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

Re: user centered design overlaps with traditional analysis?

Expand Messages
  • Jeff Patton
    ... artifacts hinder a project in terms of productivity, infighting and politics. Or the process was to deliver artifacts, not value. ... because of
    Message 1 of 26 , Feb 3, 2005
      --- In agile-usability@yahoogroups.com, "Chris" <chris@p...> wrote:
      > Accidentially sent it out.
      >
      > To finish my thought...
      >
      > Sometimes verbal is better than written. It depends. I saw
      artifacts hinder a project in terms of productivity, infighting and
      politics. Or the process was to deliver artifacts, not value.
      >
      > Found things usually don't fail from a missing artifact. Usually
      because of communication - uncontrolled or lack of.

      Glad you finished your thought. Couldn't agree more. On my current
      agile project we document stuff we know about a story and post it on
      a wiki. Our expert users sit feet from us. We've noticed a problem
      where the expert user finds out an important piece of information,
      then updates the document to include it. Now folks have already read
      the document, and don't think to scan it for changes/revisions. At
      the end of the iteration when we've missed an important detail, and
      can see it right there in the document, our first thought is "why
      didn't we just talk about this?" The artifact was present - but gave
      folks a false sense of security.

      thanks for your post.

      -Jeff
    • Chris
      Sometimes a visual channel is not appropriate. Depends again on the audience. Myself I prefer visual. Seen others more productive with written word. It depends
      Message 2 of 26 , Feb 3, 2005
        Sometimes a visual channel is not appropriate. Depends again on the audience. Myself I prefer visual. Seen others more productive with written word. It depends on a person's logical systems and beliefs. The common problem is information complexity and organization.

        -----Original Message-----
        From: Josh Seiden <joshseiden@...>
        Date: Thu, 3 Feb 2005 07:36:50
        To:agile-usability@yahoogroups.com
        Subject: Re: [agile-usability] user centered design overlaps with traditional analysis?


        Well yes, the point is about communication, not about
        artifacts.

        The fact that the artifacts were missing an entire
        dimension of communication is a symptom of a
        communication problem.

        But paradoxically, if the team had felt that the
        artifacts required a visual channel, then they would
        have had to think about the problem in visual
        terms--and would have been forced to think through the
        problem more completely.

        JS

        --- Chris <chris@...> wrote:

