For the first little more than a year of my Ph.D. project, I have been gathering a lot of background material, references, documents and notes, all of which is collected and stored digitally; this is structured with My Digital Workflow.

As everything is handled digitally, I don’t have stacks of papers and binders with different colors to organize my material, but instead have folders in my Dropbox, file naming conventions, notebooks and stacks in Evernote etc. to stay organized. These systems allow me to easily locate files and documents in Dropbox and notes in Evernote and therefore constitute an important part of organizing my digital material.

But folders and stacks don’t always provide the ideal overview and context; for example, I name all my references according to the last name of the first author of the reference (more on this here), and then in my Dropbox organize references according to the first letter. So, for instance, references by first authors Chen and Cao go in the same folder. However, these references don’t necessarily have anything in common, except for the first letter in the last name of their respective first authors, and the organization in folders, while convenient, does not provide sufficient overview and context.

This is where indexing, or tagging, becomes important. Tags provide context and information on content in documents, notes and references and may thus be used for searching and sorting material to provide the overview that folders don’t necessarily give.

I have recently streamlined the tags that I use for my references and notes and in this post share some thoughts on my tagging system and how I have implemented it. I believe this will be inspirational for Ph.D. students and others working with long-term (research) projects that require substantial amounts of background material and literature.

Conclusions

My recommendations for a tagging system, which I elaborate in the following sections, are the following:

• You Don’t Always Know the “Right” Tags from the Beginning: Don’t be afraid to wait implementing a tagging system until you have an overview of your project.
• Choose Tags that Add Value and Context to Your Work: Don’t just tag for the sake of doing it; choose and apply tags that make your work better and more efficient.
• Use Your Tags Consistently: Write a list of the tags you intend to use, and don’t invent new tags unless you can add it to at least a handful of references.
• Invent Your Own Tagging Conventions: Define your own conventions for applying tags – and stick to these as you go along!
• Index Everything You Read: Make sure to index and potentially tag everything you read; you might need it later.

Finally, I provide some details on JabRef, PocketBib and Simplenote that are the software and services I use for managing references and notes with their associated tags.

You Don’t Always Know the “Right” Tags from the Beginning

When I started my Ph.D. project in 2012, I had an overall idea of what my work would consist of and what kinds of systems and structures I would get to investigate during the project. But I obviously didn’t know all the details of the work to be done and of the research field my work would be part of, and I therefore didn’t have a detailed overview of how to structure my references. In short, I didn’t initially know the “right” tags.

In my literature database, I therefore started tagging the entries as they were added to the database on the fly. I tried to maintain an overview of the tags I used, but hadn’t thought my tagging system thoroughly through – because I didn’t know the “right” tags at the outset!

Choosing the “right” tags requires an overview which I didn’t have when I started my Ph.D. project. If you have a detailed overview of the work you will be doing right from the beginning, it should be possible to implement a structured and consistent tagging system. But if not, which is probably the situation in most newly started projects, don’t be afraid of not knowing the “right” tags; when you get the overview, you can implement the tags that work for you.

Choose Tags that Add Value and Context to Your Work

Once you have the overview of your work and project, it’s time to choose tags. Tags should be chosen to add value to your work; don’t just tag your references for the sake of doing it, but because they are a tool for searching and structuring your material.

Also, don’t necessarily tag all references and documents; it’s better to not tag a reference than to invent a tag to make sure everything is tagged. My rule of thumb is that if I don’t add a tag to at least a handful of references, the tag is too narrow and shouldn’t be introduced.

Use Your Tags Consistently

When you start tagging your references, make a list of all the tags you intend to use beforehand. If, in the process, you realize to have forgotten an important tag, you should of course just add it to your list. But having the bulk part of the list of tags before starting to tag the material ensures overview and consistent use of the individual tags which will ultimately improve the quality of your tagging system. My tags, that I keep in a note in Evernote, can be seen in the picture below.

Invent Your Own Tagging Conventions

Another important aspect of tagging is to use some predefined tagging conventions; without such conventions, tags might be assigned inconsistently, and this lowers the quality and applicability of the tagging system.

