Hello
I do not understand the idea of graphical window by digest.
At moment for me it’s only a result table in an other illustration.
But I think that wrong and I can use this feature better if I understand the idea.
- Searching to a word
- Results in topics
- If I choose one topic from the list this topic is black coloured in the middle of the graphical window.
- In the graphical window there are additionally links to other results
- Some results are in orange or blue colour
WHY ?
- If I move with the mouse over an other result I see that some connection lines are highlight. But this lines highlight lines only to 1 to 3 other results.
WHY ?
- Have the distance lines from top result to the other orange or blue results a significance ? I mean less distance more relevancy ?
Sorry for my english.
Jochen (.de)
Now I have read something in DA Help and understand more about the mind-map style.
But I think the AI of DA is not so brainy.
In german I search Internet fast search to.
Fuss AND Verstauchung NEAR Behandlung
In English
(feet AND sprain NEAR therapy)
In left window I see 25 topics.
And lot of results in the mind map windows are stupid like:
print
valvular transplant
justice
broadcast
shop
prostata
disliked
…
Maybe I’m to stupid for use of Boolsche Algebra.
I use this not so long.
I will do the same search in google to see the result.
Jochen (.de)
DEVONagent’s AI is looking at the terms used and their patterns of use in the text of the found pages. The “Topics” list is derived from that analysis, not from the terms, operators and syntax of the search query.
The graphical presentation in the Digest view dynamically presents the patterns of use of terms in all or in subsets of the found pages. Note that you can also enter a term that’s not in the Topics list; if that term is present in one or more of the found pages the graphical pattern and the set of digests that contain that term will be displayed.
Is that AI “smart”? Obviously, in analyzing the textual content of the found pages it’s doing some interesting things. It presents a graphical network of the relationships of terms in some or all of the found results.It allows you to click on any of the graphical elements and see a new focus of relationships.
Is that useful to you, the user? That will depend on your interests as well as the content of the found pages. It’s an interactive tool that allows you to “play” with the usage of single terms in that set of search results. It’s up to the user to decide whether that “play” produces something that’s useful.
Example: Suppose I’ve pulled down a set of results related to a scientific discipline. If I enter the last name of a a scientist who is mentioned, I may find (depending, of course, on the content of the pages) tips about that person’s role in the historical development of the field, his funding sources and the institutional setting(s) in which he worked. Enter a different scientist’s name and different information might be presented. Is that kind of information useful? Depends on my interests at the time.
Thanks for your answer.
I will read it but my english translation need time.
I’m not so perfect in english.
Jochen (.de)
In my search example:
Fuss AND Verstauchung NEAR Behandlung<
the results of DA and AI are not so clever, I think.
I would see results to do something against my problem.
Maybe I’m search wrong ?
I use DA not so often and it can be that I must more work with this app.
But my search is a practical example for me at the moment and your example:
…Example: Suppose I’ve pulled down a set of results related to a scientific discipline…
is a theoretical example.
Jochen (.de)
Ah, languages.
For grins, I used Babelfish to translate your search query into English. This was the result: “Foot AND verse dipping NEAR treatment”. I suspect the meaning of one or more terms was scrambled in translation. 
The example I gave, while theoretical because it was generally described, reflected some actual searches I’ve made, especially one dealing with management of Malaria.
The AI routine is merely looking at the terms used in the search result documents, both individually and collectively.
The AI routine isn’t trained in any discipline. But it’s clever at pattern recognition, and can do some of those types of analysis much more rapidly than a human. It’s up to the human user to determine whether any of the relationships proposed by the AI are significant. Because the visual display in DEVONagent works on much smaller data sets than does the See Also AI routine in my large main DT Pro database, I don’t find it as useful as See Also. But sometimes it does suggest something that gets me thinking. That’s what it’s all about.
Jochen, I just tried your query:
Fuß AND Verstauchung NEAR Behandlung
Please note that Fuß (leg) is spelled with an ‘ß’, not ‘ss’ in German. ‘Fuss’ is the English word for ‘a display of unnecessary or excessive excitement, activity, or interest’
The resulting topics were: Fuß, Behandlung, Verstauchung, Bänder, Verletzung, Arzt, erfolgt, Umknicken, Bluterguss, Schwellung, Verletzungen, Gelenks, Schmerzen, Stelle, Schützen, Kühlen, wichtig,… which sounds more or less like what you were looking for.
Side note: In the Settings drawer (opens with the toothwheel toolbar icon) I have unchecked ‘Ignore Umlauts’.
Hello Eric
Thanks for your feedback and your help.
It’s your goal 
( I think that’s a phrase in english ?)
Maybe I’m to impatient with the app.
Jochen (.de)