I do this rigorously with tremendous success. I can find anything, instantly. Yes, it takes a little work up front. For me, the payback is obvious and extreme.
I have to get round to finishing my site so please excuse its current incompletion, but hey, if you're interested: http://johnnydecimal.com
I use this system at work and I have otherwise very competent Project Managers stare at me like I'm some sort of wizard when I navigate to the right folder, the first time, every time, and find exactly what I'm looking for. Email, our shared drive, SharePoint - I use it everywhere. I would be a mess without it.
I just read the stuff on johhnydecimal.com and am near-instantly hooked! I've signed up for your mailing list, and am really pumped up to see when this is finished, and when more stuff gets added...and I appreciate that i can start implementing your recommendations immediately. Well done and kudos, I must say!
If you think about it, folders are really like limited tags where the only relationship you are allowed is "and then".
It would be nice to see a system that would present a tag set in the appearance of folders for all "and then" relationships but then let you mix in things like "and also", "or", and "but not".
This is the nodal graph structure used in ibGib. Each datum is an immutable snapshot with internal an data map and named rel8ns to other ibGib. So each ibGib is a folder and a file. It is ends up technically being a merkle tree graph, because the `gib` is a sha 256 hash of the contents. "Tags" per se are only another named relation that ive just implemented in my feature branch yesterday: https://github.com/ibgib/ibgib/commit/776d053005bd5269892c57...
Interesting. I'm working on an image tagging system that basically does the same. It grew out of the fact that I'm the family historian and I have 40k family pictures spanning 8 households that go back to 1840. Gets hard to find that one picture you vaguely remember seeing once.
+1 for hierarchical tags. I've been organizing all my bookmarks (about 12K going back more than a decade) with Usenet-style tags like comp.lang.python and sci.bio.paleontology for several years now. It works great, especially since Pinboard lets me autocomplete by tag prefix.
One of the things that I think most services like Gmail missed when they tried to replace folders with tags is that although the same item can belong to multiple categories, the categories themselves usually don't have a completely flat relationship to one another. An article can relate to both Python and paleontology, but that doesn't change the fact that Python is a programming language and paleontology is a subfield of the biological sciences. This will be very useful later when I'm trying to remember something that uses some programming language to analyze data in some subfield of the biological sciences, for example.
Tags look cute on a screenshot when you have a few hundred items. But when you have tens of thousands of items, you need a bit of hierarchy to protect your sanity.
Nested labels were only introduced in 2010 or 2011. Can't remember exactly when, but they certainly weren't available when I first started and then stopped using Gmail. Something as basic as how to organize your email should not have been an afterthought.
It was probably a concession, not an afterthought. Google's whole thing with Gmail was "keep all your email, don't bother organizing, search will solve all."
tmsu (https://github.com/oniony/TMSU/) is working on that. It's a project to give tags on all the files you care about and then lets you query them either manually on the command line, or through a virtual file system where each "subdirectory" is actually a "and this tag as well". For more complicated queries such as the one you want, it is possible to directly cd into a complicated query; see https://github.com/oniony/TMSU/wiki/Virtual-Filesystem#queri... for instance.
Looks like this is a 2005 study from William Jones, et al.
If folks are interested in this line of research, Jones has a 2007 book called Keeping Found Things Found that takes a more comprehensive look at personal information management.
The ability to organize email into folders is one of the main reasons I prefer Outlook over GMail. But as the paper mentions, "Folders can obscure as well as organize", which is where tags come in handy. What I would really like - for both email and files - is the ability to create both folders and tags quickly.
Gmail has folders; I have the same ones I used to have in Outlook. They're called labels, and the only practical difference I've noticed is that you can have more than one label on a message, unlike folders where a message can only be in one. You can nest labels one inside another and get a nice hierarchical list of them on the left just like Outlook.
Yes, it does now, and I can't for the life of me remember when that happened. But this was a capitulation after Google long insisted that tags were totally superior.
Notice there now are options like "Move to", which basically removes the tag for the current folder - I mean tag - and applied a new tag to let people treat Gmail mostly as if it is folder based, because as it turns out a lot of us insists on folders. But that functionality didn't use to be there. Both the nesting and "acts almost like folders" actions were added at some point long after Gmail was launched.
I agree with you on documents, which tend to be small in number so you can index them manually using folders. But what about the case where you get hundreds of messages a day? If you get 300 messages a day you'll accumulate over 100K messages annually. Manual indexing seems like madness. I don't even try rule based tagging at that point.
In this latter case the gmail model is great. I leave everything in the inbox and depend on gmail (a) never to lose it and (b) to find it easily through search. In fact you can use gmail as a doc store--if there's something I really value I make it an email attachment and depend on gmail to dig it up if necessary.
I think of search as a default functionality (although it was actually pretty bad in older versions of Outlook). But from a usability standpoint, software shouldn't pigeonhole users into one method or another. Since everyone's work flow is different, it's all about providing users flexibility.
> But from a usability standpoint, software shouldn't pigeonhole users into one method or another. Since everyone's work flow is different, it's all about providing users flexibility.
I'm not sure if I agree. I think it might be better to software to excel in one specific workflow rather than attempt to cater everyone. Of course this requires the market to provide different options for different workflows.
I get hundreds of messages a day. I don't accumulate over 100k messages annually as very few of those hundreds of messages a day are necessary to retain. What isn't automatically filtered, I look at, and either delete, archive or categorise.
I don't leave everything in the inbox for the reason that I can't find what I want, in part because the search is awful. A large part of the problem is that unlike websites where pagerank and context provided by inbound links allow a lot of intelligence, with my e-mail it is often a struggle to find the right search terms, and so I want to add context in the forms of categorisation to the messages I actually have a reason to keep.
Gmails poor folder support kept me from using it for my work account for many years, until they "surrendered" from their original hardline tagging stance and added more support to allow you to treat tags like hierarchical folders, for this reason.
I can't imagine keeping an unstructured inbox like what you're describing - I wouldn't find a thing. I always despair when I have to resort to Gmails search as it almost invariably means I'm in for frustration.
A good folder hierarchy is key to my OCD sanity. I never truly trust search alone. Tagging is not reliable as it depends on your state of mind being 100% consistent every time you add a tag to a file. Reliably defining scope and organizing by category within scope is much more consistent. Nice Post!
If the goal is to organize things for easy future browsing, the folder approach works well. But, if the goal is to clearly see what the next actions are and make progress on them, it is not a good approach since the latest and important items are buried somewhere deep inside all the folders.
http://www.zim-wiki.org/ is a personal database I use to organise notes, projects and scripts. Simple cross-note linking and automatic TODO list generation are particularly useful features I use frequently. The 'database' is stored markdown-formatted files that can be easily version controlled and edited from the command line.
I've done some Zim evangelism on HN before ([1], [2]), so I'll merely link those here and mention that I'd love to answer any questions about how I use it.
I'm on a MacBook. Is there some type of virtual folder type that I can use that behaves like any other folder? Something where the files and folders within don't actually 'live' there.
I don't like tags because they're not tangible. If they could live somewhere on the desktop where I could click them, or be nestable, I would use them.
You can create a Smart Folder that is based on one or more Tags (or any other combination of metadata, but Tags are useful in that you can manipulate them directly).
I have to get round to finishing my site so please excuse its current incompletion, but hey, if you're interested: http://johnnydecimal.com
I use this system at work and I have otherwise very competent Project Managers stare at me like I'm some sort of wizard when I navigate to the right folder, the first time, every time, and find exactly what I'm looking for. Email, our shared drive, SharePoint - I use it everywhere. I would be a mess without it.