Enhanced search capability

DEVONthink Feedback
1

There used to be a very cool search program, which I think was even based on HyperCard. When you entered a word, it would produce a list of all the documents with that word in it AND list of all the other words that appeared near (within X words) that word in the database. You could then click on one or more of those words and it would limit the list of documents to those containing the focal word and one or more of the selected words within X words of each other. You could keep doing this to get a narrower and narrower set of documents.

As I remember, it also showed a one line excerpt from each document, showing the search terms in context.

Particularly if you have lots of documents (1000’s of emails for example) with simlar content, this can be incredibly powerful and it seems that, based on what DT already shows (Concordance), much of the infrastructure is already there.

2

What you are describing sounds very like “On Location,” which I used steadily in the late 80s and early 90s. It worked through indexing files, which took a long time, but when up to date, you could quickly find every file that contained an instance of a text string. Even better, you could see the context of that string in the file, and also the entire text of the file. It was very fast, even on old processors. The company was in Cambridge, Mass but I can’t recall their name now. They were early Mac developers but went to the Dark Side later on. (This software had nothing to do with HyperCard.)

3

In DT Pro 2.0 you will be able to formulate search queries just as in DEVONagent 2.x.

So you will be able to specify that a term must be NEAR/n, BEFORE/n or AFTER/n another term.

There have been a number of posts in the forum about a view that would display the search term in context of a few words, rather like the Preview search pane. No promises. :slight_smile:

I have fond memories of a group of moderately large databases of water quality literature abstracts that I put together using a proprietary Hypercard-based “book” format. The U.S. Geological Survey used to distribute this material on CDs that were DOS-only, slow and pretty ugly in performance and appearance. I managed to capture all the references for countries with which I was doing environmental science exchange projects. For example, there were several hundred research abstracts about Egyptian water quality studies.

The “book” of Egyptian water quality abstracts looked beautiful, included image and sound files, and performance of searches was instantaneous even on my old PowerBook 170. Better yet, I could keep a “library” of such collections and easily switch them. I recall doing similar collections for Poland, Chile and the Philippines. (The downside was that creating such a HyperBook was slow and tedious.)

Search results were presented in context, and clicking on a result jumped to the relevant page. And a search query could be saved. I did that project in support of graduate student training at the University of Alexandria in Egypt. (Apple Computer donated computer equipment to the University of Alexandria in support of that project in 1993.)

I gave a demo of those water quality literature projects to the U.S. Geological Survey computer folk and it generated a good deal of excitement. I think that encouraged them to make the literature more widely available on the Web instead of just on clunky (and expensive) CDs.

I used to do a great number of negotiations with the U.S. EPA, so I used this same tool to analyze federal regulations. It gave me a considerable edge, as I could often find citations favorable to my side more quickly than they could find contrary citations. :slight_smile:

Some of you old Mac hands may remember that Michael Crichton’s book Jurassic Park was released as a HyperBook. I used the same software that produced that computer version. In many ways the software was more advanced than any I’ve seen in current eBooks.

Of course my DT Pro database on environmental topics is enormously larger and more powerful than those HyperBook projects. But I do miss the contextual line presentation of search results, I’ll confess. Especially as I may need to do some careful parsing of federal regulations for an upcoming project. :slight_smile: