On 30/09/2011 12:01 PM, Invisible wrote:

>
>>> What I actually /said/ was that computers (by which I mean fully
>>> autonomous computational devices)
>>
>> What do you mean by "fully autonomous computational devices"?
>
> Well, that's the killer, isn't it?
>

AI, I suppose.

> An abacus can add. But only if an intelligent human is operating it. All
> the smarts are in the operator; the abacus is just a memory tool, really.
>
> A pocket calculator can add. Even if it's operated by raindrops getting
> the keys. The smarts are in the device.
>
> Now, how the heck you formulate that into some kind of coherent
> definition......
>

When you do, submit it as a PHD thesis. ;-)

Actually I don't think that I agree with your comparison between the 
abacus and pocket calculator. I think that the latter is just a better 
type of the former.


>> I am not familiar with ...
>
> The number of steps required to find what you want is proportional to
> the logarithm of the size of the thing you're searching.
>
> If you search for a definition by looking at every page in the book
> until you find what you're after, the amount of work is obviously
> directly proportional to the size of the book. On the other hand, if you
> look it up in the index and then go straight to that page, the amount of
> time is proportional to the logarithm of the number of pages.
>
> In the former case, if you double the number of pages, you double the
> time required to find anything. In the latter case, doubling the number
> of pages has little effect on the lookup time. (Assuming the book was
> already large to start with, of course.)

Well put, that completely reverses what I thought you meant.


>
> I wasn't trying to be funny. I was trying to phrase a definition which
> would encompass things that don't look like modern hardback "books".
> Things like rolls of parchment or stone tablets, etc.
>

Fair point considering. :-)


>
> I'm well aware that there have been many computers that used decimal
> instead of binary. (A few even used other number systems.)

I've worked on at least one system that used BCD (Binary Coded Decimal)


>
> Now, you need to be a little bit careful here: There is a distinction to
> be made between devices that can "calculate" things, and devices which
> are considered to be "computers". As far as I'm aware, none of the old
> "analogue computers" were actually Turing-complete, and so strictly they
> are /calculators/ rather than /computers/.
>



alter its input or operating instructions or what?

for each problem they needed to solve. But it could be done.


> The history of /calculators/ obviously goes back far, far beyond the
> history of /computers/. (Again, depending on how general your
> definitions are.)

Indeed.

-- 
Regards
     Stephen

Post a reply to this message