P.S
This has my interest as I'm writing a data synchronisation algorithm and
could help add to conflict resolution...
On Thu, Jul 3, 2008 at 12:14 AM, Scott Barnes <[EMAIL PROTECTED]>
wrote:
> Kym,
>
> The math on that wouldn't work. As if Steve is looking for a 98% static vs
> 2% change score, it could throw out irregular results around that. As that
> algorithium is more for repeated word search patterns then actual difference
> in word adjustment / composition.
>
> Steve, As I see it you're looking for the boolean statement, "did they
> change yes/no" vs time stamp, as if they are changing Yours to You'rs then
> in turn this meant 1 character change. Isnt this a case of
>
> if( OldTotalNumWords != NewTotalNumbWords) { do logic here}.
>
> If you however want to display a percentage then what qualifies as a needle
> mover? in my case "Yours to You'rs" how do you weight this accordingly? is
> that a character divided by word count + or - delta?
>
> This is actually an interesting question :) well done Steve :0
> On Tue, Jul 1, 2008 at 4:32 AM, Kym Kovan <[EMAIL PROTECTED]>
> wrote:
>
>>
>> Steve Onnis wrote:
>> > What I want to do it implement like a "Recently Updated" type of
>> function
>> > into an app but I don't want to go by the Date Modified stamp because
>> they
>> > might just be fixing a typo or something in there which doesn't really
>> class
>> > as a major update.
>>
>>
>> I am wondering if a tool we wrote as part of a SPAM checking engine
>> could be put to use here. It used the triad idea (out of IBM IIRC) but
>> working on entire words rather than characters.
>>
>> It parsed a string into sets of three words: 1,2,3;2,3,4;3,4,5; etc ,
>> put each triple word set into a structure and in the case of the SPAM
>> engine it then compared those triplets with an existing structure with
>> values for each triplet to get a score of goodness or badness.
>>
>> If you did that type of process on one string and then repeated it on
>> the other then you will be able to compare the structures and pick out
>> word sets that are different. Then you can look at the differences and
>> score the amount of change in each case, sum them up for a grand
>> difference score and with some suitable scaling is close to what you are
>> looking for is it not?
>>
>> I am happy to share the code. Its all custom tags as its part of our
>> mail server which is running CF5 and it runs fast, it has to :-)
>>
>>
>> --
>>
>> Yours,
>>
>> Kym Kovan
>> mbcomms
>>
>>
>> >>
>>
>
>
> --
> Regards,
>
> Scott Barnes
> Rich Client Platform Manager
> Microsoft.
>
> http://blogs.msdn.com/msmossyblog
--
Regards,
Scott Barnes
Rich Client Platform Manager
Microsoft.
http://blogs.msdn.com/msmossyblog
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"cfaussie" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/cfaussie?hl=en
-~----------~----~----~----~------~----~------~--~---
