Lars Gullik BjÃnnes wrote:
> Andre Poenitz
>
<[EMAIL PROTECTED]>
> writes:
>
> | On Mon, Nov 29, 2004 at 02:36:47PM +0100, Lars Gullik BjÃnnes wrote:
>>> Alfredo Braunstein <[EMAIL PROTECTED]> writes:
>>> | // Find position closest to (x, y) in cell given by iter.
>>> | + // Used only in mathed
>>> | DocIterator bruteFind2(LCursor const & c, int x, int y)
>>> | {
>>> | double best_dist = 1e10;
>>>
>>> 1e10 is not good... (but not your code)...
>>
> | It is mine.
>>
>>> double best_dist = std::numeric_limits<double>::max();
>>>
>>> would be a lot better.
>>
> | Well, better... amybe. A lot?
>
> Well... yes.
>
> It shows that you want to start the valua at max, not at some
> arbitrary number.
>
> | I don't expect monitors with more than 100000 pixel resolution in either
> | direction to turn up soonish..
>>
>>> | + << " best: " << best_dist << endl;
>>> | if (d <= best_dist) {
>>> | best_dist = d;
>>>
>>> best_dist = min(best_dist, d);
>>
> | Why?
>
> At least in my op. it shows intent a lot better.
But note that there is more code inside that if block, so you cannot remove
the "if". And
if (d <= best_dist) {
best_dist = min(best_dist, d);
...
does look silly (to me).
> | If I read such code I always mentally translate it to the other version.
> | So at least for me this is certainly less readable.
>
> I am the opposite end.
> And especially this last one where you can keep w const and initialize
> once.
I agree with Lars here. (still, I think that there are more interesting
things to do than to discuss these details ;-)
Regards, Alfredo