        >
        > Accidentially sent it out.
        >
        > To finish my thought...
        >
        > Sometimes verbal is better than written. It depends.
        > I saw artifacts hinder a project in terms of
        > productivity, infighting and politics. Or the
        > process was to deliver artifacts, not value.
        >
        > Found things usually don't fail from a missing
        > artifact. Usually because of communication -
        > uncontrolled or lack of.
        >
        > -----Original Message-----
        > From: "Chris" <chris@...>
        > Date: Thu, 3 Feb 2005 14:47:45
        > To:agile-usability@yahoogroups.com
        > Subject: Re: [agile-usability] user centered design
        > overlaps with traditional analysis?
        >
        >
        > Hmmm... Been my experience that documents and
        > artifacts are communication channels geared to a
        > particular audience... That it is an agreement of
        > what is to be done. If I were to write something for
        > the CEO, it would look very different than if it was
        > for a programmer. Even if I was conveying the same
        > information. Sometimes verbal works better than
        > written to
        >
        > -----Original Message-----
        > From: "Joshua Seiden" <joshseiden@...>
        > Date: Wed, 2 Feb 2005 23:18:24
        > To:<agile-usability@yahoogroups.com>
        > Subject: RE: [agile-usability] user centered design
        > overlaps with traditional analysis?
        >
        >
        >
        > > I've been making the assertion over the last
        > couple months that
        > what
        > > traditional analysts are doing overlaps to a large
        > degree with
        > what
        > > UCD folks do. I'd be curious if others agree.
        >
        > [snip]
        >
        > > And finally, can anyone tell any stories about how
        > they've
        > effectively
        > > [or ineffectively] combined analysts and UCD
        > people?
        >
        > Jeff,
        >
        > I agreee that there is a lot of overlap but as Kay
        > said, a lot of
        > the foundational assumptions are different.
        >
        > What's maddening is that the goals of the analyst
        > and UX person
        > are similar, the differences in assumptions are
        > rather subtle,
        > and the tools similar.
        >
        > Witness the thread last year on use cases that
        > Alistair and I
        > participated in. It took a long time to work out the
        > similarities
        > and distinctions between the narrative tools he
        > calls Use Cases,
        > and the one I call Scenarios. And frankly, a few
        > months later,
        > I've still got a pretty slippery grasp on the
        > subject.
        >
        > Now, a story:
        >
        > On my current project, I'm working with a team that
        > is building a
        > supply-chain management application. The application
        > will be
        > used by an organization of 500+ people, working in a
        > global
        > sourcing operation. The team (composed of 6 outside
        > business
        > analysts, 6 inside business analysts and user
        > representatives, 3
        > inside technology analysts, and a large vendor team
        > with
        > expertise in this business area) worked for 8 months
        > producing
        > functional spec documents. When they saw the first
        > builds coming
        > in from the development team, they realized they
        > needed someone
        > to help with UI design, and called me in.
        >
        > Here's what I found: The team has completed about 10
        > functional
        > spec documents, covering 10 of the 20 modules of the
        > system. Each
        > spec doc is about 80 pages long. Each contains use
        > cases,
        > business rules, and detailed field definitions. They
        > are detailed
        > and carefully considered documents. But they're not
        > enough.
        >
        > Some observations:
        >
        > 1. The functional specs have no drawings in them.
        > They are text
        > documents, with long paragraphs, complex indented
        > bulleted lists,
        > and long multi-page tables. Thus to understand the
        > spec each
        > reader must work hard to visualization the system in
        > his or her
        > head.
        >
        > 2. The specs describe *the way the team wants the
        > software to
        > work* rather than the business problem to solve. In
        > other words,
        > the specs are not problem descriptions,
        > ("requirements"
        > documents) but solution descriptions. To describe a
        > solution,
        > one has to envision a solution, which is what this
        > team did, but:
        > -- they embodied their vision in text, not pictures
        > -- the visioning is thus implicit, not explicit.
        > -- There is no way to be certain that anyone was
        > envisioning the
        > same thing.
        >
        > 3. Some of the specs have been written by analysts
        > with a
        > background in software, but others are written by
        > business
        > experts. Some have good visualization skills, some
        > don't.
        > Regardless, none are experts in visualizing software
        > Uis, so each
        > imagined certain functions working in certain ways.
        > Some way
        > database tables. Some saw the relationships between
        > objects. Some
        > saw Excel. Some saw SAP. Regardless of the relative
        > visualization
        > skills of each reader, there is a high probability
        > that none will
        > be visualizing the same thing.
        >
        > Conclusion: This team created and documented a
        > design, but had no
        > designer working to visualize that design--to turn
        > the abstract
        > concrete. Futhermore, they didn't think they needed
        > one! It
        > wasn't until they saw the project starting to head
        > south that
        > they realized they had to augment their team.
        >
        > So how does this tie back to Jeff's original point?
        >
        > The analysis team did excellent work, but they were
        > missing an
        > entire "dimension" of specification. Without a
        > visual channel,
        > the specs were incomplete, and thus unworkable. I
        > couldn't have
        > done my work without the analysis in place, but if I
        > had come on
        > early, my needs would have changed the structure of
        > the analysis.
        > We probably would have covered the same
        > material--and perhaps
        > made most of the same decisions, but we would have
        > organized the
        > work and the work product very differently, and we
        > would have
        > saved a lot of wasted effort. I think the reverse is
        > true as
        > well--if I had been working on the problem for eight
        > months
        > without "traditional analysis", I would have missed
        > a dimension
        > of specification, and would have had to go back,
        > restructure, and
        > re-spec. (But I never would have gotten 8 months
        > down the road
        > first ;-)
        >
        > JS
        >
        >
        >
        === message truncated ===




        Yahoo! Groups Links









        Sent wirelessly via BlackBerry from T-Mobile.
      • Josh Seiden
        ... The big problem is that it was not complete. Most of the work that was there was appropriate. 5-10 percent was misguided because it made incorrect
        Message 3 of 26 , Feb 3, 2005
          >
          > In your story you describe what the analysts
          > produced as a design
          > specification - one missing a visual channel. I
          > wonder if you think
          > the design was an appropriate one?

          The big problem is that it was not complete. Most of
          the work that was there was appropriate. 5-10 percent
          was misguided because it made incorrect assumptions
          about the user interface.

          But because it did not specify a user interface, it
          left open a giant risk factor.

          > I suspect to
          > help this company
          > out that you had to backtrack a little and learn
          > about about the
          > people using the software and what they were trying
          > to accomplish by
          > doing so. Is that true? If so, how was that effort
          > met? Did you
          > find any surprises doing that?

          I did have to backtrack, and I'm still doing it. But
          the timeline is such that not much backtracking is
          possible.

          The team include a great many users, ex-users, and
          subject matter experts, so I am reviewing the work
          with them.

          In this case though, it's not that the team didn't ask
          the right questions or gather the right data. They
          just did a different form of analysis that I needed to
          do.

          So my "backtracking" is really a matter of reviewing
          the work with them and performing a different kind of
          analysis on the data.

