Is Our Current File & Folder System Flawed?

Since personal computers have first appeared on the market, their file systems have been organized in a very similar way to the method we use for organizing things in the physical world. Libraries, historical archives, hospital records, and many other things are organized in a hierarchical manner to make it easier to conjure up any information we would need at any point in time. Similarly, computers work with one root directory and a hierarchy of subfolders that let us easily browse to any file we wish. We’ve been using this system for the better part of all recorded history, which begs the question: Can we do better?

To answer the question we asked earlier, we need to think about whether the file and folder system has flaws at all. It turns out that there is actually one enormous issue with the way we organize our data: The typical file on a computer today is more than a simple document; it is multimedia that can be organized in several ways, with many attributes attached to each piece of data. Films – to point out one example – have a genre, lead actors, a soundtrack, a director, and many other things that can make it difficult to classify them into a single-attribute data structure.

What criterion do you choose to ultimately organize your movie collection with? If you organize it by genre (Action, Comedy, Thriller, etc.), you can only find a film easily by using that particular aspect when browsing through your collection. In other words, it would take you quite a while to find a film directed by someone who likes to venture into various genres.

filesfolders-solu

In October 16, 2015, a journalist from the BBC went to see a couple of Finnish developers making a computer system that is meant to revolutionize the way we organize our files. Named Solu, this small touchscreen-based device can connect to your monitor and features a map-style file system. There are no folders, and there are no files. All you see are icons showing projects and how they are divided. There aren’t many details on this system, and it’s unlikely that Solu will reach mainstream use. All we know is that the developers told the journalist that the file system promotes productivity. This isn’t extremely helpful, but I suspect that it is just a project map leading to a number of apps grouped by function and importance. It runs, after all, on a very heavily-modified version of Android.

filesfolders-winfs

Since 2003, Microsoft has been developing a new file system that would replace the file-folder hierarchy we all know and love. WinFS was supposed to be released with Windows Vista in 2006, but development was significantly slower than expected. It promised a relational database hierarchy, which would organize files based on various relationships that they possess to one another. One of the drawbacks of this system was the fact that it presented several complications to programmers who want to make full use of it in their new releases. Yes, many other file systems exist that try to replicate this, but they exist in a vacuum somewhere on the internet for a reason.

While relational databases might make it easier for people to browse in their computers, they present an incredible challenge to programmers who would much rather use the file-folder hierarchy. The truth is that using relational databases complicates things excessively.

Many attempts have been made in the past to replace the hierarchy we currently use to browse our files. Most of them have seen limited amounts of success in some niche communities, but the tried-and-true method of using files, folders, and subfolders overcomes its drawbacks with its simplicity for both the user and the developer. Although it is indeed difficult to categorize all types of data, it is easy to learn to use and it doesn’t require a lot of convoluted code.

Do you know of any alternatives that could potentially beat the traditional file system? Tell us about it in a comment!

9 comments

  1. Any system needs to preserve the means to “archive” information in a future-readable way and not lock data into a proprietary format that may die away in a few years. While hierarchical storage is just fine, what we need are good approaches to indexing, retrieving, and displaying that information in an intuitive and easy manner. There are many approaches already existing out there – but none have taken off on a wide scale since they all seem to have their limitations.

  2. Yes, our file system is flawed–it is too simplistic. I’ve often wished I could have subfile categories that are color-coded or distinct in some way–like file names in bold, or using different icons. Each user could control the naming of these subfiles to meet his/her own needs. Looking at my Docs Lib, I wish I could toggle on/off all the file folders listed so that I could see a listing of their subfiles–and compare these subfiles at a glance, instead of having to look through each folder, one at a time.
    The problem is probably my own fault, as I have tons of files on the same general topics, but was too busy as I went along to combine them into folders. Or to combine them into meaningfully labelled folders within folders. But I do so much research and writing on different computers and tablets, ipads, etc., and have duplicates and numerous revised docs (listed by date) that when I download them into my “work” computer, the whole Docs Lib is a tangled mess. Maybe I just need some instruction on how to create a better filing system. Maybe I need more patience and time(!) Anyone have a good system or manual to suggest?

    • Shelley, I very much agree with you that having the ability to use colors and graphics instead of the current folder icon would be GREAT! As I read your post, I thought about how I have arranged my own files. As a teacher, I need folders by subject, by section within subject, and a central way to track all of the files and projects my students send to me. Add to this the fact that I taught at more than one college, and the arrangement gets REALLY interesting. I have found that it is easiest way to organize my stuff (folders, files, etc.), is to plan the arrangement first and then to follow the plan religiously. Trying to decide where a file SHOULD have gone after the fact can be daunting! The other option is to setup your arrangement after the fact and spend a LONG weekend moving stuff around. To have a color option would have saved me many hours of re-arranging files that I misplaced in the wrong folder.

      • @Shelley and Frank T:
        Attila Orosz wrote an article for MTE about Tagspaces, a personal file organizer. Maybe Tagspaces can provide the filing structure you are looking for?

        The other solution would be to create a relational database for your records.

  3. I know that this is not much, but graph structures can be mimicked the way C# or Java does it–by using interfaces. In a directory system, that would require using links. They appear with arrows superimposed on their icons. These could be stored in directories that represent the other classifications the file falls under.

    I know that this isn’t perfect, but it works.

  4. I use Q-dir。It can split into 2 windows or 4 windows, etc. You can also config the color. give it a try

  5. The file/folder system is not obsolete. One thing it does is to ensure that there is a path to the object.

    However, consider a set of pictures. The first, most obvious way to store them is by date. This creates a place for every image.

    However, this makes it hard to find pictures of Sarah. So we keyword images so we know which images have pictures of Sarah.

    In some photo programs we can create smart albums. We can define a smart album “Has keyword ‘sarah'” Then whenever we tag a new image with Sarah, it automatically is viewable in the smart album.

    In this way we can tag images. But we have to take the time to do the tagging.

    In a broader view, we have projects. Probably starts life as a folder and grows. If we are doing lots of similar projects we will hash out a structure: header files, libraries, documentation, source code, binaries, resources. If we are *really* clever we implement a tag system that is not only hierarchical for the tags, but that objects inherit tags from enclosing folders. E.g. All the documentation is not only tagged documentation, but also tagged Project Foo.

    Now I can define a view: Show me all projects documentation that was managed by the same guy who managed the documentation for Project Foo.

  6. “Is Our Current File & Folder System Flawed?”
    No. It follows the KISS principle or Occam’s Razor, if you prefer a more sophisticated name.

    It may be too simplistic for keeping track of many of today’s files but it is not flawed. It is very convenient to have have our files cross-referenced under multiple subject. However, now you’re talking a relational database. How many computer users are willing and/or capable of designing, programming and maintaining even the least complicated RDBMS? The problem with a relational database file system is that there would be as many versions of the system as there are users because each user has his or her own requirements, and/or ideas, on how to organize their data. Just look at the current flat file system. Each one of us has different folders and sub-folders.

    “We’ve been using this system for the better part of all recorded history”
    We’ve been using this system because physical objects can be stored/filed in only one place at a time. A library cannot shelve the same physical book by author, by title and by subject.

Comments are closed.

Sponsored Stories