Loading ...
Sorry, an error occurred while loading the content.

Re: [agile-usability] user expertise and progressive usage (was RE: norman)

Expand Messages
  • Jon Kern
    if you read carefully, you ll see my idea was always based on a per-feature (or group) concept, so to speak. slider that would allow features to go from
    Message 1 of 21 , Nov 3, 2005
    • 0 Attachment
      if you read carefully, you'll see my idea was always based on a per-feature (or group) concept, so to speak.
         " slider that would allow features to go from novice to power user"
      Not across the board, as I agree it would be bad UI.

      The idea is that you can allow the user to peel back the wrapper and access more features at their pace. The system would magically track their level of comfort and expertise and frequency of use. All based on metadata recorded about the UI/features and usages...

      One of the features we built for Together to control the UI was a role-based configuration. An analyst will need to have very different parts of the app available than, for example, a developer. An analyst doesn't need the code editor, debugger, and the like. Of course, it was available if it was really needed. It was just a way to control the default features that were visible and front-and-center, so to speak.

      Of course, a better goal is to provide better UIs. But, it does become very challenging for big honkin' tools like Together or OptimalJ or even IDEs to accommodate varying levels of expertise as a new user goes from novice to power user.

      What sorts of UI design rules of thumb apply for these types of highly complex apps?
      -- jon
      
      


      Larry Constantine said the following on 11/2/2005 10:01 AM:
      Jon Kern wrote:
      ----------
      a valuable addition to something like Word would be the magic customizer
      slider (that i have invented in my head for years when i was building a very
      complex, somewhat horrid at times UI in the incarnation of TogetherSoft's
      ControlCenter UML modeling tool) that would allow features to go from novice
      to power user in terms of: visibility and depth and degree of user control.
      ----------

      It's cute, been proposed many times by many people in one form or another,
      but whether it's a slider or a series of radio button, it turns out to be a
      bad idea.

      (1) A user is not across the board novice, intermediate, or expert in
      interaction style, but varies from one part of the UI to another. Typical
      users are improving intermediates for the 10-20% of the UI they use fairly
      regularly, expert for a few percent of over-learned features, and novices
      for the remaining bulk. But that changes over time and even with respect to
      what they are using the system for at a given moment.

      (2) Having the entire configuration of the user interface change because the
      user slips the slider up or down a notch is extremely disruptive user
      experience.

      (3) Most users do not actually know at what level they are operating or what
      they should tell the system. Even making the choice on setup causes great
      anxiety, as many users fear that if they set the level too low, they will be
      prevented from doing things but if they set it too high they will be
      overwhelmed. Interestingly, when allowed to choose between so-called short
      menus (with reduced options and only basic features) and full menus, the
      vast majority of users prefer long menus.

      A far better approach is based on the progressive usage model (the ski-slope
      model, as it is sometimes known) which supports continuous and incremental
      in-context adaptation of the UI by users to fit evolving interaction style
      and level of expertise. (Covered in our book and several papers.)

      --Larry Constantine, IDSA

    • Jon Kern
      Nice find with that photo app Jon, oh man is that bad. It s almost too painful to compare it to something like Picassa 2. btw: yes, it is a bit exploratory
      Message 2 of 21 , Nov 3, 2005
      • 0 Attachment
        "Nice find with that photo app Jon, oh man is that bad. It's almost too painful to compare it to something like Picassa 2. "

        btw: yes, it is a bit exploratory as far as the UI goes, i think. i bet it is going for a different audience and is purposely trying not to look like every other photo program on the planet (like Picasa). They wanted to make the ability to organize and tag photos through simple "who/what/where" info via drag-n-drop or simple type to get a list of tags or create tags. and to make it so your "Mom" can even share photos with friends.

        it's actually easy to use and I don't mind not having a stupid "file" or "edit" or "help" menu. click one button to share one or more photos... click tags to sort: "Jonny" "Mountain" "Nepal" to narrow down to all pics of Jonny, then Jonny with Mountains, then Jonny + Mountains + in Nepal.

        And praise for Word? That squiggly is what IDEs have been doing for a while: catching errors as you make them.

        -- jon
        
        


        Jade Ohlhauser said the following on 11/2/2005 1:12 PM:
        Well said Larry.
         