          > Was there
          > functionality designed into
          > the spec that wasn't needed? Was there
          > functionality left out of the
          > spec that was?

          This team has been very detail oriented, so the specs
          represent a documentation of every possible thing that
          could ever happen. (What if a fire alarm sounds while
          the user is eating a low-fat tuna sandwich while
          entering this data? Is that different from the case in
          which the tuna sandwich is made with regular tuna
          salad?")

          My starting place has been to separate out the
          probable from the possible.

          > And, finally, assuming some of that
          > stuff whas true,
          > giving bad news about an expensive-to-create design
          > spec usually
          > results in the death of a messenger. Assuming some
          > of that went on,
          > how was your resulting advice met?

          Well, the project leaders have been very receptive to
          my work. They saw the first UIs coming in, and saw the
          future of the project suddenly turn black. They have
          spent a lot of time and money on this and are willing
          to spend a little more to ensure success.

          One measure of success: the executive sponsors saw my
          first design proposals, and demanded a *project delay*
          in order to ensure that the improved UI designs were
          included in the first release. The sponsors went to
          the well for a very large dollar amount to fund the
          extra three months of development work that this
          change requires.

          > Thanks again for the post.

          Thanks again for the list.
        • Sue Heim
          ... Can t your subscribe to the Wiki, so that changes to any document are sent out to the appropriate parties? This would ensure, anyways, that folks who ve
          Message 4 of 26 , Feb 3, 2005
            Jeff:
            >Glad you finished your thought. Couldn't agree more. On my current
            >agile project we document stuff we know about a story and post it on
            >a wiki. Our expert users sit feet from us. We've noticed a problem
            >where the expert user finds out an important piece of information,
            >then updates the document to include it. Now folks have already read
            >the document, and don't think to scan it for changes/revisions. At
            >the end of the iteration when we've missed an important detail, and
            >can see it right there in the document, our first thought is "why
            >didn't we just talk about this?" The artifact was present - but gave
            >folks a false sense of security.

            Can't your subscribe to the Wiki, so that changes to any document are sent
            out to the appropriate parties? This would ensure, anyways, that folks
            who've already read the doc will be notified that there is a revision, and
            they oh yeah, maybe they oughta go check it out...?

            ...sue
          • Jade Ohlhauser
            I like this conversation because I ve found spec documentation to be the most important design & development tool I have. And I try to use *agile documentation
            Message 5 of 26 , Feb 3, 2005
              I like this conversation because I've found spec documentation to be the most important design & development tool I have. And I try to use *agile documentation creation*. I think it would be disastrous to simply take a design team, spend x months writing a complete spec down to the finest detail, then "throw it over the wall" to the developers. Our specs grow along with the development the code (in fact we keep them in source control).
               
              When it comes to specs I believe width is more important to depth. Makes sure the spec is end to end, but it's OK if you can't go into detail on everything in between right away. However, where there is a lack of details or an unknown make sure it is clearly identified. This way we can get started reviewing the spec and talking about the feature sooner. With the unknowns or potential issues flagged it means they can't "go unspoken" until the end of development and blow up. As development goes on the spec also starts "filling in"
               
              For me I think the most valuable things of written specs are they force you to think of the complete solution and they remove ambiguity.
               
              Some things I've learned about good specs:
               
              1. Every feature/module starts out with a 1 sentence summary. That headline style introduction is then followed by a 1 paragraph summary. It's important to ease people into the details like this, as I think Joshua knows all too well :)
               
              2. In addition to the summary, is paragraph (or so) summarizing the business need for the feature. It's here where I defend (in brief) why this feature should be built including mentioning specific customers if applicable. Now you could argue that developers don't need this information, but I've found it to be valuable for everyone to know the business influences. The same applies to the next section... 
               
              3. Every feature also includes a sentence (or so) that says what the goal of this feature is. What is the user trying to accomplish? Of course this will be covered extensively in the meat of the spec, it's good to ease people in and provide them with the big picture view. This also gives people the challenge the spec, i.e. does the feature as spec meet the goal it says it's supposed to.
               
              4. Those first 3 may seem to overlap, but that's OK. Giving people a little extra info is far better than not giving them enough or expecting them to build a visual model of the feature from details only. Imagine trying to visualize a new building from blueprints alone. It's possible, but how much easier is it when you also get a little cardboard model. Also, the above 3 each approach the overview of the feature from a different angle: the functionality, the reason for building it, and the end goal of the user.
               
              5. Each spec document must have 1 author only. Many people will contribute, but 1 person needs to own the document and control the addition, editing, and removal of content.
               
