Re: Interesting article by Joel Spolsky: The Perils of JavaSchools


Willem wrote:
> wooks wrote:
> ) But "ordering" the list isn't going to resolve rounding errors is it.
>
> The list *is already ordered*. The nitpick is not that you are going to
> resolve the errors, the nitpick is that if two lists with the same members
> have a different ordering, then summing them could give a different result.
>

so if you are going to have rounding errors regardless why does it
matter???


> ) Whichever order you process the list you are still going to have
> ) rounding errors.
>
> But they will be different. Which is the whole point.
>
> )> This seems like a key point -- the ordering exists, but may not be
> )> significant.
> )
> ) ...and if you abstract away the insignificant bits from your
> ) definition, you simplify and lose nothing..
>
> What part of 'may' did you not understand ?

On the contrary methinks you have a problem understanding the concept
of insignificant.

> You can only abstract it away if it were *never* significant.
> By your very own example of reversing a list, it is obvious that
> in some cases, it *is* significant.
>
> Furthermore, your definition did *not* abstract away the ordering.
>
> And finally, if you take a definition of 'list' and take away the ordering,
> what you end up with is the definition of something else that is *not* a
> list. My best name for it would probably be a 'bag'.
>
> Almost all real-world examples of lists have an ordering to them,
> whether it is used or not.
>

http://en.wikipedia.org/wiki/Abstraction_%28mathematics%29

PS Pay very close attention to the very 1st sentence.

.

Relevant Pages