        I agree having a single control that changes the UI or other sort of interface "expertise" modes is not a good idea. You're asking the user to make a decision without a lot of information on what the consequences and rewards are and what their own needs are. Also, it's a disruptive jump moving "up to the next level". It's not quite back to square one, but it's a real and/or perceived step backwards in the user's journey to application mastery. Finally, I think if there's something so bad about something that it needs to be removed from the UI for a so called "beginner" user, then it should probably be redesigned.
         
        Nice find with that photo app Jon, oh man is that bad. It's almost too painful to compare it to something like Picassa 2.
         
        And if I may slip in one more UI praise for Word that in my opinion makes up for a lot of bad, the red squiggly. I'm talking about underlining misspelled words right away instead of waiting for a spell check. Brilliant. Of course, they ended up taking the superficial element of the concept too far with the Smart Tags. Talk about missing the point.
         
        Jade Ohlhauser
        Product Manager
        RPM Software                                 
        www.rpmsoftware.com 403-265-6727 x704
         


        <cut>

      • Ron Jeffries
        Hi Jon, ... Reading incredibly carefully, I see the word slider , not sliders , suggesting strongly that there is just one across the board slider. ... Oh
        Message 3 of 21 , Nov 3, 2005
        • 0 Attachment
          Hi Jon,

          On Thursday, November 3, 2005, at 5:06:30 AM, Jon Kern wrote:

          > if you read carefully, you'll see my idea was always based on a per-feature (or group) concept,
          > so to speak.
          >    " slider that would allow features to go from novice to power user"
          > Not across the board, as I agree it would be bad UI.

          Reading incredibly carefully, I see the word "slider", not
          "sliders", suggesting strongly that there is just one across the
          board slider.

          > The idea is that you can allow the user to peel back the wrapper
          > and access more features at their pace. The system would magically
          > track their level of comfort and expertise and frequency of use.
          > All based on metadata recorded about the UI/features and usages...

          Oh ... and I guess from this that you mean an internal slider, not a
          visible one, where some all-knowing being is deciding how I'm doing
          and exposing new features for me, not a slider that I the user can
          adjust?

          > One of the features we built for Together to control the UI was a
          > role-based configuration. An analyst will need to have very
          > different parts of the app available than, for example, a
          > developer. An analyst doesn't need the code editor, debugger, and
          > the like. Of course, it was available if it was really needed. It
          > was just a way to control the default features that were visible
          > and front-and-center, so to speak.

          And was that good? Did actual users really like it? A situation
          today that is like that is Visual Studio 2005, which holds all kinds
          of things back in the Team Version that should in fact be available
          to kids on the street in the Pick Up at Circuit City and Learn C# in
          Your Free Time version.

          VS 2005, to me, is trying to control how I work, based in part on my
          pocketbook and in part on some Vision From Redmond about how I
          should work, rather than letting me control how I work.

          This isn't the same as your example but it's an example of how very
          far awry incremental exposure can go.

          > Of course, a better goal is to provide better UIs. But, it does
          > become very challenging for big honkin' tools like Together or
          > OptimalJ or even IDEs to accommodate varying levels of expertise
          > as a new user goes from novice to power user.

          Yes, it really does. I'm not at all sure that hiding things works
          ... when I use another person's version of Office that isn't
          configured like mine, I often look for things that I'm being
          protected from because the program doesn't realize that a Word God
          has just sat down at the keyboard. I suppose it should be reading
          the Microsoft Customer RFID chip in my brain. Maybe next version.

          > What sorts of UI design rules of thumb apply for these types of
          > highly complex apps?

          I, too, look forward to hearing this answer!

          Ron Jeffries
          www.XProgramming.com
          Example isn't another way to teach, it is the only way to teach.
          --Albert Einstein
        • Desilets, Alain
          And if I may slip in one more UI praise for Word that in my opinion makes up for a lot of bad, the red squiggly. I m talking about underlining misspelled words
          Message 4 of 21 , Nov 3, 2005
          • 0 Attachment
            Message
            And if I may slip in one more UI praise for Word that in my opinion makes up for a lot of bad, the red squiggly. I'm talking about underlining misspelled words right away instead of waiting for a spell check. Brilliant. Of course, they ended up taking the superficial element of the concept too far with the Smart Tags. Talk about missing the point. 
             
            -- Alain:
            Yeah, I LOVE this kind of unobtrusive decorator. When used properly, they can provide a lot of additional information right away when it becomes relevant, but without interrupting the normal flow of work.
             
