FOLDER3D

A Graphical File Management System Supporting Visualisation of File

Relationships

Saturnino Luz

∗

Department of Computer Science, Trinity College Dublin, Dublin, Ireland

Masood Masoodian, Bill Rogers, Simon De Schutter

Department of Computer Science, The University of Waikato, Hamilton, New Zealand

Keywords:

File management systems, 3D user interfaces, Information visualisation, Hierarchical folders, File relation-

ships, Virtual folders, Window management systems.

Abstract:

The desktop metaphor with its hierarchical structure of folders is the basis of almost all graphical ﬁle manage-

ment systems. Despite this popularity, these systems suffer from several problems, including the restrictiveness

of the single inheritance structure of hierarchical ﬁle management. Although various alternative systems have

been proposed, none of these have gained popularity. We argue that the reason for this failure is that these

systems have generally proposed complete alternatives to the hierarchical system, thus ignoring many of its

positive aspects. In this paper we describe a 3D graphical ﬁle management which complements conventional

2D hierarchical folder structures by allowing visualisation of alternative ﬁle relationships.

1 INTRODUCTION

Xerox Star (Smith et al., 1982) introduced to the

world the concepts of the graphical computer desktop,

making it possible for users to directly manipulate

ﬁles and folders as they would in a real physical desk-

top. Despite its popularity, this metaphor suffers from

several shortcomings, including the well-known prob-

lem of single inheritance in a hierarchical structure

used for management of ﬁles within nested folders.

Different operating systems have attempted to solve

this problem by introducing “aliases”, “shortcuts”, or

“symbolic links” to give the users the possibility of

moving between different parts of the ﬁle manage-

ment structure across different paths. In terms of the

graphical visualisation of ﬁles and folders, however,

the problem remains largely unsolved, with all the

major ﬁle management systems showing ﬁles in a sin-

gle inheritance hierarchy of nested folders.

Although alternatives to the desktop metaphor

have been proposed, none of them has been adopted.

We believe that the main reason for this is that

these alternatives have aimed to replace the desktop

∗

The authors, listed in alphabetical order, have con-

tributed equally to the ideas presented in this article.

metaphor completely. In this paper we describe a

3D graphical ﬁle management system, which aims

to complement, rather than replace, the single in-

heritance ﬁle management systems by maintaining a

wider range of relationships between different docu-

ment ﬁles, and providing for a better visualisation of

such relationships through the use of dynamically cre-

ated virtual folders.

2 FOLDER STRUCTURES

Shortly after Xerox Star was released (Malone, 1983)

published his paper on the study of how ofﬁce work-

ers manage their physical desktops. The results of this

study had clear implications for the design of future

systems which based on the desktop metaphor. Un-

fortunately, however, almost all modern graphical ﬁle

management systems have, for the most part, ignored

Malone’s interesting ﬁndings.

Among the results of this study was the obser-

vation that there are two groups of people: those

who manage their desktop “neatly”, and those who

are “messy”. One reason for this difference is re-

lated to the primary type of tasks performed by peo-

149

Luz S., Masoodian M., Rogers B. and De Schutter S. (2010).

FOLDER3D - A Graphical File Management System Supporting Visualisation of File Relationships.

In Proceedings of the International Conference on Imaging Theory and Applications and International Conference on Information Visualization Theory

and Applications, pages 149-152

DOI: 10.5220/0002824801490152

 SciTePress

ple, with some tasks requiring strict “ﬁling” of docu-

ments while others rely on “piling” of documents in

loosely deﬁned groups. Piling of information on the

desktop assists with the process of “reminding” and

helps users manage the cognitively demanding task

of categorising their ﬁles. Malone suggests that to

improve the process of categorisation computer sys-

tems need to allow multiple classiﬁcation as well as

deferred classiﬁcation of information.

Studies of people’s ﬁle management habits on

computers have identiﬁed similar problems with clas-

sifying information. For instance, (Nardi and Bar-

reau, 1997) report that, in their studies, there was a

preference for location-based search for ﬁles, people

archived very little information, and the placement

of ﬁles had a critical reminding function. However,

(Fertig et al., 1996a) argue that the main reason for

these ﬁndings was due to the limitations of the desk-

top metaphor itself. They argue that users may have

different preferences when they “are provided with

richer and more functional interaction environments.”

A more comprehensive study by (Ravasio et al.,

2004) showed that the users of systems based on the

desktop metaphor ﬁnd it impossible to “impose the

same type of hierarchical structure on the screen as

in the ﬁle system”, and because of this they tend to

use “the concepts of thematic proximity and docu-

ment type to group the content on the desktop”, which

is then mainly used as temporary storage. (Ravasio

et al., 2004) also point out that users view information

from at least three different angles: (1) task oriented:

focusing on task to be accomplished, (2) context ori-

ented: focusing on other documentsand tasks at hand,

and (3) content oriented: focusing on the actual infor-

mation contained in documents. The strict hierarchi-

cal folder structure is therefore clearly restrictive in

that a ﬁle can only be placed in one folder, and as

such, can only be viewed from one perspective.

Several attempts at improving this situation have

been made. Lifestreams (Fertig et al., 1996b; Fer-

tig et al., 1996a), for instance, replaces the spatial

aspect of desktop, and acts like a diary of items

which can be viewed across time. One important fea-

ture of Lifestreams is the support for different data

views through “stream ﬁlters”, which divide infor-

mation into dynamically created sub-streams. Sub-

streams are therefore a form of virtual ﬁle organisa-

tion in which, unlike a hierarchical folder structure,

a ﬁle can be in different places. Lifestreams is sim-

ilar to “semantic” systems, such as MIT’s Seman-

tic File System (Gifford et al., 1991). A different

view of time-oriented ﬁle management is provided by

Time2Hide (Lepouras et al., 2008) a system in which

items that the user has not accessed for a while grad-

ually fade on the desktop.

Systems which aim to enhance the spatial group-

ing of related items, such as Data Mountain (Robert-

son et al., 1998), are generally based on the “pile”

metaphor (Mander et al., 1992). Piles are also used in

Presto (Dourish et al., 1999), but Presto allows auto-

matic grouping of items based on their attributes and

provides a space called Vista (similar to a desktop) on

which items can be placed in piles.

The problem with most of these prototypes, and

other systems which allow organisation of informa-

tion using a logical (Dourish et al., 1999), tempo-

ral (Fertig et al., 1996b; Fertig et al., 1996a), or spa-

tial metaphor (Rekimoto, 1999a; Rekimoto, 1999b;

Robertson et al., 1998), is that they propose radi-

cally alternative ways of organising information, as

replacements for the hierarchical folder system, often

without taking into account many of their beneﬁts. A

study by (Jones et al., 2005) has identiﬁed some of

the positive aspects of folder hierarchies. According

to this study folder hierarchies are more than just a

means to re-access information, and they have infor-

mation value in themselves. As well as organising

ﬁles, folders also “represent an emerging understand-

ing of the associated information items and their var-

ious relationship to one another.” The results how-

ever also show that additional information, particu-

larly those that are common across multiple folders,

often have to be “squeezed” into tree hierarchies.

There is, therefore, a need for enhancing the capa-

bilities of the folder structure by supporting relation-

ships between ﬁles across multiple locations as well

as facilitating the visualisation of such relationships.

3 VIRTUAL FOLDERS

In general, relationships between ﬁles can be deﬁned

in terms of their content and their context of use. We

are particularly interested in how content and contex-

tual information can be used as the basis for better

visualisation of ﬁle relationships, and as such, allow

for more effective browsing of ﬁle groupings, while

at the same time retaining the familiar nested folder

tree hierarchy which can provide an effective means

of traversing between folders (Golemati et al., 2008).

Although content information can be automati-

cally extracted by the system, there are some limi-

tations to doing this completely automatically. Any

proposed content-based system should therefore al-

low an easy way of manually associating ﬁles based

on their content. The folder system itself is one way

of doing this, though only in a limited single inheri-

tance hierarchy. In contrast, a dynamically generated

IVAPP 2010 - International Conference on Information Visualization Theory and Applications

150

virtual folder would allow creation of many different

groupings of a single ﬁle within different categories.

Contextual information, on the other hand, can be

extracted automatically much more easily. Examples

of useful context information related to a ﬁle include

links to other ﬁles created or accessed about the time

that a particular ﬁle itself was created or accessed

(Dourish et al., 2000). Once again, although phys-

ical folders are a means of placing contextually re-

lated ﬁles together, dynamically created virtual fold-

ers would allow creation of many different contexts

for a single ﬁle depending on the users real-time ac-

tivities, needs, or preference settings. Another way

of enhancing the hierarchical folder structure is to use

of the ﬁle attributes which are best recalled by com-

puter users (e.g. ﬁle location, type, format, associated

events, etc) and could aid ﬁle retrieval (Blanc-Brude

and Scapin, 2007).

