Tuesday, April 27, 2010

Workspaces, Part iv - The things I heard from you

A few days ago, I asked OMG!Ubuntu! readers how they used their Virtual Desktops. The results were, ultimately, very very interesting and not exceedingly surprising; at least not generally speaking. I've been talking quite a bit about workspaces (i, ii, iii) here on my blog, and I have to assume that you're at least somewhat interested if you're here. :D

So what did I find out? I asked 6 questions, and I want to look in particular at the results of 4 of these questions. For a review, the 4 questions I want to look at are:
1. How many Virtual Desktops do you use?
2. How do you navigate within them?
3. What method do you use to use them?
4. What features do you want in virtual desktops?

The first of these questions is the easiest to look at. 37 of you used 4 workspaces, far and above every other number. Check out the chart:

So, admittedly a small-ish sample size of 84 users in this graph (I got another 4 or 5 late admissions, but I already had these saved to png and everything! Ok, so really there was just nothing very different about the late entries). Theres actually more to this than just which one has the most users; theirs definitely some other information we can pull from this. First of all, not a single person used 7 workspaces. Not one. For much of the way across the board, odd numbers have less users than their even counterparts (2 has more than 1, 4 more than 3, 6 more than 5). But once we get up to 9, it has more users then 8. What does that mean?

Quite honestly, I think this speaks to one very clear point. Users want order in their desktop. For most people, it makes the most sense to use even numbers. A majority of users who used 4 desktops had them laid out in a wall, not in a cube or even a line. 16 used a wall, 12 used a line. This is not, by any means, a conclusive study, or a conclusive poll. But it does show an interesting trend, especially given many of the people who used 4 desktops had set applications or tasks assigned to the workspaces, but usually one was set to miscellaneous use, or they stated that they didn't use the 4th one at all.

So, is there anything conclusive? I think that it shows peoples basic trend is towards even number of desktops, and regardless of even or odd, 4 and less are the most practical for most users use.

Lets go back to the charts! Heres one, which looks at peoples basic mode of use for their workspaces:

Use number one, again with a huge margin, is having Assigned Applications or Tasks to workspaces. I know, thats vague. But what we often got was something similar to this: "I keep my browser in workspace #1, put my current documents in workspace #2, Email and IM in #3, and my media player in workspace #4."

Especially once you consider that basically all of the statements really say the same thing. Virtual desktops are being used often for one application, run fullscreen, just to get it out of the way. I mean, that really makes sense, considering you can only use one workspace at a time. But I was surprised that there weren't more people who were running 4 or 5 or more documents all at the same time, and switching entirely based on the exact task. But it seems my original assumption was off; While I tempered my original analysis by saying that it was not either or, it seems that no is really either or when it comes to task-centric or workspace-centric. Rather, it seems a very large majority of users (contained within the polled group of, notably, nerds) have some applications which claim their own space, and use one or two workspaces only for varied, task centric use of virtual desktops. So, while not directly lining up with my original thoughts, it does line up with some more underlying thoughts: that users of Virtual Desktops are using them basically to keep full screen apps like evolution, rhythmbox, and pidgin open and running.


This shows the primary ways of control that users make use of. Now, this one is not a 1:1 chart, as many users use two or more of these control systems. However, we do see some trends. As a whole, its all about the keystrokes. Expo is up there, but there were only 6 people who only made use of Expo, while the other 21 people mentioning expo use it in addition to keyboards shortcuts and/or dragging windows. I would say, easily conclusively, that almost everyone uses keyboard shortcuts the most. Many users only used something besides the keyboard to take a window along with them, such as dragging a window to the screen edge or using expo.


These are the things people want. Honestly? I only see two things here: People want to be sure that their work flow won't get totally screwed to crap by gnome-shell, and some people recognize a desire to see their workspaces be distinct in some ways: Some, just to be distinct in appearance, some, to be totally distinct in terms of appearance, controls available, basic set up. This, for some people, is app specific, some task, some actually work and social specific, giving the user easily switchable workspaces for different uses.

Well, thats some good quality info, in my opinion. More, later!


  1. Posted this on omgubuntu, but could be worth mentioning here as well. A more up-to-date version of what i describe here is what I would like to see:


  2. Awesomely layered out. If the survey would've reached out to maybe a little more Persons it would be more representative.

    Still an awesome starting point for what future Ubuntu UI Design changes should focus on.

  3. I guess people using more than one workspace were more likely to participate in the survey than people not using multiple workspaces.

  4. Jimmy Cederholm, that is a very interesting idea. It has a bit of the taste of one of the few really awesome Mac applications that never made it to other platforms, DragThing. I loved it and it was one of the few things I've missed since switching to GNU/Linux.

  5. Nice Work. I'm amazed at the amount of work you put into this! :)

  6. I must be really odd, since I use 7 workspaces.

    Other than that, I agree with everything.

  7. "Users want order in their desktop. For most people, it makes the most sense to use even numbers."
    I think this conclusion is insufficient and overlooks the arithmetic, geometry, human weaknesses, and imposed defaults involved.

    First the maths. Put simply, it's much easier to make a grid from even numbers. Let's look at the arrangements:

    1: 1 single workspace
    2: 1 across by 2 down, or 2 across x 1 down
    3: 1x3, 3x1
    4: 1x4, 2x2, 4x1
    5: 1x5, 5x1
    6: 1x6, 2x3, 3x2, 6x1
    7: 1x7, 7x1
    8: 1x8, 2x4, 4x2, 8x1
    9: 1x9, 3x3, 9x1

    In each case, the even number has more combinations than the preceding odd number, and with more combinations comes a greater likelihood of use.