            The Eclipse IDE is great in that way. For example, as soon as I type a syntax mistake, the culprit token is underlined, plus a red stop sign and a light bulb appear in the margin. Clicking on the red stop sign shows the exact error message from the compiler, and clicking on the light bulb brings up a list of automatic fixes for the error (ex: replace "okBitton with okButton").
            ---- 
             
            Jade Ohlhauser
            Product Manager
            RPM Software                                 
            www.rpmsoftware.com 403-265-6727 x704
             


            From: agile-usability@yahoogroups.com [mailto:agile-usability@yahoogroups.com] On Behalf Of Larry Constantine
            Sent: Wednesday, November 02, 2005 8:02 AM
            To: agile-usability@yahoogroups.com
            Subject: [agile-usability] user expertise and progressive usage (was RE: norman)

            Jon Kern wrote:
            ----------
            a valuable addition to something like Word would be the magic customizer
            slider (that i have invented in my head for years when i was building a very
            complex, somewhat horrid at times UI in the incarnation of TogetherSoft's
            ControlCenter UML modeling tool) that would allow features to go from novice
            to power user in terms of: visibility and depth and degree of user control.
            ----------

            It's cute, been proposed many times by many people in one form or another,
            but whether it's a slider or a series of radio button, it turns out to be a
            bad idea.

            (1) A user is not across the board novice, intermediate, or expert in
            interaction style, but varies from one part of the UI to another. Typical
            users are improving intermediates for the 10-20% of the UI they use fairly
            regularly, expert for a few percent of over-learned features, and novices
            for the remaining bulk. But that changes over time and even with respect to
            what they are using the system for at a given moment.

            (2) Having the entire configuration of the user interface change because the
            user slips the slider up or down a notch is extremely disruptive user
            experience.

            (3) Most users do not actually know at what level they are operating or what
            they should tell the system. Even making the choice on setup causes great
            anxiety, as many users fear that if they set the level too low, they will be
            prevented from doing things but if they set it too high they will be
            overwhelmed. Interestingly, when allowed to choose between so-called short
            menus (with reduced options and only basic features) and full menus, the
            vast majority of users prefer long menus.

            A far better approach is based on the progressive usage model (the ski-slope
            model, as it is sometimes known) which supports continuous and incremental
            in-context adaptation of the UI by users to fit evolving interaction style
            and level of expertise. (Covered in our book and several papers.)

            --Larry Constantine, IDSA

          • Jon Kern
            ... Ron Jeffries said the following on 11/3/2005 8:59 AM: Hi Jon, ... per-feature (or group) concept, ... user ... Reading incredibly carefully, I see the
            Message 5 of 21 , Nov 3, 2005
            • 0 Attachment
              :=)

              Ron Jeffries said the following on 11/3/2005 8:59 AM:
              Hi Jon,

              On Thursday, November 3, 2005, at 5:06:30 AM, Jon Kern wrote:

              >  if you read carefully, you'll see my idea was always based on a per-feature (or group) concept,
              > so to speak.
              >     " slider that would allow features to go from novice to power user"
              >  Not across the board, as I agree it would be bad UI.

              Reading incredibly carefully, I see the word "slider", not
              "sliders", suggesting strongly that there is just one across the
              board slider.
              Yeah, but I knew what I meant :=0
              And, there could be the overarching "slider" based on measured brain wave activity with the skull cap hat shipped with the product.

              > The idea is that you can allow the user to peel back the wrapper
              > and access more features at their pace. The system would magically
              > track their level of comfort and expertise and frequency of use.
              > All based on metadata recorded about the UI/features and usages...

              Oh ... and I guess from this that you mean an internal slider, not a
              visible one, where some all-knowing being is deciding how I'm doing
              and exposing new features for me, not a slider that I the user can
              adjust?
              Hey, I never committed to building this, it was just a fleeting idea that has wafted around the rather shallow UI design center in my brain...

              I mentioned this idea more as a manifestation of the reality that is presented by a complex (possibly overly so) tool.

              > One of the features we built for Together to control the UI was a
              > role-based configuration. An analyst will need to have very
              > different parts of the app available than, for example, a
              > developer. An analyst doesn't need the code editor, debugger, and
              > the like. Of course, it was available if it was really needed. It
              > was just a way to control the default features that were visible
              > and front-and-center, so to speak.

              And was that good? Did actual users really like it? A situation
              Well, I also did it because a major firm was willing to commit just over $1M in product purchases. So, at least they used it, maybe.

