urfriendben

Problem

Current filesystems in local machines take too long to find a particular file in archive or even cannot be found if user is unable to recall where in the drive he has stored it or what it was named.

Solution

A filesystem without folders, all files stored in a single level, and each file has tags attached for categorisation and easy search. Uploading of files and administering with tags is made efficient, resulting in a quick storage solution and easy access.

DESIGN PROCESS

Ideation

Coming up with a scope that resolves the problem.

Ideation

Redesigning user workflow in uploading, searching and updating file should take less time and require less actions (clicks) to complete.

Ideation

This should minimally do all that current filesystem can do and further enhance with ability to share much like any cloud drive available.

User Study

Conducted on members of NUS in an hands-on observational study.

Students and Lecturers

Targeting mainly computing members, we seek to evaluate the time taken in navigating a traditional filesystem. With a given computer, how long does it take to find a particular document by searching for a file given its content description (e.g. Year 2 Semester 1 CS3240 Tutorial 1) by clicking through folders manually and searching. Our study showed that almost 90% of users do not recall file name but judging by descriptions, look for relevant folders and step by step looking through the folders for the file and more than 75% of users actually prefer using the search.

Common Actions

Users are then given an alternate computer with the same files, however, this time all files are in a single folder. Based on how tech savvy users are, and through observing the various actions they do to find a file in a flat filesystem, we gather the various actions users are accustomed to so as to include them in our system as users are already familiar with similar actions. All users in the study immediately went for the search and tried all description words given in the scenario. Searching multiple keywords got them the correct file while searching all keywords at once would not. On average users took 4 tries in searching for a particular set of keywords to find the file.

PROTOTYPING

Product

A working prototype was developed as part of Orbital module and no commercialised product was developed. A testing product for proof of concept was developed and showcased with the basic features (uploading and updating) and core search (basic keywords search).

"Folders"

Users were given the ability to upload multiple files related to a set of keywords at once. For example, uploading a tutorial, lecture and lab file all at once, they are able to tag all these files under a certain module and week of the semester. Following that, subsequently adding tags for each individual file.

Searching

Searching files via the tags created meant similar files can easily be found just by a tag. Each user has their own impression of how the tag system can help improve their efficiency of work. Most who tried out the system felt they would commonly use keywords like revision to mark a file for revision when reading week comes. Others like professors felt it would help them since they could tag files under a particular word they remember and these files can easily be found at a later time. Clean UI was also something most liked where despite the tags, the filesystem was not cluttered with so much information (i.e. tags fo the file) however, they were hidden unless more details is shown about a file.

REFLECTION

Usability

Many other features could still be added to make the system a more powerful and easier to use, such as, auto complete tag suggestions for tags that have already been created. However, this proof of concept has really shown us that design is a cummulative process and what we've done is just the first step. Many suggestions have come in on how the system can improve add further improve the efficiency of a user workflow. Prototyping became an important step as these improvements could not have surfaced if the basic model of a new filesystem was trialed. Visualisation can only bring you so far, but something hands-on can help people generate many other ideas without having to constantly keep in mind (in visualisation) how the product is expected to work.