    Even domination is further accounted for by the necessity of one single line for odd numbers less than nine—the first odd multiple of the first odd number greater than one. As a line grows, the maximum travel between desktops increases. Adding rows/columns reduces that distance, so is preferable. And when a user adds a second row/column, they end up with an even number of workspaces (even if they don't need all of them) because odd numbers just don't divide by two.

    This grid use also explains why 9 is more popular than 8: in a 3x3 arrangement, the user has a maximum of two workspaces to move across to reach another, compared to three or seven, respectively, for a 2x4 or 1x8 grid. And they get an extra workspace to boot!

    And now the human weaknesses, which could also be described as striving for efficiency. I've already mentioned that limiting the maximum travel is preferable as it saves effort and time, explaining the popularity of a desktop wall over a line—and I'd bet that those line users used fewer workspaces. But another factor comes into play: the user has to keep track of where everything has been put. As the number of spaces increases, the demands placed on the user become tougher, and popularity wanes.

    The seven-workspace arrangement falls at a particularly user-unfriendly nexus of all these factors, requiring a displeasingly long single line as well as being overly taxing on memories. It's hardly surprising nobody claimed to use that number.

    Finally the defaults. The spike at two workspaces is accounted for by being the default. Some people are happy with what's given to them, and see no reason to change. Four is similarly blessed, as the default for the Compiz cube, as well as being the older panel default, brought forward by long-time users.

    So yes, users want order, but, perhaps more importantly, they also want to keep track of and move between windows with a minimum of effort. Because computers are supposed to make things easier, right?

  8. cldx3000: I just didn't know how to get it to more people. Just used what voice I had.

    Anonymous said: Actually we had people who used just one workspace participate as well. Some of the most constructive comments came from them. And my goal was not to see how many people use them, but how they are currently used.

    Jimmy Cederholm: I bookmarked it, and I'll just have to let it ruminate in my head a bit. I like it though. Similar to things I have considered.

    Orjanv, Xubean, Tom: Thanks! Been working hard for a week or so. :D

    Marc: Exactly my point. People want order, and within the constraints of a geometric, 4 sided screen, it makes perfect sense. Your comment does not contradict my conclusion, its exactly something i considered. People actually often stated they had four but only used three, but it looked better. That important data just like everything else I found. It means I wouldn't want to impose something that feels off balanced, for one. So thank you for the analysis, really, but thats exactly the thought process I went through, and just didn't post here. The exact order of commonality of workspaces is exactly as you've described. Regardless of actual need, people set up a certain number of desktops based on other factors, and then try to acclimate themselves to that particular layout.

    Interestingly, many people who used 2 claim they, at one time, used far more, but scaled back.

    but, yes, your comment has valid points, I just wanted you to know I went through the same process and wasn't surprised by these points either. :D

  9. Please don't think I was contradicting your general conclusion—I described it as "insufficient", not incorrect.

    I agree that users want order, but I was trying to point out that users don't choose to have an even number of workspaces to achieve that order; what they choose is ease of use, via a grid arrangement. That choice then results in an even number only as a side effect. You refer to the use of an even number of desktops twice, as if it's actually important, when what you should have referred to at those points was use of grid layouts.

    In this regard, you appear to make a basic mistake of inexperienced analysts, in confusing cause and effect, akin to writing that "people drive cars to burn petrol", rather than "people drive cars to travel quickly, burning petrol as a consequence". You say that my comment matches your thought process, but you didn't include those details in your post, which is why I used the word "insufficient", because in the professional reports of analyses I'm more used to dealing with—which I appreciate you may not have been trained to write, and probably weren't aiming for anyway—simply presenting data followed by a conclusion isn't enough. Without an explanation of why the data lead to any particular conclusion, those conclusions cannot be properly judged. It doesn't matter how brilliant or insightful your thoughts are, if you don't explain them, then how are we to know how much faith to put in the conclusions you declare?

    In adding my comments I was trying to help the analysis by providing (what I thought was) that explanation. If I've done that accurately, then I'm glad. And if I can help out with any future analyses, I'd be glad to do that too.

  10. Kind of late but heres my reply

    Video Reply because i dont like typing =)


  11. Have you considered that Panel application shortcuts are counterproductive compared to the Desktop selector (virtual desktops brings vmware and xen to mind)?

    In my opinion the 'workspace' metaphor suits best. My computer is a printing press, a lab, a library, a communications center, it's an entire factory housed in a little box. Each workspace should be project oriented and, as you mentioned, would not need to be organized suitably to be cloned and the same setup used by the other workstations.

    At present, we cannot use the desktop selector in this way because we need a way to clone the tools across the workspaces in an efficient manner. If I want Chromium on space 1 and 3 I can open two instances and place them there but I wouldn't want to have to do that each time I rebooted.

    Furthermore, there's a severe limitation to our project environment in that we must duplicate various tools and distribute them individually across work areas with no easy way to share the clipboard. That most commonly translates to web-browser data, in my case.

    I overcome the most part of these problems by using Chromium. I can drag a tab out, down to the next desktop and onto the other instance of Chromium. Storing it as a virtual machine image allows me to save and restore the state of my desktops, apps and data while I reboot the host Windows machine. I call it Ubuntoogle and at present I carry it on my key-chain USB device so I can use it and demonstrate the awesomeness that is Ubuntu, on any Windows machine, wherever I may roam.


Note: Only a member of this blog may post a comment.