[3m"When I can just say to a computer what'I want done and it does it, then I'll'consider it's finally time to buy a+computer."
[0m How often have you heard'that or similar from those not wanting'to learn how to use a computer? Often I'bet and the irony is that if such a'machine existed, they'd just hate using'it anyway. What's more, it's very easy'to demonstrate such a machine, either'conceptually, as in this article, or in'practice. You don't believe me? Then
read on...
'What these people want of course is'some sort of secretary in the computer'who will listen to what they have to'say and then translate this into the'results they desire. To achieve this'would require some pretty sophisticated'AI, but as this isn't quite ready yet,'we can simulate it very successfully by'the simple solution of plonking a real,'human-type secretary down in front of a'decent, present-day computer. This'secretary would need to be a wizz with'the computer and a good range of the'software available for it, but many'such people exist. (Even if one would'prefer not to be in the same room as'them.) Thus we now have a computer that'can listen to the determinedly computer'illiterate and do their bidding on
command. So far, so good.
'But what would happen in practice? To'use this system to write a letter would'probably work reasonably well. Just'dictate it and our secretary will type'it on a wordprocessor, tidying it up as'they did so. But then, what if some'clip-art was wanted amongst the text?'No problem I guess, even if it had to'be drawn by the secretary. But what if'the user (for want of a better word)'had a specific scene in their head they'wanted drawn? Not so easy now, is it?'In fact, anything specific on a'computer is usually a lot easier to do'than describe. And that's the crunch.'Our user would after a very short time'be saying "Here, let me do it." Which'is a lot easier than saying (and'getting right) "Place an ellipse on the'screen at 324 pixels across and 122'pixels down with a width of 37 pixels,'a height of 97 pixels and using an'orange that's quite bright and more
towards red than yellow."
'That said, this scenario would make an'interesting experiment. How about'having a computer screen in a room'along with a video camera facing the'user and in another room one (or better'still, many) computer experts repre-'senting our secretary. These people'could watch and listen to our user'through the video camera's output and'they would then have to try and achieve'what the user requested to the best of'their ability. Telling the user the'setup was a sophisticated AI program'might also be an interesting avenue of'exploration. Either way, it would be'interesting to see just what computer'illiterates would come up with given'their ideal way of using a computer. I'suspect computer literates though would'go nuts very quickly as they fell into'the trap of instructing the computer in'the same manner they're used to using
them.
'However, to get back to my original'argument, I don't believe talking to a'computer will ever be a viable way of'using one, no matter how good its'software. And even if one was made'workable, it wouldn't be popular due to/the fact to
[3mbe
[0m workable it'd have to'be making 90% of the value judgements'itself, and most users wouldn't be'happy with that level of lost control.'And I suspect the determinedly computer'illiterate would be the least happy of
(.txt)