Folder Im Folder

Thoughts Design Technology

Do we still need folder hierarchies? Apple, and Oliver Reichenstein don’t seem to think so. I discuss.

Oliver Reichenstein has an interesting post over at Information Architects, talking about the way that Apple is approaching folder hierarchies in its new Mountain Lion operating system. If you haven’t read it already, I suggest you do. It’s well-written and presents some interesting arguments which will challenge some of your assumptions about how we organise our digital lives. Despite that, however, I can’t help coming away from it feeling slightly concerned. Oliver has now updated his post to clarify that the restriction on levels of folders is not mandatory, but still seems to think that this is “the future.” I’m worried by this, and I thought I’d spend some time in this post talking about why.

A short disclaimer

Before I begin, I want to make it clear that I’m not really interested in discussing the nitty-gritty details of what Apple is or isn’t doing in this, or any other, version of OSX or iOS. What I’m going to talk about here is the ideas and assumptions involved in the thought process that ends with restricting users to only having one folder level (I know it’s voluntary — we’ll get to that in a minute).

The problem of visualisation

In his post, Oliver says:

Folders tend to grow deeper and deeper. As soon as we have more than a handful of notions, or (beware!) more than one hierarchical level of notions, it gets hard for most brains to build a mental model of that information architecture. While it is common to have several hierarchy levels in applications and file systems, they actually don’t work very well. We are just not smart enough to deal with notional pyramids.

This is, essentially, the core of his argument: that because we can’t visualise or retain some kind of mental model of a complex hierarchical system, that system is not a good way to work. The problem with this, I think, is that it over-simplifies the way that folder hierarchies work for most us.

Traditionally, computer filesystems (as they appear to the user) mirror the kind of information archicture found in a filing cabinet. Files exist within folders which, themselves, exist within drawers and boxes. The only difference being that, on a computer, we can have infinitely recursive filing cabinets. Oliver is quite correct in pointing out that this can introduce some problems. We can often fall into the trap of thinking that more folders equates to more organisation. This can have some quite bizarre consequences. I can remember more than one client I’ve worked with who put just about every file into its own folder so that the location and contents of everything was represented in the file system. In such a world, the file itself almost becomes invisible, with the folder becoming the unit of information. Such users visualise the folder almost as the title of a document, with the file its contents. In an example like this, it is easy to see Oliver’s point — infinite depth is not found in the real world, which ultimately means that we often don’t deal with it very well on a computer.

The problem with that argument, however, is that (disastrous organisation schemes as above notwithstanding) it’s not a true picture of how we interact with deep hierarchies. It is true that we mostly can’t visualise the entire hierarchy but that misses the point. We don’t need to visualise the entire hierarchy at once in order to be able to use it. The way that file managers work allows us to drill down into a hierarchy and see only those parts of it that are relevant to our current task or train of thought. I’ll admit this may not be the most efficient way of dealing with files for every task, but it is useful from time to time, particularly when you don’t know exactly what you’re looking for.

To use a different metaphor, imagine how hard it would be to find a book in the library without the Dewey system. The Dewey Decmial System is, in essence, a hierarchy. It allows you to start with a topic, and drill down into sub sections, authors and eventually, to browse a small selection of books. If you’re not sure exactly what you’re looking for, this kind of process is essential (as is the library search database — they work together). Hierarchies allow us, particularly when working in teams, to share not only files, but the thought processes and mental models that go with them. They are not, perhaps the best way of doing it all the time, but they work really, really well for some things.

Trying to picture notional systems with several levels is like thinking three moves ahead in chess. Everybody believes that they can, but only a few skilled people really can do it. If you doubt this, prove me wrong by telling me what is in each file menu in your browser.

I’m going to leave that file menu question alone for now because it doesn’t really deserve a response. Suffice to say that if menus required us to be able to remember every item, we wouldn’t be using them. Let’s look at the chess metaphor instead.

Being able to think three moves ahead might make you a better chess player, but not being able to doesn’t stop you playing chess, you simply deal with one move at a time. Claiming we should remove the folder hierarchy because most users can’t visualise most of it, is like claiming that because most people can’t think more than two moves ahead, every chess game should consist of only two moves.

The problem with geeks

Oliver claims that folder hierarchies are a hangover from an earlier stage of computing designed around a different set of users.

The folder system paradigm is a geeky concept. Geeks built it because geeks need it. Geeks organize files all day long. Geeks don’t know and don’t really care how much their systems suck for other people. Geeks do not realize that for most people organizing documents within an operating system next to System files and applications feels like a complicated and maybe even dangerous business. Remember that autoexec.bat file?

Geeks think in our own little boxes. If one comes up with a new pattern for a common interaction, others will naturally be generally skeptical. That’s usually a healthy attitude for usability issues. But that’s not the real reason why geeks are skeptical. It’s because we are smart asses. We are the people that put salt and pepper on the pizza before trying it, because we just know best.

While I agree that the history of personal computing is littered with examples of things which have been designed in a way that will make users feel like they are about to break things, I must take issue here. Not everything designed by geeks is easy to use (regular expressions anyone?) but that doesn’t mean that we should go to the other extreme and make life difficult for the technical users (I know, it’s voluntary — just hold on).

