, while a lot of others are rather dry reading. (but its been a while since a read a manual…)

  3. Thats the thing – you shouldn’t have to ask it to tell you, it should know. One of the few interfaces I can think of that did this sensibly was Wordperfect, there was a menu system with commands, when you started typing the command it would show the menu if you were typing slowly, or if you knew it you could type it quickly and it wouldn’t bother displaying the menu. This meant that you could learn and quickly use the commands you used a lot while still have accessibility to the commands you rarely used.

    I don’t think I’ve seen anything do this quite so well since, though there clearly has been some innovation in UIs since the DOS text interface that was built upon.

    Some cameras are better than others, mostly the interface is constructed in terms of how things are implemented, not in terms of how they’re used. Its difficult for users to get sensible mental model of how things are going to behave. Creating enchanting interfaces is about giving the user the mental tools to understand how the device will behave without having to understand how it works.

  4. The point is that this type of user doesn’t care about interface design

    This type of user doesn’t need a computer, they don’t have the kinds of problems that computers help to solve.

    This is one of the problems with general purpose computers, people think they need them, when actually they don’t really have a problem to solve with them.

Leave a Reply

Your email address will not be published.