Malone on How People Organize their Desks

In reading this article, I was distinctly reminded of a gif that demonstrates the history of the desktop computer ( i.imgur.com/zBANAHm.gif ). Many of Malone’s insights about reminding and finding are aspects of today’s modern computer, however, I’d argue that some the benefits he describes (files and folders/piles) haven’t brought the benefits he had hoped, and in some respects the messiness of the digital desktop has lost some things.

With respect to reminders, I’d argue that the abstraction of folders as well as “desktop” clutter contribute a desensitization of visual reminders. Desktops are either so overwhelming cluttered that is impossible to attribute any sort of reminder from the information or so abstracted that receiving a reminder requires a great deal of effort. There’s a parallel that can be made in today’s world for push notifications– we receive so many of them that it’s so easy for us to disregard them. I’ve ignored 3 during this post– “Oh hey there, Duolingo Push Notif; That’s what’s up. *Clears notifs*”.

We can search through mounds of files and piles easily on the computer, more so than the piles we have on our desks. But it’s a lot more difficult to find something for which you can’t remember where you placed it and you don’t know what classifications were given to it. For example, consider starting a new file for quickly jotting down notes. You then save the file without much thought placed into where the file will be saved or its name. Months later, you attempt to retrieve the said file without remember much of anything about it except that it had important notes in it. Search fails here- how do you search for something when you know it exists but you don’t know where or what to look for? Modern search on desktop computers also doesn’t index the majority of file contents. In the physical realm, you may not know what the file is that you’re looking for, or even where but the location is limited by your physical space. In the digital realm, this isn’t the case. In addition, scanning a file is more burdensome than simply glancing at a hard copy.

Another element that seems to be misrepresented is that organizing information on the computer in the same manner which we do in the physical world allows for more precise and easy organization. While it’s true that we can do more since the space is no longer limited, I’m not sure that it adds more to the process than search functionality. Messy people aren’t necessarily compelled to be more organized on a digital computer because the functionality exist. And organization is still just as laborious on the computer– e.g. spending 8 hours organizing categories, folders, etc digitally.

The desktop metaphor is interesting but perhaps this metaphor isn’t necessarily the best. We’ve been habituated to organization styles of the physical world– files and piles are very much rooted in our notions a world with limited space, but our minds (at least not consciously) as well as computers don’t have this limitation. Humans organize information through connections between thoughts and ideas. On the surface, things may seem unrelated but there’s a graph of associations through which we can quickly find information. Perhaps this model is a more fruitful approach to computer information system design.

I agree with the above points about different modes of storage being appropriate for different situations. One of the issues with storage in most traditional designs for computing systems is that we're confined to folders and subfolders and file names. This scheme works fairly well for people working within defined processes, where every file comes with a specific folder it should be a part of, but it doesn't always work so well when projects are nebulous, exploratory, or creating new categories of things. Often times in research we start out with a different set of categories than what we end up with.

This issue is particularly pronounced in the literature review process, as our understanding of what the form of the network should be emerges from our understanding of the individual pieces of research; the network should naturally shift and change shape and structure as we explore more of the individual pieces. A traditional folder/subfolder structure isn't going to be very good at holding these. I've already re-mapped my library of papers and re-named my categories more than once while I've been here. Certain tools help with this process by providing multiple ways of categorizing literature (e.g. folders, tags, labels), but it still takes quite a bit of work to re-map everything when a change in perspective takes place. As discussed above, a better way to go about doing this might be to support fuzzy retrieval, but I think there is still opportunity to explore the creation of useful network structures (and visualizations) beyond the classic folder trees.

In some ways, I think it's hard to remove the example of modern GUI desktop interfaces when reading a paper like this. I have a hard time understanding what exactly the interfaces were like when they made these suggestions. I assume most computers were using DOS/Terminal style text entry at this point in time which really does highlight how misaligned file/folder computer structures are from the way people organize and think about things. I do agree that a lot of the issues lie in the need to make organizational choices before the entire structure is known. However, I think modern systems have adopted a lot of useful approaches though. The 'Recent Documents' list in most applications is useful. Being able to reorder file contents by category/name/date/etc.

I think a lot of neat/messy is connected to how well defined a problem/workflow is. If something is going to be done a lot, there's benefit to optimizing it. If it's research, then spending too much time optimizing something that never be repeated is not a good use of time. Also, the ability to procrastinate by organizing messy things should be considered before we optimize digital information storage/retrieval too much.

