The purpose of &> is to unbox atoms of the operands.  If you don't
need to do that, "0 is better.

In contrast, u&.> was better than  <@u"0 even for unboxed
operands, last I looked.

Henry Rich

> -----Original Message-----
> From: [EMAIL PROTECTED] 
> [mailto:[EMAIL PROTECTED] On Behalf Of 
> Richard Donovan
> Sent: Thursday, June 12, 2008 9:53 AM
> To: Programming forum
> Subject: [Jprogramming] "every" adverb
> 
> 
> I have been playing around with the adverb "every" supplied 
> with the system:
> 
>    d=:i.10
>    < d
> --------------------┐
> │0 1 2 3 4 5 6 7 8 9│
> L--------------------
>    every
> &>
>    < &> d
> --T-T-T-T-T-T-T-T-T-┐
> │0│1│2│3│4│5│6│7│8│9│
> L-+-+-+-+-+-+-+-+-+--
>    < "0 d
> --T-T-T-T-T-T-T-T-T-┐
> │0│1│2│3│4│5│6│7│8│9│
> L-+-+-+-+-+-+-+-+-+--
>    d=:i.1e5
>    (< &> d) -: (< "0 d)
> 1
>    ts
> 6!:2 , 7!:[EMAIL PROTECTED]
>    
>    10 ts '< &> d'
> 0.426803 1.50782e7
>    10 ts '< "0 d'
> 0.0517512 7.33485e6
>    
> 
> It seems that "every" forces the preceding verb to take on 
> the ranks of verb> (open)
> i.e. 0 0 0 
> 
> As you can see from the ts tests I ran, it seems that "every 
> &>" makes a serious time
> difference to the expression, and also uses more space.
> 
> If the only purpose of &> is to force the verb ranks to be 0 
> 0 0 why not just use the
> expression    v "0 which is not only more obvious as to what 
> it's doing but also seems
> much quicker and leaner?
> _________________________________________________________________
> Great deals on almost anything at eBay.co.uk. Search, bid, 
> find and win on eBay today!
> http://clk.atdmt.com/UKM/go/msnnkmgl0010000004ukm/direct/01/--
> --------------------------------------------------------------------
> For information about J forums see 
> http://www.jsoftware.com/forums.htm

----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to