Seeking annotation-organizational approach for DTP

There is a lot of pushback against tagging so I think its worth reviewing when it is useful.

The important distinction to be made is that folders are for storing documents and tags are a way of storing metadata about those documents.

The purpose of adding metadata to a document is to make that document findable and classifiable in another context. If you are going to tag you have to have a clear understanding what those contexts are. Once you understand these contexts then things like tag categories and auto-complete of tags help keep the contexts consistent.

There are two approaches to adding metadata with tags:

  • You know actually what metadata you want to add to the document before you even start. There are obvious examples of this sort of metadata such as the date the document was created, the author of the document, the addressee of the document and so forth. In much the same way as I would not want a document in a database that did not have meaningful title, I would not a document in the database which reflected an interaction between people unless it properly reflected the sender and receiver as metadata.
  • You dont know what sort of metadata may be useful before you start working through the documents. A common example of this is the type of document. After reading through a bundle of document you realise that they are composed of invoices, handwritten notes, typed transcripts. etc. and you decide that these are useful categories to add to the document. This is the sort of metadata you might choose add on a second review of the documents

It may be that the type of work you do does not require metadata at all. For less than a 100 documents, the visual representation of the layout of the documents in DT combined with ones own memory means its almost not certainly worth adding any metadata. If you are dealing with documents that use clear and distinct terminology then the AI may also be good enough for you. Where however you are dealing with thousands of documents, meticulous metadata can be invaluable to finding connections or even in locating documents or annotations which you can only identify by a single feature.

Adding metadata can seem unproductive and boring and frequently is. I think if you have many documents and the resources it is worth employing someone to add the metadata for you. Its may not be directly worth your time to add the recipient, addressee and date to several hundred pieces of scanned correspondence but when its done for you, its invaluable. You can process the documents and see the connections between them in half the time it would otherwise take. Instantly you can choose to see only correspondence between two people of interest without laboriously searching through the documents. Expand this to thousand of documents and the benefits only multiple. If you export your notes to Aeontimeline (or I imagine something like Tinderbox), tags groups are the only real way to export the context of the notes.

I have written about tagging and metadata in other contexts which may be interesting if you are new to the topic:

Using annotations to build a chronology
Tags and folders are different

Whether tagging is useful to you will depend on what you need from your workflow.


(As a side note, in the legal field coding, as tagging is termed, has well established hierarchies and methodologies (Unitisation, Subjective, Objective, Physical, In Text) and whole software products in the windows world are built on making this process efficient and consistent. The procedure is that you code the first hundred documents and then the AI processes the remaining thousands of documents and codes them automatically for you (based on its interpretation of your initial coding), presenting you with only those that it has a low confidence about its code assignment, for review. The AI in these products is truly impressive but then so is their price!)