Fussy check of cell contents

[access_mk]

Hi

Can anybody help me in following task...

I want check/compare two cells containing alpha numeric sentences. If the cells are having more than 3 words in common or the meaning of the cells is same irrespective of the words then it sould return some thing like 'DUP' in the corresponding row
A B C
1 depth of 3 metres depth of 3 metres Dup
7 decimetres 7decimetres

2. daybeacon BnOr daybeacon BnOr Dup

3. the legend of lights F.R. the legend of lights F.R. Not duplicate
35m by Q.R. 35m 6M by Q.R. 7M point (a) eventhough
few words are same

4. obstructions 5.1 Obstn obstructions 3.5 Obstn Not Dup

5. light to, Fl(3)10s17M light to, Fl(3)10s17M Dup (B5 is same
as A5 with additional
info)

so the required function/macro should think and decide the duplication.
Is that possible... if so kindly help me out... check attachement for clear picture of data

[tony h]

Some obvious thoughts are:
1. how accurate does the result need to be? And what is the risk if you get it wrong?

2. I think getting a straight Dup or non-dup is going to be pretty difficult first time. What you can get is an indication on a scale

3. metrics going into the scale could be:
- same number of words
- same exact word
- number of words that occur in both (this differs from the previous in matching multiple occurances of one word in the first phrase to one or more occurances in the second phrase)
- same words in the same order

4. you may find you want to remove certain words (eg the a etc)
5. or want to group words into phrases

your thoughts on this would help in finding an answer