              6. There shouldn't be any rigid structure or formatting rules. For our specs, the summary, need, and goal are required, but they are followed by the large "details" section which can be any format, allowing the author to do what's best for the given feature. We do have a set style template everyone uses for the fonts and such, but we don't impose any rules on things like how deep to nest x type of list or you have to include a list of every user type, etc. Every feature has unique requirements so this gives the author the freedom to decide what's best. It also makes the writing effort about the content, not the document itself.
               
              7. It's OK to be informal or even attempt to be funny. This makes the spec easier to read and easier to write.
               
              8. When there is a long list of variables or something, it should be move to an appendix at the end of the document. This helps keep the reading flow. The same idea applies to using footnotes where applicable for off-topic notes, term definitions, etc.
               
               
              Our specs:
              We have various functional spec Word documents, each containing several features. Each one of these documents has a matching technical spec document. These documents reside in our StarTeam source control. Along with the functional specs there are page sketches (using less now) and HTML mockups (using more now).
               
              Jade Ohlhauser
              Product Manager
              RPM Software                          
              www.rpmsoftware.com 403-265-6727
               


              From: Chris [mailto:chris@...]
              Sent: Thursday, February 03, 2005 8:28 AM
              To: agile-usability@yahoogroups.com
              Subject: Re: [agile-usability] user centered design overlaps with traditional analysis?

              Accidentially sent it out.

              To finish my thought...

              Sometimes verbal is better than written. It depends. I saw artifacts hinder a project in terms of productivity, infighting and politics. Or the process was to deliver artifacts, not value.

              Found things usually don't fail from a missing artifact. Usually because of communication - uncontrolled or lack of.

              -----Original Message-----
              From: "Chris" <chris@...>
              Date: Thu, 3 Feb 2005 14:47:45
              To:agile-usability@yahoogroups.com
              Subject: Re: [agile-usability] user centered design overlaps with traditional analysis?


              Hmmm... Been my experience that documents and artifacts are communication channels geared to a particular audience... That it is an agreement of what is to be done. If I were to write something for the CEO, it would look very different than if it was for a programmer. Even if I was conveying the same information. Sometimes verbal works better than written to

              -----Original Message-----
              From: "Joshua Seiden" <joshseiden@...>
              Date: Wed, 2 Feb 2005 23:18:24
              To:<agile-usability@yahoogroups.com>
              Subject: RE: [agile-usability] user centered design overlaps with traditional analysis?



              >
              I've been making the assertion over the last couple months that
              what
              >
              traditional analysts are doing overlaps to a large degree with
              what
              >
              UCD folks do.  I'd be curious if others agree.

              [snip]

              >
              And finally, can anyone tell any stories about how they've
              effectively
              > [or ineffectively] combined analysts and UCD
              people?

              Jeff,

              I agreee that there is a lot of overlap but as Kay said, a lot of
              the foundational assumptions are different.

              What's maddening is that the goals of the analyst and UX person
              are similar, the differences in assumptions are rather subtle,
              and the tools similar.

              Witness the thread last year on use cases that Alistair and I
              participated in. It took a long time to work out the similarities
              and distinctions between the narrative tools he calls Use Cases,
              and the one I call Scenarios. And frankly, a few months later,
              I've still got a pretty slippery grasp on the subject.

              Now, a story:

              On my current project, I'm working with a team that is building a
              supply-chain management application. The application  will be
              used by an organization of 500+ people, working in a global
              sourcing operation. The team (composed of 6 outside business
              analysts, 6 inside business analysts and user representatives, 3
              inside technology analysts, and a large vendor team with
              expertise in this business area) worked for 8 months producing
              functional spec documents. When they saw the first builds coming
              in from the development team, they realized they needed someone
              to help with UI design, and called me in.

              Here's what I found: The team has completed about 10 functional
              spec documents, covering 10 of the 20 modules of the system. Each
              spec doc is about 80 pages long. Each contains use cases,
              business rules, and detailed field definitions. They are detailed
              and carefully considered documents. But they're not enough.

              Some observations:

              1. The functional specs have no drawings in them. They are text
              documents, with long paragraphs, complex indented bulleted lists,
              and long multi-page tables. Thus to understand the spec each
              reader must work hard to visualization the system in his or her
              head.

              2. The specs describe *the way the team wants the software to
              work* rather than the business problem to solve. In other words,
              the specs are not problem descriptions, ("requirements"
              documents) but solution descriptions.  To describe a solution,
              one has to envision a solution, which is what this team did, but:
              -- they embodied their vision in text, not pictures
              -- the visioning is thus implicit, not explicit.
              -- There is no way to be certain that anyone was envisioning the
              same thing.

