SC, Rev,and RB speed test

[David Vaughan]

On 17/04/2004, at 5:36, Chris Yavelow <Chris at yav.com> wrote:
>
> ... In the meantime, I did put some realistic data up on my speedtest 
> page:
> http://www.yav.com/speed.html
>

I took Chris data and examined it. My solution below exploits that 
knowledge of the data but if it is genuinely a typical search set (most 
search material being three words or so) then the following does the 
trick.

Using the algorithm on Chris' web page I process the material in 530 
ticks on my machine (Chris does it in 494). With the code below this is 
cut to 51 ticks so it should run in <50 ticks on Chris' machine, an 
all-time record.

The trick is to exploit "intersect" (look it up) which is amazingly 
fast at, er, intersecting. The intersection itself executes in around 
50 milliSeconds or about 3 ticks. It is all the preparatory work which 
takes the other 48 ticks.

on mouseUp
   -- local tFind,tSearch,tLine,tMatchList, sMS, sTicks
   set cursor to watch
   put empty into field "TheResults"
   put the milliseconds into sMS
   put ticks() into sTicks
   put field "SearchTextList" into stList
   put field "targetText" into tMatchList
   -- stripping quotes from both sets saves problems with single quotes
   --  or with quoted strings being treated as a word
   replace quote with empty in stList
   replace quote with empty in tMatchList
   -- Prepare the target text in all consecutive  triplet groups
   put empty into k1
   put empty into k2
   repeat for each word inWord in tMatchList
     put k2 into k1
     put k3 into k2
     put inWord into k3
     put true into tMatch[k1 && k2 && k3]
   end repeat
   repeat for each line inLine in stList
     if the number of words of inLine = 3 then
       put true into tSearch[inLine]
     else
       -- Do a simple search for groups other than three words
       if inLine is in tMatchList then
         put inLine & return after MatchList
       end if
     end if
   end repeat
   -- Now for the Revolution!
   intersect tSearch with tMatch
   put keys(tSearch) & cr & MatchList into field "TheResults"
   put the milliseconds - sMS && "ms" & return && "(" & ticks()-sticks 
&& "ticks)" & return before  field "SpeedRecords"
end mouseUp

Hope this is interesting, not to mention helpful.

regards
David

> Thanks for the comments,
> Chris