4 FOLDER3D PROTOTYPE

In order to investigate the issues discussed in the

previous sections, we have developed a proto-

type 3D graphical ﬁle management system, called

FOLDER3D, which provides mechanisms for creat-

ing alternative grouping and visualisation of related

ﬁles using the concept of virtual folders. FOLDER3D

prototype is based on the Looking Glass 3D desktop

environment (SUN Microsystems, 2009), and uses its

window management infrastructure. Figure 1 shows

a frontal view of the user interface with its 3D ﬁle

management object containing four panels, which the

user can rotate along the vertical and horizontal axes.

All panels are equally interactiveand display synchro-

nised views of different components of the ﬁle sys-

tem structure. The user can select and navigate across

the folder hierarchy by selecting any icon on any of

the panels. The front panel represents the user’s cur-

rent focus which can be either a folder which the user

could further explore (causing the side panels to be si-

multaneously updated) or a ﬁle whose properties are

displayed on the panel, as shown in Figure 1. The use

of simulated perspective in FOLDER3D implements

a distortion-oriented technique similar to the well-

known techniques of bifocal display and perspective

wall (Leung and Apperley, 1994) which ﬁts in natu-

rally with the platform’s 3D metaphor.

The interface object can be rotated arbitrarily (us-

ing the 3D manipulation primitives available in the

platform) or ﬂipped into pre-deﬁned positions which

highlight each of the different views supported by the

browser through the “handles” provided, which are

represented in the interface by the arrows on the bot-

Figure 1: Front view of FOLDER3D.

tom (left and right) and top of the panel.

Flipping the object to the left brings into focus the

indented tree panel (left panel of Figure 1), which

displays a traditional view of a folder structure as a

hierarchy. From this viewpoint, the user can easily

navigate the ﬁle system while at the same time main-

taining a peripheral display of the details of any ﬁle

selected during this navigation task.

The panel on the right-hand side of the selected

ﬁle panel displays the icons corresponding to the

ﬁles and folders situated on the same level as the se-

lected ﬁle in the ﬁle system hierarchy. Navigation is

also possible through this panel, following a standard

“zoomable” interface interaction mode. However, it

is more likely that the user will use it to inspect ﬁle

properties, which are displayed on the left as the user

moves the cursor over the ﬁle icons.

As the user navigates the ﬁle system, they might

want to explore ﬁles which are related, either in terms

of the content or context, to the ﬁle currently in fo-

cus. FOLDER3D supports the visualisation of related

ﬁles on the virtual folder panel located at the top of

the properties panel. As with the other panels, the

related-ﬁles panel can be highlighted by rotation of

the FOLDER3D object, as shown in Figure 2.

Figure 2: Top panel showing related ﬁles in a virtual folder.

In order to allow visualisation of related ﬁles,

FOLDER3D deﬁnes a set of properties for each ﬁle in

FOLDER3D - A Graphical File Management System Supporting Visualisation of File Relationships

151

the system. These properties, which include the stan-

dard ﬁle properties common to most ﬁle management

systems (e.g. ﬁle type, size, ownership), also list a set

of related ﬁles. Although the current version of our

prototype only allows the list of related ﬁles to be cre-

ated manually, we will also provide a drag and drop

mechanism for interactively moving ﬁles from differ-

ent folders into the virtual folder of a ﬁle. It will also

be possible for the users to select from a list of cate-

gories so as to determine which virtual folder should

be displayed in the top panel of a selected ﬁles. Ex-

amples of such categories include the list of ﬁles that

are related temporally, share common keywords with

the selected ﬁle, have the same properties, etc.

5 CONCLUSIONS

We presented a graphical ﬁle management system

based on a 3D desktop metaphor which combines in-

dented lists, zoomable icon panels and ﬁle property

visualisations with a representation of ﬁle relation-

ships that can be deﬁned across and above the stan-

dard directoryhierarchy. Althoughthese relationships

can be inferred according to ﬁle content (keywords,

semantic similarity, etc) or context (time, task, etc),

FOLDER3D provides a visual representation to sup-

port the deﬁnition of such relationships by the user as

they explore the structure of the ﬁle hierarchy.

We are planning to extend FOLDER3D to support

dynamic creation of folder and ﬁle relationships by

drag-and-drop, to allow selection of categories of re-

lationships displayed on the top panel, and to enable

the use of FOLDER3D as a dynamic desktop back-

ground.