              3. Some of the specs have been written by analysts with a
              background in software, but others are written by business
              experts. Some have good visualization skills, some don't.
              Regardless, none are experts in visualizing software Uis, so each
              imagined certain functions working in certain ways. Some way
              database tables. Some saw the relationships between objects. Some
              saw Excel. Some saw SAP. Regardless of the relative visualization
              skills of each reader, there is a high probability that none will
              be visualizing the same thing.

              Conclusion: This team created and documented a design, but had no
              designer working to visualize that design--to turn the abstract
              concrete. Futhermore, they didn't think they needed one! It
              wasn't until they saw the project starting to head south that
              they realized they had to augment their team.

              So how does this tie back to Jeff's original point?

              The analysis team did excellent work, but they were missing an
              entire "dimension" of specification. Without a visual channel,
              the specs were incomplete, and thus unworkable. I couldn't have
              done my work without the analysis in place, but if I had come on
              early, my needs would have changed the structure of the analysis.
              We probably would have covered the same material--and perhaps
              made most of the same decisions, but we would have organized the
              work and the work product very differently, and we would have
              saved a lot of wasted effort. I think the reverse is true as
              well--if I had been working on the problem for eight months
              without "traditional analysis", I would have missed a dimension
              of specification, and would have had to go back, restructure, and
              re-spec. (But I never would have gotten 8 months down the road
              first ;-)

              JS









              Yahoo! Groups Links









              Sent wirelessly via BlackBerry from T-Mobile.



              Yahoo! Groups Links









              Sent wirelessly via BlackBerry from T-Mobile.
            • Chris
              Agreed. I use risk with priorities to guide efforts. I see change and the impact of change as risks. Putting together a very detailed artifact is a risk at
              Message 6 of 26 , Feb 3, 2005
                Agreed.

                I use risk with priorities to guide efforts. I see change and the impact of change as risks.

                Putting together a very detailed artifact is a risk at different levels. Political, business, project survival, etc.

                Reducing the risks sets up the environment for success.

                -----Original Message-----
                From: "Desilets, Alain" <alain.desilets@...>
                Date: Thu, 3 Feb 2005 13:28:14
                To:"'agile-usability@yahoogroups.com'" <agile-usability@yahoogroups.com>
                Subject: RE: [agile-usability] user centered design overlaps with traditio
                nal analysis?

                When it comes to specs I believe width is more important to depth. Makes sure the spec is end to end, but it's OK if you can't go into detail on everything in between right away.

                -- Alain Désilets
                I am confused by this statement... it seems to contradict itself.

                To me, doing things "end-to-end" means depth-first, not breadth first. More precisely, it means you focus on one particular user task that the system needs to support, and implement it without worrying about the rest (or at least, not until the rest is proven to be actually needed). In ExtremeProgramming, this has been captured by quaint phrases like YAGNI ("You Ain't Gonna Need It") and "Implement the Simplest Thing that Could Possibly Work".

                I have found that to work very well. You design a small part of the system, implement it, deploy it, gather feedback about it, then refactor it, or augment it with new functionality. Each of these cycles takes about one to two weeks. Note that refactoring includes changing the UI so that it is easier to use given the current set of features supported by the system.

                In my opinion, this approach is the only way to gather information about ACTUAL needs, especially when you are designing a new and innovative product where no-one (including users, managers and yes, usability experts) really understand what is required.

                In my opinion, any process where more than 10% of the project time is spent upfront on gathering requirements and producing design documents, does not qualify as Agile.


                Alain Désilets, MASc
                Agent de recherches/Research Officer
                Institut de technologie de l'information du CNRC /
                NRC Institute for Information Technology

                alain.desilets@...
                Tél/Tel (613) 990-2813
                Facsimile/télécopieur: (613) 952-7151

                Conseil national de recherches Canada, M50, 1200 chemin Montréal,
                Ottawa (Ontario) K1A 0R6
                National Research Council Canada, M50, 1200 Montreal Rd., Ottawa, ON
                K1A 0R6

                Gouvernement du Canada | Government of Canada



                Yahoo! Groups Links
                To visit your group on the web, go to:
                http://groups.yahoo.com/group/agile-usability/
                To unsubscribe from this group, send an email to:
                agile-usability-unsubscribe@yahoogroups.com
                Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
                Sent wirelessly via BlackBerry from T-Mobile.
              • Chris
                Forgot to mention that this way of thinking is similar to some eastern philosophy. Can t recall which one. ... From: Chris Date: Thu, 3
                Message 7 of 26 , Feb 3, 2005
                  Forgot to mention that this way of thinking is similar to some eastern philosophy. Can't recall which one.

