Timing (was wordOffset, repeat loop, speed?

Mark Brownell gizmotron at earthlink.net
Mon Jan 6 15:16:01 EST 2003


Wow!
63 milliseconds; 206 hits; 390345 characters ...looks like I have a  very
fast parser. It even works well with empty space, it gives me exactly what I
needed. The array contains all first character numerical valuse for each
hit.

So what about multiple OS uses for the licensed person of the single pro
version license? -- anyone?

Thanks for the help, "offset()" did the trick.


on 1/6/03 2:46 PM, Mark Brownell at gizmotron at earthlink.net wrote:

> I just did an exact duplicate test in Director using the textCruncher Xtra
> where the same 400 kbyt text doc took 2 ticks to build my array. It also
> handle any combination of characters as the text to find. I will look into
> "offset()" to see if I can get both things working the way I want it. When I
> did this in Realbasic I got almost the same results as textCruncher Xtra. I
> believe that all it is is part of the string class in C++ being added as an
> Xtra, in Director. I hope I can build this fast parsing system becuse I like
> what I have seen so far in RR.
> 
> on 1/6/03 2:24 PM, Dar Scott at dsc at swcp.com wrote:
> 
>> 
>> On Monday, January 6, 2003, at 11:58 AM, Mark Brownell wrote:
>> 
>>> 2.) I noticed that it won't work with multiple words or empty space
>>> between
>>> characters. I would like to work with the numericle offset of
>>> characters and
>>> place that info gathered into an array if possible.
>> 
>> Would offset() do what you want?
>> I would expect this to be faster, since going to the start position
>> will have constant time.
>> 
>> (Your method probably has mc^2 time, call it E ;-), where m is the
>> number of words to find and c is the length of your string in
>> characters.  Using offset() should have mc time.  Why squared?  I'm
>> just guessing, but I assume wordOffset() has to count up to the
>> starting word each time.)
>> 
>> Dar Scott
>> 
> 
> _______________________________________________
> use-revolution mailing list
> use-revolution at lists.runrev.com
> http://lists.runrev.com/mailman/listinfo/use-revolution




More information about the use-livecode mailing list