REFERENCES

Blanc-Brude, T. and Scapin, D. L. (2007). What do people

recall about their documents?: implications for desk-

top search tools. In IUI ’07: Proc. of the 12th intl.

conference on Intelligent user interfaces, pages 102–

111. ACM.

Dourish, P., Edwards, W. K., LaMarca, A., Lamping, J.,

Petersen, K., Salisbury, M., Terry, D. B., and Thorn-

ton, J. (2000). Extending document management sys-

tems with user-speciﬁc active properties. ACM Trans-

actions on Information Systems, 18(2):140–170.

Dourish, P., Edwards, W. K., LaMarca, A., and Salisbury,

M. (1999). Presto: an experimental architecture for

ﬂuid interactive document spaces. ACM Transactions

on Computer-Human Interaction, 6(2):133–161.

Fertig, S., Freeman, E., and Gelernter, D. (1996a). “ﬁnd-

ing and reminding” reconsidered. SIGCHI Bulletin,

28(1):66–69.

Fertig, S., Freeman, E., and Gelernter, D. (1996b).

Lifestreams: an alternative to the desktop metaphor.

In CHI ’96: Conference companion on Human fac-

tors in computing systems, pages 410–411. ACM.

Gifford, D. K., Jouvelot, P., Sheldon, M. A., and O’Toole,

Jr., J. W. (1991). Semantic ﬁle systems. SIGOPS Op-

erating Systems Review, 25(5):16–25.

Golemati, M., Katifori, A., Giannopoulou, E. G., Daradi-

mos, I., Vassilakis, C., Lepouras, G., and Halatsis, C.

(2008). An interview-based user study on the use of

visualizations for folder browsing. In Proceedings of

the 12

International Conference Information Visual-

isation, 2008. IV ’08, pages 106–112.

Jones, W., Phuwanartnurak, A. J., Gill, R., and Bruce, H.

(2005). Don’t take my folders away!: organizing per-

sonal information to get ghings done. In CHI ’05:

CHI ’05 extended abstracts on Human factors in com-

puting systems, pages 1505–1508. ACM.

Lepouras, G., Papatriantafyllou, A., Katifori, A., and Dix,

A. (2008). Time2hide: spatial searches and clutter

alleviation for the desktop. In AVI ’08: Proc. of Ad-

vanced visual interfaces, pages 355–358. ACM.

Leung, Y. K. and Apperley, M. D. (1994). A review and tax-

onomy of distortion-oriented presentation techniques.

ACM Trans. Comput.-Hum. Interact., 1(2):126–160.

Malone, T. W. (1983). How do people organize their desks?:

Implications for the design of ofﬁce information sys-

tems. ACM Transactions on Information Systems,

1(1):99–112.

Mander, R., Salomon, G., and Wong, Y. Y. (1992). A “pile”

metaphor for supporting casual organization of infor-

mation. In Proc. of CHI’92, pages 627–634. ACM.

Nardi, B. and Barreau, D. (1997). “ﬁnding and reminding”

revisited: appropriate metaphors for ﬁle organization

at the desktop. SIGCHI Bulletin, 29(1):76–78.

Ravasio, P., Sch¨ar, S. G., and Krueger, H. (2004). In pur-

suit of desktop evolution: User problems and practices

with modern desktop systems. ACM Transactions on

Computer-Human Interaction, 11(2):156–180.

Rekimoto, J. (1999a). Time-machine computing: a time-

centric approach for the information environment. In

UIST ’99: Proc. of the 12th annual ACM Symp. on

User interface software and technology, pages 45–54.

ACM.

Rekimoto, J. (1999b). Timescape: a time machine for the

desktop environment. In Proceedings of CHI ’99,

pages 180–181. ACM.

Robertson, G., Czerwinski, M., Larson, K., Robbins, D. C.,

Thiel, D., and van Dantzich, M. (1998). Data moun-

tain: using spatial memory for document manage-

ment. In UIST ’98: 11th ACM symposium on User

interface software and technology, pages 153–162.

ACM.

Smith, D. C., Irby, C., Kimball, R., Verplank, W., and

Harslem, E. (1982). Designing the star user interface.

Byte, 7(4):242–282.

SUN Microsystems (2009). Project Looking Glass.

https://lg3d-core.dev.java.net/, (10th November

2009).

IVAPP 2010 - International Conference on Information Visualization Theory and Applications

152