                  -----Original Message-----
                  From: "Chris" <chris@...>
                  Date: Thu, 3 Feb 2005 18:51:18
                  To:agile-usability@yahoogroups.com
                  Cc:"Pehura, Urgent" <urgent@...>
                  Subject: Re: [agile-usability] user centered design overlaps with traditional analysis?

                  Agreed.

                  I use risk with priorities to guide efforts. I see change and the impact of change as risks.

                  Putting together a very detailed artifact is a risk at different levels. Political, business, project survival, etc.

                  Reducing the risks sets up the environment for success.

                  -----Original Message-----
                  From: "Desilets, Alain" <alain.desilets@...>
                  Date: Thu, 3 Feb 2005 13:28:14
                  To:"'agile-usability@yahoogroups.com'" <agile-usability@yahoogroups.com>
                  Subject: RE: [agile-usability] user centered design overlaps with traditio
                  nal analysis?

                  When it comes to specs I believe width is more important to depth. Makes sure the spec is end to end, but it's OK if you can't go into detail on everything in between right away.

                  -- Alain Désilets
                  I am confused by this statement... it seems to contradict itself.

                  To me, doing things "end-to-end" means depth-first, not breadth first. More precisely, it means you focus on one particular user task that the system needs to support, and implement it without worrying about the rest (or at least, not until the rest is proven to be actually needed). In ExtremeProgramming, this has been captured by quaint phrases like YAGNI ("You Ain't Gonna Need It") and "Implement the Simplest Thing that Could Possibly Work".

                  I have found that to work very well. You design a small part of the system, implement it, deploy it, gather feedback about it, then refactor it, or augment it with new functionality. Each of these cycles takes about one to two weeks. Note that refactoring includes changing the UI so that it is easier to use given the current set of features supported by the system.

                  In my opinion, this approach is the only way to gather information about ACTUAL needs, especially when you are designing a new and innovative product where no-one (including users, managers and yes, usability experts) really understand what is required.

                  In my opinion, any process where more than 10% of the project time is spent upfront on gathering requirements and producing design documents, does not qualify as Agile.


                  Alain Désilets, MASc
                  Agent de recherches/Research Officer
                  Institut de technologie de l'information du CNRC /
                  NRC Institute for Information Technology

                  alain.desilets@...
                  Tél/Tel (613) 990-2813
                  Facsimile/télécopieur: (613) 952-7151

                  Conseil national de recherches Canada, M50, 1200 chemin Montréal,
                  Ottawa (Ontario) K1A 0R6
                  National Research Council Canada, M50, 1200 Montreal Rd., Ottawa, ON
                  K1A 0R6

                  Gouvernement du Canada | Government of Canada



                  Yahoo! Groups Links
                  To visit your group on the web, go to:
                  http://groups.yahoo.com/group/agile-usability/
                  To unsubscribe from this group, send an email to:
                  agile-usability-unsubscribe@yahoogroups.com
                  Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
                  Sent wirelessly via BlackBerry from T-Mobile.


                  Yahoo! Groups Links





                  Sent wirelessly via BlackBerry from T-Mobile.
                • Jade Ohlhauser
                  What I meant by wide over deep is that it covers the end to end experience , or say path , of the feature without having to go deep on each feature. In other
                  Message 8 of 26 , Feb 3, 2005
                    Message
                    What I meant by wide over deep is that it covers the end to end "experience", or say "path", of the feature without having to go deep on each feature. In other words, make sure everything that needs to be talked about is there, without necessarily having each point fleshed out. Kind of like an outline.
                     
                    Other people have said this better than me and I think this point was from Cooper's Asylum, but I may be wrong (too many good books out there).
                     
                    Jade Ohlhauser
                    Product Manager
                    RPM Software                          
                    www.rpmsoftware.com 403-265-6727
                     


                    From: Desilets, Alain [mailto:alain.desilets@...]
                    Sent: Thursday, February 03, 2005 11:28 AM
                    To: 'agile-usability@yahoogroups.com'
                    Subject: RE: [agile-usability] user centered design overlaps with traditional analysis?

                    When it comes to specs I believe width is more important to depth. Makes sure the spec is end to end, but it's OK if you can't go into detail on everything in between right away.  
                     
                    -- Alain Désilets
                    I am confused by this statement... it seems to contradict itself.
                     
                    To me, doing things "end-to-end" means depth-first, not breadth first. More precisely, it means you focus on one particular user task that the system needs to support, and implement it without worrying about the rest (or at least, not until the rest is proven to be actually needed). In ExtremeProgramming, this has been captured by quaint phrases like YAGNI ("You Ain't Gonna Need It") and "Implement the Simplest Thing that Could Possibly Work".
                     
                    I have found that to work very well. You design a small part of the system, implement it, deploy it, gather feedback about it, then refactor it, or augment it with new functionality. Each of these cycles takes about one to two weeks. Note that refactoring includes changing the UI so that it is easier to  use given the current set of features supported by the system.
                     
