I stumbled across your blog as I have been researching how to best simplify and future proof my workflow: leading to text files, NVAlt and Simplenote. One issue with these solutions is the flat folder structure, tags get you so far but you are still presented with a long list of files. So your and AmberV’s excellent posts made a lot of sense as a means of providing additional, searchable, robust, easy categorisation.

So I gave a lot of thought to how best to adopt this system for myself. I first mapped all the types of docs I have and the groupings that made most sense to me. After many hours of thinking, rereading your post multiple times to clarify the theory, moving things around etc, I ended up with effectively the same six tags as you. It feels like this was my own doing rather than being influenced by you, but either way it does perhaps mean that this system can be more readily transferable to different people or situations than folder structures.

Next was the question of sub-tags. I came at this in a different way. My approach rests on two principles: ease of remembering, and allowing automatic cross comparisons between wholly different documents. The answer, for me, has been to have only a single sub-tag (either 1, 2, 3, or 4) and for the meanings of these sub-tags to be identical across all tags. This obviously removes some information that your system allows, but for me the benefits will hopefully outweigh the cons.

Getting into the detail, I was struck by AmberV’s use of internal and external for several of her top tags and in your system 1′s and 3′s often denoting similar meaning across tags (so I do not profess to be wholly original). I mapped out what I thought could be the most useful binary distinctions between my documents and came up with: internal/by me vs external/by others; abstract/reflexive vs practical; informal/personal vs formal/official; draft vs completed. The first two of these resonated the most as both readily applying to just about everything but also allowing useful distinctions across all tags. Two sets of binary tags leads to four combinations in total, so to reduce typing one number would denote a combination of both variables.

So my system uses the following: 1 = by me and reflexive (e.g. diary entries, charting my weight, wedding speeches, fiction writing, meeting notes, letters to friends) 2 = by me and practical (e.g. tasks lists and project work, made-up recipes, meeting actions, job applications, letters to the bank) 3 = by others and reflexive (e.g. friend’s fiction, health reports from the doctor (as these just record my health rather than require action – potentially financial statements could fit here too), articles about the creative process or say meditation, letters from friends) 4 = by others and practical (e.g. recipes from the web, letters requiring action, training material)

I realise there is a blurred line between these sub-tags, as there is with any system, but it makes sense to me. My file name structure is very similar to yours, though of course a bit more detail may be required in the prose name. I have also added at the end of the name the option for additional metadata (e.g. which chef wrote the recipe). I previously have used [] to denote this sort of information but I realise that ( ) makes more sense so that spotlight searches on “[…” will only bring up tags.

I am only just starting to use this system and have yet to proliferate across all files (where my first step is to get consistent dating in the file names) so I have still to see the benefits of this system, meaning it may well change.

Finally, again, thanks for the inspiring and insightful analysis and practical help, and I hope this comment is of some interest to you or other readers.

Good questions. I’m writing a reply as a post that should be up shortly.

I’ve been trying to get my head round your post and AmberV’s comments, plus the forum posts. I understand the filename system. I have a few questions:

1. Do you apply your system to all files you create on your computer including proprietary or only text files?

2. The folders are causing me an issue. Since the filename pretty much covers the file, there would seem no need for folders, except that you would end up with a single folder with a massive amount of data. Would it be best to use a few folders that cover broad areas such as ‘work’, ‘family’, etc.. This one really baffles me.

many thanks

http://macpowerusers.com/2010/03/mpu-023-workflows-with-merlin-mann/

For instance, in my EagleFiler library, I store fonts, some dmgs of applications I keep when I install a new machine, and other binary data. As I’m slowly migrating data to the system, I’m wondering if it should all go in.

Thanks,

Alan

I’ll keep playing with this system for a couple weeks, and see how well it goes. I’ll probably come back here for additional questions ;-)

Thanks again,

Alan

In my documents folder I have what I think of as my root folders…

• Admin (Bills, finance, taxes and such)

• Commerce (Ephemeral project files)

• Gaming (yep. I’m one of those)

• Organizer (Of less and less use over time)

• Reference (Static items like System notes, shortcuts, graphics files)

• Topics (The biggie, home to 100+ topic sub-folders)

• Writing (The other biggie, where all my work resides)

Why? I like having a specific environment for things like my Writing work. Some day I’ll post a screen shot, but I’ve added icon graphics and background pictures to all my root folders reflective of their individual purpose. My Writing folder has a zen like sprig of wheat on rice paper and its icon is a star burst (kind of).

I’m trying to implement something similar to what you did, inside EagleFiler (with the insurance that if I stop using it, the file naming convention will preserve all the metadata). One thing I did not find clear from your description is where you store your files: you mention aliases very often, so do you keep everything in one folder then do aliases from there?

I’m trying to implement a system similar to this one, and I was wondering if you also integrated administrative stuff filed away (such as bills or bank statements), and if so in which category. (I’m tempted to create a new one for them, but I realize that the strength of the system is in keeping down the number of categories.)

Thanks,

Alan

I avoid them for the sake of sanity. If the system expands past the point of easy recall I will stop using it. The breakthrough for me was the flight to abstraction the system provides. Since info gathering is by design a diverse and hard to predict activity, the code system has to work at a level of abstraction applicable as far as I can foresee. More and more I’m using in-text multimarkdown tags to augment foldering for categories. But by far the best finding tool I have is a well crafted full text search using spotlight.

>>How do you manage bibliographic references?

Zotero in Firefox

>>Have you tried “hiperlinking” documents in this file systems? (while making this system platform independent and application independent)

One of the features I mis in “everything bucket” applications is wikilinks, the easy link between documents. Scrivener has an internal wikilink, Journler did as well. DEVONThink has a very good one. TinderBox, of course is based on it.

Right now I add the names of files in text to a document, if i feel there is value in a link – someone recommended adding index cards of links as separate file – then using spotlight against the file name. Sounds complicated. It is very easy. I have a rough draft of a post about this process because it is not yet a solved issue. Stay tuned.

After beign a self-nonpaid-software tester for the sake of finding the ideal application to read, record, anotate, and retrieve any information, I’ve ended almost without knowing relying on my learning application (supermemo). The reasons did not seem obvious at first. However, more and more I kept relying in it,because of the individuality of each file to keep articles, the possibility of editing on place and the fact that even if this program did not exist in the future I wouldn’t loose any file. It can manage images to, so it seem an ugly (very bad UI) long term soulution. Then I realized that because of troubles with backup it wouldn’t be so. But my file sistem did have a consistent backp procedure. That was the Aja! moment.

I started relying on files, as well almost a year ago, but I’m not on MAC, which by does not leave me a decent option to “Scrivener” or “DEVONthink” and this fact makes me look with comptempt my OS. From your experiences I can tell, I would have endned at the same end of the road – “File System Commonplace”. I still envy Scrivener though.

My format follow similar principles, but it uses keywords also. By using Desktop applications I’m finding less need to include the author and I’m thinking on changing to categories insted of keywords, but I’m not certain if it is a long term solution.

File System Commplace format: 2010.02.25-Title_Autor-Key1.Key2.Key3.ext

Have you included keywords on your naming convention before? If so, why did you stoped doing it? Is it that you preffer two word code, primarily, for the sake of typing speed?

How do you manage bibliographic references?

Have you tried “hiperlinking” documents in this file systems? (while making this system platform independent and application independent)

Good luck!

PS: I probably sent a previous unfinished comment, please disregard and deleted it.

DDB: Thanks Andreas Schmidt. Excellent article. (I’m always humbled to be mentioned in the same paragraph as Alex Payne and, the unnamed but linked to, AmberV)

Domo arigato!!!!