I just downloaded DEVONthink and am trying out its boolean search capabilities. What want to be able to do is find instances of two terms that are NEAR each other, and not see instances that are not NEAR each other, all in one search. Say I use this search:
real NEAR life
I want to see highlighted only instances where those two terms are NEAR each other. I don’t want to see highlighted instances where they are not NEAR each other. I don’t see a way to do this.
Searching for ‘real NEAR life’ will find all instances where the two terms are within 10 words of one another. If you want to tighten up that criteria, change the search to ‘real NEAR/n life’, where n=the number of words proximity.
I find that it finds all documents that have those NEAR each other, but if I look at one of the found documents, I notice that DEVON highlights all instances of the two terms no matter how close they are. I want DEVON to highlight only the ones that are NEAR each other. So far I have found only one search product that does that (dtSearch), but it only works on Windows.
That would be very very helpful
I played with dtSearch and found that the ability to restrict highlighting to only those terms that meet the NEAR criterion can (depending on content) produce very confusing results. In one case, where NEAR/n was fairly large and there were a number of overlapping occurrences of terms meeting the criterion, almost the entire document was covered in highlighting. The dtSearch approach works well for simple cases but there are logical possibilities that can overwhelm it.
Well, I have yet to find a problem doing this with dtSearch during my evaluation period of the product right now. It’s the reason I still have Fusion running on my Mac for research purposes.
Using DEVINthink Pro Office -
The highlighting of all occurrences of a word when using NEAR is a very bad thing and completely non-logical. This makes a search for
term1 NEAR term2
actually return
(term1 OR term2) WHEN text1 CONTAINS (term1 NEAR term2)
If I search a cookbook for the words “dough NEAR chocolate” I don’t want it highlighting the five hundred occurrences of “dough” which are not near “chocolate”. If the term “dough” gets highlighted no matter what, then the NEAR function is absolutely worthless.
In reference to Bill_DeVille’s comment about dtSearch: the text between the search terms needn’t be highlighted at all (which is the way DT seems to handle it), which would obviate the problem of large areas being highlighted.
PLEASE FIX THIS!
By the way, this is otherwise a really great program guys!