Information Architecture

One of the key ideas in Oliver’s post is that each app has its own paradigm and its own way of interacting with files. This is certainly true. I know that my own filesystem has a variety of different organisational schemes depending on the nature of the files in any given folder. This is a good thing. An even better thing is when each app is able to represent the files under its control in a way that makes sense for its particular workflow. A music player such as iTunes is a good example of this, and Oliver talks about that in his post. It makes sense, within a music player to think in terms of genres, artists, albums and dates rather than in terms of folders. That’s fine.

The problem arises when we have to interact with the same files in multiple apps, which may or may not share the same paradigm. If all my music files are in my top level ‘Music’ folder, that may work well for listening to my music, but what happens when I want to use some of that music in a video editing app which doesn’t share my music app’s paradigm, or doesn’t allow me to drag and drop from one app to another?

Organising my music library into rudimentary artis/album folders doesn’t give me all the flexibility that I have when I use my music player app, but it does allow me to find things when I need to, or when I need to use my files in an app that doesn’t exist as part of the happy Apple ecosystem (leaving that whole question very determinedly out of this post). More importantly, it doesn’t actually stop my music player from giving me all the other information as well, because my music player doesn’t care about my file structure; it just looks for music files.

Now, you could argue that actually, I don’t need two levels of folders there. I could simply organise my music into Music/Artist or Music/Album folders and then use filename prefixes to do the rest. That’s fine, except that filename prefixes are a type of hierarchy, and if you’re determined not to have more than one level, you shouldn’t use them. If you’re going to use them, why not just reintroduce another folder level and be done with it?

The file manager is an app

The other issue here is that, while it’s fine to have each app interacting with files in its own way, it’s easy to forget that this should also apply to the file manager. The file manager is an app, and it has its own paradigm: folders and files. This shouldn’t just apply to fully fledged file manager software either. When you open up the file picker (or the Document Library) you are entering a particular paradigm. I argue that the traditional file management paradigm should be available to you (although not exclusively) even within a different app. A file picker isn’t always the best way to find files, but it should be there.

How do we solve this?

Let me be clear: I think different file management paradigms for different apps is a good thing. I agree with Oliver that there are some tasks for which the infinitely deep folder paradigm just doesn’t work. I also think, however, that choice is a good thing, and that assuming that all your users are going to share your particular paradigm for how to work with files is a very bad thing. That’s actually what Oliver was talking about when he complained about geeks and their legacy, only turned on its head. If geeks gave us the autoexec.bat syndrome, lazy or patronising app developers will give us the “How do I open this file in another app?” syndrome which will lead to convolutions like excessive use of filename prefixes, or duplicate files and folders.

If there’s one thing we should never underestimate here, it’s the lengths that even unskilled users will go to in order to interact with things in a way which makes sense to them. We can have some control over it by designing well thought-out interfaces, but we can never imagine and deal with all the possible use cases.

The argument that we need to simplify the underlying file system in order to present a more usable interface to users is simply fallacious. We don’t need to do that. We can use metadata (and in fact we already do) to present different paradigms in different places without changing anything. Folders themselves are, in fact, metadata already, so why get rid of them if they are useful in some paradigms or for some users?

Removing folder hierarchies without providing a good alternative, that works without app developers having to build it into their apps, makes a system not simple, but simplistic.

But it’s voluntary!

Let’s return to Apple for a minute. Assuming that you agree with Oliver when he claims that multiple levels of folder don’t work, there remains the question of what to do about it. Apple has a long history of improving interfaces because it thinks they’re broken. This is a good thing — more power to them. What’s not so good is that they often do this by breaking other things.

You can argue that it’s voluntary, and Oliver does, but we’re still expected to believe that this is the future of filesystem management. The reality is that if you want to use Apple products, it’s a good idea to use iCloud. It makes sense and, increasingly, as OSX and iOS slowly merge, it’s going to be essential. That doesn’t sound very voluntary to me.

The risk is that we end up with two filesystems; one for your personal files which you store in iCloud, and another for your work projects and more technical stuff. This assumes that my personal files aren’t as complex as my other stuff. Is this true? Is my personal life and its digital manifestation like that? I’m not sure. What I am sure about is that I would like to be able to organise my stuff in whatever way I want, even if that is inefficient in other people’s eyes.

Not everyone needs folder hierarchies, and many people use them badly. Gentle nudges in the form of default folders and better metadata can help enormously. Limiting everyone to just one level of folders won’t hurt those, more malleable users, but it will really hurt those who really, truly need a hierarchy. This seems a shame, as it’s not needed in order to provide any new features.

Regardless whether you’re talking about Apple products, or not, the idea that we can and should remove folder hierarchies because it’s complicated is concerning. Oliver’s right that the information architecture you have to do to deal with folders is hard. It’s nothing, however, in comparison to the information architecture you have to do to work with a flat-file system when you’re metadata won’t do the job for you.

Let’s make metadata better; not filesystems worse.