In my list of tags, I have some central tags, like fmm, phc and plasmonics, that define subclasses of tags. For example, fmm_3d, fmm_asr and fmm_blochmodes belong to the class of fmm tags. The tag fmm_blochmodes is assigned to references that deal with Bloch modes using the Fourier Modal Method (fmm). Whenever fmm_blochmodes is assigned to a reference, I also assign the tags fmm and blochmodes to it. On the other hand, a reference might deal with Bloch modes, but not the Fourier Modal Method, and for this reference I only assign the blochmodes tag. This way I can filter all my references on Bloch modes using the blochmodes tag.

The above outlines one convention for assigning tags to references. You might feel like doing something different, but make some conventions and stick to them. As I will get back to in the sections on JabRef and PocketBib, this convention is convenient so also think of the software and apps you plan to use before setting up your own tagging system.

Index Everything You Read

A big project, like a Ph.D., might consist of several subprojects that each comes with references, notes and documents. It’s important to organize and index all of this, in particular articles you read in part or fully, so that you easily find them later.

Not long ago I found a couple of articles that I had read as background for one of my own subprojects, but that hadn’t been included in my literature database. I don’t need the material right now, but since I spent time finding and reading it they have now been included in the database – with some suitable tags from my list above.

JabRef: References and Tags on the Computer

As mentioned previously, I maintain my literature in a BibTeX database in JabRef. In JabRef, I have added a custom field “tags” to all entries (details on setting this up here), and using the tagging system discussed in the previous sections I add tags in this field for most of the entries in the database.

After tagging all the references, I have set up Dynamic Groups based on the “tags” field; this creates one group for each tag that contains references with the specified tag. Whenever a tag is removed from an existing reference, it is removed from the corresponding tag group, and when I add tags to new references in the database these automatically appear in the tag groups. In JabRef, these dynamic groups appear to the left, as seen in the picture below, and provide quick access to references in any tag group.

It is also possible to filter the references based on several tags simultaneously. Suppose, for example, I want to see all references with tags fmm and phc; this is done by searching for “tags=fmm and tags=phc” in JabRef (without the quotes). If I want to see all references related to photonic crystals (tag phc), I could search for “tags=phc or tags=phc_active or tags=phc_blochmodes or tags=phc_experimental or tags=phc_qds or tags=phc_singlephotons or tags=phc_slowlight”. But due to my convention of tagging all entries in the phc subclass with phc, I can also simply search for “tags=phc”. This makes my tagging convention convenient for filtering references based on both the individual subclass tags, like phc_active or phc_blochmodes, and based on the subclass title, phc.

PocketBib: References and Tags on the iPad

The literature database is stored in my Dropbox, and using PocketBib (also discussed previously) I can access it on my iPad. PocketBib is able to read the tags as well as the dynamic groups; these appear to the left in the app as shown in the picture below.

It is possible to edit entries in PocketBib, and documents can be opened directly in the app, if the associated PDFs in Dropbox are linked to the entry in the database (details on doing this here). Therefore, I recommend the app to anyone looking for a solid BibTeX manager for the iPad to be able to work on the go.

Simplenote: Tags for my Work Log

As discussed in detail earlier, I write a work log almost every day to document my work and to reflect on problems, ideas and progress in my work. I write these notes and logs in Simplenote since it provides a minimalistic and simple interface that allows me to focus on the writing.

One thing making Simplenote easy to use is its lack of folders to organize the notes; these all occur in one long list that may be sorted in various ways. Its efficient search function makes it quick to find notes containing specific words, and the lack of folders therefore in general isn’t a problem. However, sometimes I wish to see all the notes and work logs for a specific part of my work, and this is where the Simplenote tags come in handy; it is possible to only show notes with a specified tag, and this therefore makes for a lightweight folder system in Simplenote.

Like in my literature database, I had initially added tags as I went along, and this didn’t provide a very efficient way of filtering my notes. I therefore went over all my tags, removed a lot of them and sticked to the tags shown in the picture below. These might very well change in the course of time, but they at least provide some relevant context in my notes. For example, all notes relating to quasi-normal modes, which is something I’m working on right now, can be filtered using the quasinormalmode tag.

I also keep a lot of notes in Evernote, but I haven’t implemented a tagging system in there just yet; its folders and stacks in the first place provide a structured way of organizing the notes.