I agree with previous posts in that digital desktops don’t help much with information classification, people who are messy with their desks may also be less organized with a digital computer.
I think one thing that is not mentioned in the paper about categorization is how to get rid of files/stuff that the person isn’t using. Although with a digital computer, space is no longer an issue, but in my own experience of organizing desks, getting rid of stuff that I’m not using is very important for keeping my desk neat. Especially when I clearly know that I can access the file somewhere else. For example, sometimes I classify papers into different folders and give them a name, but I ended up finding that when I need to find a paper, searching it in google scholar is much quicker than retrieving it from my own computer.

Building upon what has been discussed about fuzzy retrieval. In the case of classifying paper, one helpful way is helping/enabling the users to add some personalized notes to each document, which is dynamic as the user’s understanding of the information changes, thus creating links between the user’s mindset and the files. For example, when the person thinks about something that’s not directly related to the document, the document can be retrieved as well. I think associating the document with some keywords that already exist in the user’s mindset could help.

Looking around the desks in my office, some are quite cluttered (shoutout to myself and Xu), and none of them look like the clutterless desk from the end of the GIF Mike shared. I'd be curious to see a study comparing/contrasting physical and digital organizational styles-- I feel like there could be a lot to learn from that.

But to other people's points, should we really look at the physical desk to help us understand how to organize digitally? It seems like we're pretty clueless when it comes to organizing our desks; as the paper states, even people who 'successfully' desk organizers are inefficient in terms of the costs/benefits of time spent organizing and value received from it. Rather than finding ways that digital systems can improve on an existing kind of shitty model, maybe we should be trying to come up with a new model entirely.

Is the desktop just a lazy analogy? Like, we put the computer on top of a desk, so then we thought, yeah, let's try to replicate what we did before with this new thing. But we're not quite doing the same things, at least not anymore. On computers, we're using distributed tools, we're communicating and collaborating with other people, we're foraging for new info, we're engaging in virtual worlds, we're accessing enormously large archives, we're accessing resources during off-work hours (entertainment, etc.). These aren't things that were ever confined to our desktops in the non-digital era, so why use desktops as the primary analogy? Perhaps instead, if we want to think about how to organize such systems, we should think about how we might organize a backpack while participating in a scavenger hunt (okay, terrible example... but the point is, there might be a better analogy out there). I also second Nathan's suggestion that maybe we shouldn't even be concerned with organization at all, as better search tools may ultimately be more beneficial to retrieval of information. (All this said, also cognizant that this paper is from 1983. So at the time, I think this study and the analogy 'made sense.' But maybe we now need a different model).

Is it safe to consider that piles are the resultant of over-categorisation? As in, when the object being classified falls into multiple categories, the user would tend to mark it as a distinct object, incidentally, the pile is a collection of distinct objects. The object in the pile could either fall into "important" or "trivial" section. My argument is that whenever there is a possible over-categorisation, the user feels that the object is either too important to be classified as a particular category or too trivial to fall under any category. Therefore, a pile could be generalised as a collection of [important/trivial] object regardless of the category (here, the classification/categorisation is of no importance).

So, when the author talked about multiple-classification solving the issue of classifying, the importance of the object classified is "finding", rather than "reminding". But, the piles signify reminding and it functions on the level of importance ahead of object's class. How can the classifier objectively pile such objects? Can a classifier circumvent this scenario as efficiently as humans?

I agree that the desktop metaphor seems to have outlived its usefulness. It may have been useful early on to transition people from physical desktops to virtual workspaces, but when there are so many aspects that fall short in important ways (one example: lack of clear prioritization), it may need rethinking. As others have said, for finding, our existing file hierarchies are useful enough (though that's more often due to the searchbar than any folder-based categorization), and the chronological data address one of Malone's points about the usefulness of chronological organization. But, as others have said, reminding seems to be missing (other than notifications... but that isn't really a feature of the file structure and organization). That is, notifications and alarms are active reminders, but what's missing is the spatial ability of a file placed in the corner of a desktop to passively remind you. Personally, I haven't seen my computer desktop in weeks. Is my workspace messy? (perhaps). However, when I was teaching, there was a (unfortunately) near-constant flow of papers coming in (from students) and going out (after being graded), with other piles of papers of varying priority (to be graded, to be graded today, etc). I'd like to think I was fairly decent at keeping my physical paper organization process organized. So why do I find the digital version so much more difficult? Maybe the consequences of messiness are lower... with searchability, I can instantly find anything (that I can remember an appropriate keyword for), or search chronologically in a way I couldn't with paper piles.