[REBOL] /dup and /deep Re:(2)

2000-03-22 Thread sqlab

> [EMAIL PROTECTED] wrote:
> >Is there a refinement /deep for word with the refinement/dup ?
> >
> >If not, I think, that /dup as refinement for words like insert should
> be
> >omitted in favour of the refinement /dup for the word copy.
> 
> Boy would that increase time and memory usage!
> 
I use that to build a hierarchical message skeleton with repetitions of
blocks inside of blocks with empty data (null values). The position determines
the meaning. Without copy/deep I could not insert or change certain parts
without changing others. Storing the position as an further element of an
item would require an extra coded renumbering after insert or delete. 

> But, if you want to do that, to replace a theoretical
> insert dest copy/dup/deep src num
> with
> loop num [dest: insert dest copy/deep src] dest: head dest
> which should be about as fast.

Of course I use my own comparable routine.)

> 
> Perhaps you should consider a copy-on-write strategy as a time
> and memory saving measure. I know that many don't think about
> such things today, but copy/deep takes time. Time and memory
> considerations are even more important in interpreted code.
> 
> Brian Hawley
> 

If I want faster execution instead of easy development, I am using a C
compiler or an other adequate tool.


regards
AR

AR


-- 
Sent through GMX FreeMail - http://www.gmx.net




[REBOL] /dup and /deep Re:

2000-03-21 Thread brian . hawley

[EMAIL PROTECTED] wrote:
>Is there a refinement /deep for word with the refinement/dup ?
>
>If not, I think, that /dup as refinement for words like insert should be
>omitted in favour of the refinement /dup for the word copy.

Boy would that increase time and memory usage!

But, if you want to do that, to replace a theoretical
insert dest copy/dup/deep src num
with
loop num [dest: insert dest copy/deep src] dest: head dest
which should be about as fast.

Perhaps you should consider a copy-on-write strategy as a time
and memory saving measure. I know that many don't think about
such things today, but copy/deep takes time. Time and memory
considerations are even more important in interpreted code.

Brian Hawley




[REBOL] /dup and /deep

2000-03-20 Thread sqlab

Hello

Is there a refinement /deep for word with the refinement/dup ?

If not, I think, that /dup as refinement for words like insert should be
omitted in favour of the refinement /dup for the word copy.

By the way, I did not get one message since two days.
So this is just a test too.

AR

-- 
Sent through GMX FreeMail - http://www.gmx.net