              The beauty was, I could configure it externally without hard-coding. Therefore, it wasn't too challenging to roll it out due to our ability to "externally" specify product feature grouping (to sell different product feature mixes).

              today that is like that is Visual Studio 2005, which holds all kinds
              of things back in the Team Version that should in fact be available
              to kids on the street in the Pick Up at Circuit City and Learn C# in
              Your Free Time version.

              VS 2005, to me, is trying to control how I work, based in part on my
              pocketbook and in part on some Vision From Redmond about how I
              should work, rather than letting me control how I work.
              Sounds like a poor implementation of the idea...

              There's the concept of misunderstanding how a user might need to use features.

              There's a concept of features per product edition... get more if you pay more.

              Then there's the magical "Intelli-Slider-per-Meaningful-Feature-Group" adaptive shelter, that doesn't prevent a user overriding feature availability in an easy manner.

              This isn't the same as your example but it's an example of how very
              far awry incremental exposure can go.

              > Of course, a better goal is to provide better UIs. But, it does
              > become very challenging for big honkin' tools like Together or
              > OptimalJ or even IDEs to accommodate varying levels of expertise
              > as a new user goes from novice to power user.

              Yes, it really does. I'm not at all sure that hiding things works
              I'm not sure either...
              ... when I use another person's version of Office that isn't
              configured like mine, I often look for things that I'm being
              protected from because the program doesn't realize that a Word God
              Yes... I have the same problem when working on a Word doc/desktop by someone who doesn't use styles or even the different "tags" that show tabs, spaces, paragraphs, etc.
              has just sat down at the keyboard. I suppose it should be reading
              the Microsoft Customer RFID chip in my brain. Maybe next version.
              I'd go for that, as long as the chip also helped me go through airports more quickly ;=)

              > What sorts of UI design rules of thumb apply for these types of
              > highly complex apps?

              I, too, look forward to hearing this answer!

              Ron Jeffries
              www.XProgramming.com
              Example isn't another way to teach, it is the only way to teach.
                --Albert Einstein
            • Jon Kern
              okay... I had something to do with this UI. i went phishing for some initial reactions. we deliberately threw out all manner of UI
              Message 6 of 21 , Nov 3, 2005
              • 0 Attachment
                <disclaimer>
                okay... I had something to do with this UI. i went phishing for some initial reactions.
                </disclaimer>

                we deliberately threw out all manner of UI convention and tried to start from scratch with a "zero mass" design... it was based on seeing our "moms" (really) struggling horribly with all manner of "standard" photo apps.

                just like the aeron chair was met with distaste at first, it grew on people after they began to really use it. you can get a feel for the software by looking at the 10-minute "walkabout" viewlet. But even better to try it out.

                if you give tidepool a spin, try to organize a few hundred images, create some stories, and share photos with friends, maybe you would have a different opinion from your first gut reaction.
                -- jon
                
                


                Jon Kern said the following on 11/3/2005 8:24 AM:
                "Nice find with that photo app Jon, oh man is that bad. It's almost too painful to compare it to something like Picassa 2. "

                btw: yes, it is a bit exploratory as far as the UI goes, i think. i bet it is going for a different audience and is purposely trying not to look like every other photo program on the planet (like Picasa). They wanted to make the ability to organize and tag photos through simple "who/what/where" info via drag-n-drop or simple type to get a list of tags or create tags. and to make it so your "Mom" can even share photos with friends.

                it's actually easy to use and I don't mind not having a stupid "file" or "edit" or "help" menu. click one button to share one or more photos... click tags to sort: "Jonny" "Mountain" "Nepal" to narrow down to all pics of Jonny, then Jonny with Mountains, then Jonny + Mountains + in Nepal.

                And praise for Word? That squiggly is what IDEs have been doing for a while: catching errors as you make them.

                -- jon
                
                  


                Jade Ohlhauser said the following on 11/2/2005 1:12 PM:
                Well said Larry.
                 
                I agree having a single control that changes the UI or other sort of interface "expertise" modes is not a good idea. You're asking the user to make a decision without a lot of information on what the consequences and rewards are and what their own needs are. Also, it's a disruptive jump moving "up to the next level". It's not quite back to square one, but it's a real and/or perceived step backwards in the user's journey to application mastery. Finally, I think if there's something so bad about something that it needs to be removed from the UI for a so called "beginner" user, then it should probably be redesigned.
                 
