Introducing file systems to students who really don't understand

Question

In this question, Michael0x2a points out something that I have observed, but never formally considered (emphasis my own):

4. The IDE often introduces several new "metaphors" that need to be explained. (For example, Eclipse has this notion of a "workspace" -- it's not immediately obvious this is actually a folder on your filesystem). This problem is further exacerbated by the fact that a small minority of students don't seem to have a firm grasp of what a filesystem is (which is disconcerting).

I have had, on occasion, to step all the way back to the idea of folders, directories, and directory trees for confused students, but these conversations always get muddled. In spite of years of teaching this material, I am always surprised when a student cannot tell me what directory he or she is in, nor link the idea of a directory to some sort of metaphorical location.

Our so-called techno-natives have real difficulties with this, since phones and taskbars hide directory structures almost entirely. This problem is especially easy to spot the first time a student is introduced to a command line, when I may have a conversation like this:

You are currently in your home directory. Type cd somedir and press enter. Great! We have changed our directory to somedir. Now type cd somesubdir... where are we now?"'

Some students know, but some are genuinely lost. The idea of going back up a directory cd .., or that ls lists the items at your current location, is so opaque that the question itself doesn't even seem to make sense.

Many beginning students seem to already come in with an intuitive understanding of folders/directories, but some do not. Are there useful analogies or metaphors that you use to teach file systems to your students? Do you address this problem in your classroom, and, if so, how?

Answer 1

The problems

Hierarchies are not obvious

First let me point out that a hierarchy is not the most obvious or best structure for storing files. It is still based on library categorization systems, where a book can be in only one place. I file could be indexed in more than one place (and sometimes is).

A tag system as used on this-site, Gmail and elsewhere, that allows you to give multiple tags to a file, is better. E.g Tag something as education, computers, filesystems, cats, boxes.

Tags should be arranged hierarchically, not files. However as educators there is little we can do about this.

Microsoft Windows, Mac OS X, Android, and iOS hide things

Of the big four major proprietary Operating Systems (Windows, OS X, Android, iOS), Microsoft Windows hides the directory structure the least. Apple is probably the biggest villain here. They made an active decision starting with the iPod and iTunes that directories needed to be hidden, and the world has been moving that way ever since.

What I say in this section applies generally to all 4, but I have given specific example for Microsoft Windows.

MS-Windows has some strange non-folder folders (Desktop, Documents, Videos, Music), sort of like symbolic links but not. This causes problems of understanding. For example, I used a tool and known that it has saved to Images, but took over 10 minutes to retrieve the file, as I did not know where on the filesystem it was.

Recent versions of MS-Windows have dropped the ⬆ up arrow (they then brought it back again), in file explorer. It only has ⬅ ➡ left and right arrows. This makes it difficult to navigate and experience the hierarchy.

Recent versions of MS-Windows, by default hide file extensions. This may not affect understanding of hierarchy but does affect understanding of file names. (While I agree that using file extensions to encode file type is a bad idea, hiding them under the carpet is just making it worse.)

Microsoft wants to store pictures with pictures, music with music, spreadsheets with spreadsheets. I want to store teaching stuff with teaching stuff: A class may have a presentation, a work sheet, some images, some source code, etc. it all needs to be kept together.

Some solutions

A teaching metaphor

I have taught some year 7 pupils (age 11 years old), by showing them a set of slide and telling them a story. The story was about how you can but a box in a box, and a box in a box in a box, and several boxes in a box. Boxes have names/labels, cats have names/collars. You can also put a cat in a box, but you must never try to put a box in a cat. Boxes are like folders/directories, cats are like files.

You can ask students: “Can a box be in more that one box?”, “Can a cat be, fully, in more than one box?”, “Can a box and a cat be in the same box?” …

If using Microsoft Windows

• Disable hide file extensions
• Do something about these Libraries (non-folder folders): In my virtual machine with Windows 7, I have disabled all of the libraries.

Learn the filesystem and the operating system

I would teach the Unix command line (UNIX, GNU/Linux, BSD). It does makes it easier to see how things work. Nothing is hidden, but things are often nicely abstracted. “It is so simple that, if you come up with a metaphor to explain how it works, it probably is how it works.” — miss-quoted from one of the founders of Unix (Thomas, Ritche or other).

I would also recommend using the dolphin file browser. Dolphin has an integrated command line, where directory changes are synchronized in both directions. This makes it easier to switch back and forth between text (command line) and graphical view and/or controller.

You are currently in your home directory. Type cd «somedir» and press Enter.

See the change in the graphical part of the display.

Click on a directory

See the change, in both GUI and command line

---

GNU/Linux with KDE, has a great GUI, and command line. Intuitive, simple, and follows the principle of least surprise.

GNU/Linux is more available (When compared to some other OSs e.g. macOS, Microsoft Windows): You have the freedom to download it. Run it. Distribute it. Study it. Modify it.

Answer 2

Its actually harder than just finding a modern user-interface which exposes a directory structure. The whole concept of a directory structure is just a convenience - there is nothing 'physical' which it relates to.

Most of us started our interactions with computers at a point where directories were the first point of interaction, so it is hard to relate to the experience gained from modern interfaces where the 'workings' are abstracted and files are located by searching rather than through an index.

I think it's now become important to teach storage structures as soon as you have a scenario which permits it. I think that a library is a good analogy - people are familiar with a large amount of structured content, and books containing different types of content.

To discuss where a document (or book) might be stored, you have potentially quite a lot of layers:

• Country
• City
• Institution (university, council library)
• Department
• Building
• Floor
• Shelf
• Shelf position
• Book title

You can also have copies of the same book - with different notes in the margins, or different pages missing. Today, files are maybe in an anonymous data centre, or on a device, or on removable storage. I think it's useful to go to this level first before introducing the idea that we have a virtual structure within any one drive, because this helps to explain why it starts to be useful to have a structure. When you only own 5 or 10 books, you can easily find them all. Once it's all your photos, and all the music you ever streamed - then it's not so easy.

At this point, you can introduce the idea of arbitrary named containers (which are nested). Explain what they're for and how they are important, then show the real-life example on the actual hardware you're using (and explain why this is abstracted from normal non-expert users).

Note: It also occurs to me that spending too much time re-enforcing legacy storage concepts is not helpful, so resorting to the command line as a primary teaching tool for this might not be good. Teach enough that if they are in an environment where the command line is important, the concepts come naturally, but don't pass on your own legacy baggage.

Answer 3

Make sure you are having your students "spelunk" in the file system and have them draw some of the tree structure. Often, this visualization helps them see the way the FS works, more clearly.

We spend a couple of days gaining an understanding of absolute and relative path. I teach this course using a Gnu/Linux server.