                    In my opinion, this approach is the only way to gather information about ACTUAL needs, especially when you are designing a new and innovative product where no-one (including users, managers and yes, usability experts) really understand what is required.
                     
                    In my opinion, any process where more than 10% of the project time is spent upfront on gathering requirements and producing design documents, does not qualify as Agile.
                     

                    Alain Désilets, MASc
                    Agent de recherches/Research Officer
                    Institut de technologie de l'information du CNRC /
                    NRC Institute for Information Technology

                    alain.desilets@...
                    Tél/Tel (613) 990-2813
                    Facsimile/télécopieur: (613) 952-7151

                    Conseil national de recherches Canada, M50, 1200 chemin Montréal,
                    Ottawa (Ontario) K1A 0R6
                    National Research Council Canada, M50, 1200 Montreal Rd., Ottawa, ON
                    K1A 0R6

                    Gouvernement du Canada | Government of Canada

                     

                  • Jeff Patton
                    I m just catching up with this list [and the rest of my life for that matter...] ... These are the failure conditions I hear Alistair Cockburn harp about
                    Message 9 of 26 , Feb 10, 2005
                      I'm just catching up with this list [and the rest of my life for that
                      matter...]

                      --- In agile-usability@yahoogroups.com, Josh Seiden <joshseiden@y...>
                      wrote:
                      > This team has been very detail oriented, so the specs
                      > represent a documentation of every possible thing that
                      > could ever happen. (What if a fire alarm sounds while
                      > the user is eating a low-fat tuna sandwich while
                      > entering this data? Is that different from the case in
                      > which the tuna sandwich is made with regular tuna
                      > salad?")

                      These are the failure conditions I hear Alistair Cockburn harp about
                      frequently... and he's winning me over. Doing design work often has
                      me focusing on the most probable "happy path" and paying a bit less
                      attention to failure conditions. Alistair asserts that use cases
                      force consideration of those failure conditions sooner. And,
                      supporting what he's saying I've seen lots of nasty stuff that causes
                      delays in development bubble out of the failure conditions. But,
                      often that nasty stuff only slightly affects UI design or user
                      interactions.

                      Is it safe to assume that fire-alarm-with-low-fat-tuna-salad was a
                      failure condition on a use case? Did you have use cases to work
                      from, and did you find that the usecases documented basic user
                      interaction reasonably - in a way useful for your design work?

                      Thanks,

                      -Jeff
                    • Jeff Patton
                      ... are sent ... folks ... revision, and ... We ve been bad at subscribing to changes - we ve had the silly belief that since they re sitting a few feet from
                      Message 10 of 26 , Feb 10, 2005
                        --- In agile-usability@yahoogroups.com, "Sue Heim" <sue_heim@m...>
                        wrote:
                        > Can't your subscribe to the Wiki, so that changes to any document
                        are sent
                        > out to the appropriate parties? This would ensure, anyways, that
                        folks
                        > who've already read the doc will be notified that there is a
                        revision, and
                        > they oh yeah, maybe they oughta go check it out...?
                        >

                        We've been bad at subscribing to changes - we've had the silly belief
                        that since they're sitting a few feet from us and they know that
                        we're working on their story they might /tell/ us. But, I'm being
                        overly harsh - our super users are good people and usually do tell
                        us.

                        The bigger problems have been that the documents they write contain
                        lots of words... lots of words that they've written independently
                        without a lot of support from others. Those words contain design
                        decisions. Programmers read the words and say to themselves, "these
                        are my requirements, I'll write the software this way." There's
                        often not much conversation about those words. But, often we find
                        out the words result in a piece of software that's clumsy to use - or
                        has contradictory information in it. In those cases we have a
                        conversation with the person who wrote the document and find that
                        what they'd written isn't really what they meant, and they're quick
                        to accept any suggestions we have for improvement.

                        At our recent reflection session we decided it would be best if the
                        documents were delievered along with a face to face conversation. We
                        continue to be reminded that documents give us a false sense of
                        security - that words in print seem "right" even though they're often
                        not.

                        thanks,

                        -Jeff
                      • Chris
                        I lived though similar situations. And the politics always ended badly. ... From: Jeff Patton Date: Thu, 10 Feb 2005 15:04:28
                        Message 11 of 26 , Feb 10, 2005
                          I lived though similar situations. And the politics always ended badly.

                          -----Original Message-----
                          From: "Jeff Patton" <jpatton@...>
                          Date: Thu, 10 Feb 2005 15:04:28
                          To:agile-usability@yahoogroups.com
                          Subject: [agile-usability] Re: user centered design overlaps with traditional analysis?



