On Jan 13, 2011, at 4:11, Felix <[email protected]>
wrote:
> From: Jim Ursetto <[email protected]>
> Subject: Re: [Chicken-users] Re: sqlite3 egg patches for chicken experimental
> branch
> Date: Wed, 12 Jan 2011 23:16:21 -0600
>> I am wondering why milliseconds->time is deprecated in the first
>> place. The internal #<time> object stores times in milliseconds
>> itself, so to simulate milliseconds->time, we have to do a floating
>> point division by 1000 just to have Chicken multiply it back by 1000.
>> What is the justification for that?
>
> Do you see this as a performance problem? Somehow, the milliseconds
> operations appear to be redundant, since seconds is the more natural
> (to me) time-unit. Probably this was a mistake.
I doubt there is any significant performance issue; it just seemed redundant to
me to do the unnecessary unit conversions. I assumed srfi-19 included
milliseconds and nanoseconds->time just for this reason, so that the user could
stay in the units he has and possibly gain a (small) efficiency boost.
Although perhaps arguing from srfi-19 is not helping my case. If you want to
remove it I will not cry, too much.
Jim
_______________________________________________
Chicken-users mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/chicken-users