                Nice find with that photo app Jon, oh man is that bad. It's almost too painful to compare it to something like Picassa 2.
                 
                And if I may slip in one more UI praise for Word that in my opinion makes up for a lot of bad, the red squiggly. I'm talking about underlining misspelled words right away instead of waiting for a spell check. Brilliant. Of course, they ended up taking the superficial element of the concept too far with the Smart Tags. Talk about missing the point.
                 
                Jade Ohlhauser
                Product Manager
                RPM Software                                 
                www.rpmsoftware.com 403-265-6727 x704
                 


                <cut>

              • Jade Ohlhauser
                Well if you re fishing then I guess you caught me :) I can say for first reaction the faux wood backgrounds and gradients around the controls really turned me
                Message 7 of 21 , Nov 3, 2005
                • 0 Attachment
                  Well if you're fishing then I guess you caught me :)
                   
                  I can say for first reaction the faux wood backgrounds and gradients around the controls really turned me off as did the non-standard widget look and file browse UI in Windows. You would think being a self-proclaimed usability guy I'd try to put my preferences aside and wait for the test data.
                   
                  Jade Ohlhauser
                  Product Manager
                  RPM Software                                 
                  www.rpmsoftware.com 403-265-6727 x704
                   


                  From: agile-usability@yahoogroups.com [mailto:agile-usability@yahoogroups.com] On Behalf Of Jon Kern
                  Sent: Thursday, November 03, 2005 9:13 AM
                  To: agile-usability@yahoogroups.com
                  Subject: Re: [agile-usability] user expertise and progressive usage (was RE: norman)

                  <disclaimer>
                  okay... I had something to do with this UI. i went phishing for some initial reactions.
                  </disclaimer>

                  we deliberately threw out all manner of UI convention and tried to start from scratch with a "zero mass" design... it was based on seeing our "moms" (really) struggling horribly with all manner of "standard" photo apps.

                  just like the aeron chair was met with distaste at first, it grew on people after they began to really use it. you can get a feel for the software by looking at the 10-minute "walkabout" viewlet. But even better to try it out.

                  if you give tidepool a spin, try to organize a few hundred images, create some stories, and share photos with friends, maybe you would have a different opinion from your first gut reaction.
                  -- jon
                  
                  


                  Jon Kern said the following on 11/3/2005 8:24 AM:
                  "Nice find with that photo app Jon, oh man is that bad. It's almost too painful to compare it to something like Picassa 2. "

                  btw: yes, it is a bit exploratory as far as the UI goes, i think. i bet it is going for a different audience and is purposely trying not to look like every other photo program on the planet (like Picasa). They wanted to make the ability to organize and tag photos through simple "who/what/where" info via drag-n-drop or simple type to get a list of tags or create tags. and to make it so your "Mom" can even share photos with friends.

                  it's actually easy to use and I don't mind not having a stupid "file" or "edit" or "help" menu. click one button to share one or more photos... click tags to sort: "Jonny" "Mountain" "Nepal" to narrow down to all pics of Jonny, then Jonny with Mountains, then Jonny + Mountains + in Nepal.

                  And praise for Word? That squiggly is what IDEs have been doing for a while: catching errors as you make them.

                  -- jon
                  
                    


                  Jade Ohlhauser said the following on 11/2/2005 1:12 PM:
                  Well said Larry.
                   
                  I agree having a single control that changes the UI or other sort of interface "expertise" modes is not a good idea. You're asking the user to make a decision without a lot of information on what the consequences and rewards are and what their own needs are. Also, it's a disruptive jump moving "up to the next level". It's not quite back to square one, but it's a real and/or perceived step backwards in the user's journey to application mastery. Finally, I think if there's something so bad about something that it needs to be removed from the UI for a so called "beginner" user, then it should probably be redesigned.
                   
                  Nice find with that photo app Jon, oh man is that bad. It's almost too painful to compare it to something like Picassa 2.
                   
                  And if I may slip in one more UI praise for Word that in my opinion makes up for a lot of bad, the red squiggly. I'm talking about underlining misspelled words right away instead of waiting for a spell check. Brilliant. Of course, they ended up taking the superficial element of the concept too far with the Smart Tags. Talk about missing the point.
                   
                  Jade Ohlhauser
                  Product Manager
                  RPM Software                                 
                  www.rpmsoftware.com 403-265-6727 x704
                   


                  <cut>

                Your message has been successfully submitted and would be delivered to recipients shortly.