The ubiquitous digital file: a review of file management research

User experiences with tagging

Reportedly, users struggle with tags, and display a preference for folders in practice, even whilst claiming to prefer tags. Power users are faster when using tags, but rarely apply more than a single tag. The meaning of these results is unclear.

It is worth questioning the significance of these studies in the context of a real non-hierarchical file system acting as a personal computer's primary file system. The studies cited only focus on email and "Web-based contexts."

Emotional organisation?

lol?????

Across two studies examining mood (Massey, 2017) it was found that momentary changes in mood can affect user’s FM behaviour, as “sad participants made significantly more folders than happy participants” (p. vii), but that there was no clear relationship between organisation and longer-term trait emotional tendency (e.g., generally happy people didn’t have significantly fewer folders or shallower hierarchies than generally unhappy people).

User infantilisarion, or actually helpful?

There is real research behind some features I personally find annoying — hiding dotfiles, hiding unused directories, ... — However, there is discussion to be had about the consequences affecting public tech literacy.

Navigation improvements

• In graphical settings, highlighting the path to search results. An implementation of this can be seen in the 'settings' app on MacOS.

• Semantic searching.

Metaphors and visualisations

Radicalism

Backward-compatibility is of the utmost importance. Users tend to do better with gradual adoption, and we don't want to invent an entire OS from scratch.

As discussed above, tagging has been investigated for its potential use in providing multiple classification of files, thus obviating maintaining a folder hierarchy. Several systems have implemented this, either by using tags without the folder hierarchy (Seltzer & Murphy, 2009) or in tandem with it (Albadri, Watson, & Dekeyser, 2016; Voit, Andrews, & Slany, 2011). The ubiquity of the tagging concept means it can be offered as an unobtrusive feature (Oleksik et al., 2009) in both local and Web-based FM systems (Hsieh, Chen, Lin, & Sun, 2008) and in document management systems (Ma & Wiedenbeck, 2009).

Most of these novel approaches have had little effect on file management beyond their initial testing. A tagging feature has been introduced to Mac’s Finder application, however, where it is offered alongside the folder hierarchy. This may be the most drastic change that file management will encounter in the near future; because current operating systems deal with files, any software that aims to replace them must still provide users with some access to them (Kaptelinin, 2003), thus prolonging the habit of managing them and therefore the need for such functionality.

In summary, several incremental and revolutionary prospects show promise for changing the nature of file management, but the traditional file and folder metaphor also seems likely to persist. It is tempting to feel that innovative software could at any moment dramatically change the way we interact with digital objects, and that files will therefore soon disappear, and yet the existence of equivalent speculation about FM in the 1980s (e.g., Burton, 1985) suggests this state of uncertainty and promise is not new. Perhaps the staying power of FM is a consequence of it being a standard, if imperfect, computing feature, much like the Sholes (i.e., QWERTY) keyboard layout, which has proven difficult to transition away from. While the above literature shows traditional FM is laden with challenges, it nonetheless provides an essential computing function by allowing users to store, organise, retrieve, share, and interact with many digital items, and FM and its challenges both seem to make for an interesting and productive object of study. Finally, as noted above, many novel and potentially beneficial FM augmentations and interfaces are never adopted by users or integrated into default FM applications. Increasing the use of such software, perhaps through more direct engagement with OS vendors, is therefore an outstanding long-term goal for the PIM and HCI research communities.

The allure of tags

Completely abandoning directories in favour of tags is a siren song for power users, which the average computer user may not benefit from. The tree-crawling performance of navigating a HFS is forgiving for the forgetful. Partial paths and their contents serve to remind the user as they browse:

As described in [cite:@teevan2004perfect]

Often it appeared as if the participant was following a path they could not quite articulate but believed to exist. In the following incident, Rachel described navigating down her directory hierarchy using cues at each level to remind her which step to take next:

Rachel: I didn’t know necessarily how to type that path name from memory and so I used the path completion. [...] I knew what its name was relative to the directory above it. I didn’t know the path down the whole tree.

Interviewer: Did you ever make any false completions, start with the wrong letter or something?

R: No.

A successful NHFS system must offer an equivalent or better exploration experience. What would graph-crawling look like?