DT for patent analysis

I have a DT database with patents that I keep around for search/compare. When engineering teams consult me for infringement, validity or novelty searches, I frequently consult with my DT database first before going out on google or using DA.

One of the ways to “search” in DT is by using the classify function. All the existing patent matters I keep in folders with names that group like technologies in a manner that makes sense to our business. So when I now use the “classify and see also” function, I get an immediate impression how the new text compares with the data I already had. This is such a time saver, you wouldn’t believe!

However, the classification function operates with all the folders in the database. Sometimes I want to limit its search by folder & sub-folder or by tag(s). Is there anyway to trick DT into doing this?

Although items can be excluded from Classify or See Also as an option in the Info panel, so that you could exclude some items for comparison to your document of interest, that’s not practical for your needs.

A possibility that you might consider would be to copy (not move) selected groups and/or the results of a search to a (temporary) new database. That would allow you to exclude items less relevant for comparison to your document. When finished, delete the temporary database.

Aren’t kludges great?

@korm - aahh, DT still has a lot of “hidden” functions for me. The exclude from classify will help me a lot. How about going “the other way” i.e., is there a way to pick the group or tag set that is included ?

@bill - yeah that is what I have been doing, not great, kludge is the right word for it.

Just as a FYI, excluding a group from See Also and/or Classification (as korm described) only affects the group-not the contents of the group. If you want to exclude ALL the documents of a group from See Also and/or Classify, you will need to select them all and make then make the change on the multiple selection in the Info panel. I mention this as some users will exclude a group from See Also/Classify and then wonder why documents and sub-groups of the excluded parent group still appear as suggestions in the See Also/Classify pane.

Kludges often can save time and effort, especially when they are simple and straightforward. In this case, what is desired is to isolate a subset of documents in a database by certain characteristics, then to compare a specific document to the others using, e.g., See Also, to look for similarities.

The reason I suggested this one is that it takes less time and effort than opening individual group and document Info panels and checking the options to exclude from Classification and from See Also–at least in my own databases. First, I’d have to decide which ones to exclude, among hundreds of groups and tens of thousands of documents. And afterwards, I’d have to remember which groups and documents had been altered that way, and revert them to their original settings.

Would it be possible to automate that process, perhaps involving something like Hazel? If so the procedure would have to be run twice, once to check the exclude options and a second time to uncheck them. I’ll ignore for the sake of argument that some of the groups included in this procedure might already be excluded for other reasons, and it would get more involved to leave them unaltered.

If a subset of groups or documents in a database can be quickly selected for copying to a new database, the kludge may be more efficient, and removes potentials for lingering effects on the original database contents. It depends on the characteristics of the database and on the question to be resolved.

Suppose I forget about groups, and create a search to filter for patents that might be similar to a patent I’m looking at. Run the search, copy the results to a new database, import into that database the patent document I’m evaluating, select it and run See Also. I look at the See Also suggestions. Depending on the issues involved, I might spend minutes or hours on comparisons and making notes.

If I find a See Also suggestion that’s pretty similar, that tips me off that I would file the new document in the same group in the original database, at least as a replicate to that location.

How long does that take to set up this kludge? Well under 5 minutes if I’m familiar with that literature and my database.

@Greg: Thanks - that point had escaped me. It seems natural to me that excluding the group would extend to the files and groups held within. I ran a quick test whereby I had a duplicate pdf in a group called ‘test’ with its setting set to exclude from classification. The same pdf was also in my inbox. Classify and see also properly excluded the duplicate from the search results. Works as expected. So not sure what you meant.

@Bill. Not that I am unwilling to use a kludge sometimes, but in my case I have some groups that always will need to be excluded from the result. In my database, I keep patents and patent related issues (infringements, studies etc.) the latter will always need to be excluded from the classification. Sure, I can set up another db for just the excluded stuff, but I already have a bunch of other databases (for other purposes), why add more? I recognize your solution and will probably apply it from time to time, but would still prefer a better, more elegant solution such as the ‘excluded from classification’ is giving me.

I do have scenarios where I need to tailor the search group further. So your idea to throw Hazel onto it got me thinking… Perhaps some scripting and label setting might help out further.

Hazel is not going to change settings in DEVONthink – Hazel operates on items in the file system, not on items inside application data such as inside DEVONthink databases. I don’t understand why it’s mentioned here.

It is possible to use a Smart Group to flip the “exclude from classification” status from true to false and vice versa. See this tutorial. Use the automation shown in the tutorial with caution – you won’t lose data, but you’ll need to think about what it the process shown in the tutorial is doing. Trial the method on a test database.

Search results only find the first instance of a replicant or a duplicate-the search results never return multiple instances of the same document. In other words, if change your ‘test’ group to not exclude from classification, I would expect that your search results will be unchanged. To test what does/does not get excluded from See Also/Classify, use similar documents instead of duplicates or replicants.

I rest my case. :slight_smile: