Classify from the global Inbox

I’m looking seriously again at DevonThink Pro Office for document management. A lot of the attraction is that the AI classification means less reliance on remembering my filing scheme (do I put car insurance documents under ‘insurance’ or ‘vehicles’ or ‘bills’?)

Like a lot of other people on the forum, I was puzzled that DevonThink doesn’t offer any classification suggestions for documents in the global inbox (also like many I thought it just didn’t work). I’ve now read several posts on this, so I understand that this is apparently by design.

This makes the workflow I would like very clunky. I would like to have a few databases - for example:
Business
Personal and household stuff,
Research projects

I scan all paper docs and correspondence to an inbox folder, and classify from there. There would be an extremely good hit rate for AI classification of many of these scanned docs across devonthink databases (for example invoices, bank statements, research papers). But if I have multiple databases I have to do a manual first pass to decide which database to put them in, move them to that database, and only then can the AI work. This seems to me to almost entirely defeat the purpose, and I am struggling to understand why it is thought to be desirable.

Why not have at least an option to allow DevonThink to attempt to classify documents from the global inbox across databases? Why would that have any disadvantage whatsoever?

It seems to me I have the choice of multiple databases, global inbox, and manual classification or avoid the global inbox, have a single database and AI classification.

Presumably the workflow I am describing does not fit the DevonThink paradigm. What am I missing? What would be a more suitable workflow?

Thanks,

Martin.

The issue of whether the AI features such as Classify and See Also should remain restricted to individual databases, or should operated across databases has been discussed a number of times, and remains under consideration by the developer. There are pros and cons.

Personally, I prefer their operation to be within individual databases, and not across databases. For example I have two large databases that contain many thousands of documents related to environmental management, policy and regulatory issues. One of them includes references about pollution cases, toxicology and health issues, policy issues and relevant regulatory and statuary issues. The other contains methodological references: sampling methodologies, sample analytical methodologies, data evaluation methodologies, quality assurance issues and cost/benefit and risk assessment methodologies.

The utility of See Also and See Related Text, as well as the Classify assistant is greatly improved by restriction to the appropriate database. Obviously, if my research interest is, e.g., health effects on humans of mercury contamination in edible fish, See Also should not bring up documents specific to sampling or analytical methodologies related to mercury, but be focussed on cases, regulatory limits and associated issues, etc.

I have a number of other databases, each of which meets a specific interest or need.

I don’t find it difficult to add new content to these databases, for filing purposes.

For example, I created a database just for the purpose of handing new OCRed PDFs resulting from scanning documents. In that database I created a number of smart groups to help automatically associate documents that share searchable characteristics. While not perfect, this greatly reduces the drudgery of subsequent filling of new documents into their appropriate database, or even appropriate groups within that database. By moving the contents of a smart group into the Inbox of the appropriate database, Classify becomes useful to aid in filing to appropriate groups.

Other cases in which I add batches of new content to the Global Inbox can be handled in this way: Select a new document related to a certain database from the list, then walk through the list looking at other documents, Command-click on others to be sent to that database, then drag the selected documents to the Inbox of the desired open database via Tools > Show Groups & Tags.

A future release will support this.

Thanks. Any guidance on when that might be?

Definitely not in 2016.

Hello! I was wondering if there has been any news on this topic. Is classifying from the global inbox something that might come this year? (A optional setting would already be enough for those who want to play with it)

I am in a similar boat where I shovel everything into the global inbox and then process it with OCR.
From then I want to quickly move things between databases: Work, Personal, Notes and a few more.

When moved in the correct inbox, the classifier is almost always spot on, so I think allowing it to also classify from the global inbox should work great. It’s currently also the fastest way to move things into the right place without mouse (with a ‘quick move’ dialog not available).

While the option may be part of a future release, we do not give timeframes on development, as it’s too dynamic a thing. Thanks!