                          --- In agile-usability@yahoogroups.com, "Sue Heim" <sue_heim@m...>
                          wrote:
                          > Can't your subscribe to the Wiki, so that changes to any document
                          are sent
                          > out to the appropriate parties? This would ensure, anyways, that
                          folks
                          > who've already read the doc will be notified that there is a
                          revision, and
                          > they oh yeah, maybe they oughta go check it out...?
                          >

                          We've been bad at subscribing to changes - we've had the silly belief
                          that since they're sitting a few feet from us and they know that
                          we're working on their story they might /tell/ us. But, I'm being
                          overly harsh - our super users are good people and usually do tell
                          us.

                          The bigger problems have been that the documents they write contain
                          lots of words... lots of words that they've written independently
                          without a lot of support from others. Those words contain design
                          decisions. Programmers read the words and say to themselves, "these
                          are my requirements, I'll write the software this way." There's
                          often not much conversation about those words. But, often we find
                          out the words result in a piece of software that's clumsy to use - or
                          has contradictory information in it. In those cases we have a
                          conversation with the person who wrote the document and find that
                          what they'd written isn't really what they meant, and they're quick
                          to accept any suggestions we have for improvement.

                          At our recent reflection session we decided it would be best if the
                          documents were delievered along with a face to face conversation. We
                          continue to be reminded that documents give us a false sense of
                          security - that words in print seem "right" even though they're often
                          not.

                          thanks,

                          -Jeff







                          Yahoo! Groups Links









                          Sent wirelessly via BlackBerry from T-Mobile.
                        • Chris
                          It s funny that you mention those failure conditions. There is this product that automatically handes those failures. They call these failures exceptions. For
                          Message 12 of 26 , Feb 10, 2005
                            It's funny that you mention those failure conditions. There is this product that automatically handes those failures. They call these failures exceptions. For design, you just have to worry about the happy paths. So far I'm really impressed by its simplicity.

                            -----Original Message-----
                            From: "Jeff Patton" <jpatton@...>
                            Date: Thu, 10 Feb 2005 14:53:48
                            To:agile-usability@yahoogroups.com
                            Subject: [agile-usability] Re: user centered design overlaps with traditional analysis?



                            I'm just catching up with this list [and the rest of my life for that
                            matter...]

                            --- In agile-usability@yahoogroups.com, Josh Seiden <joshseiden@y...>
                            wrote:
                            > This team has been very detail oriented, so the specs
                            > represent a documentation of every possible thing that
                            > could ever happen. (What if a fire alarm sounds while
                            > the user is eating a low-fat tuna sandwich while
                            > entering this data? Is that different from the case in
                            > which the tuna sandwich is made with regular tuna
                            > salad?")

                            These are the failure conditions I hear Alistair Cockburn harp about
                            frequently... and he's winning me over. Doing design work often has
                            me focusing on the most probable "happy path" and paying a bit less
                            attention to failure conditions. Alistair asserts that use cases
                            force consideration of those failure conditions sooner. And,
                            supporting what he's saying I've seen lots of nasty stuff that causes
                            delays in development bubble out of the failure conditions. But,
                            often that nasty stuff only slightly affects UI design or user
                            interactions.

                            Is it safe to assume that fire-alarm-with-low-fat-tuna-salad was a
                            failure condition on a use case? Did you have use cases to work
                            from, and did you find that the usecases documented basic user
                            interaction reasonably - in a way useful for your design work?

                            Thanks,

                            -Jeff







                            Yahoo! Groups Links









                            Sent wirelessly via BlackBerry from T-Mobile.
                          • Cummins, Darin
                            We ve only recently started writing use-cases, so take this for what it is worth. When we sit down to create the use-cases (usually with product management, QA
                            Message 13 of 26 , Feb 10, 2005

                              We’ve only recently started writing use-cases, so take this for what it is worth.

                               

                              When we sit down to create the use-cases (usually with product management, QA and someone from customer service all taking the role of “user”), it’s been interesting to see how the UI has changed from what the developer thought it should be, to something that really works for the user.  By discussing the failure conditions and how the UI is going to respond, the developers have really been re-thinking the entire front end, especially how it interacts with the backend.

                               

                              We had one project recently where a developer created the UI based on the simple requirements delivered, just the way we use to do it.  Everyone, including people from outside of development, really liked what he had done.  He was then pulled off on another project while we got together and created use-cases.  When we revisited the UI, it became blatantly obvious that it would not work and we redesigned it.  (What a concept!)

                               

                              Anyway, in that case, the failure conditions and how it was decided to be handled in the UI, actually simplified the way the backend needed to work.

                               

                              --Darin

                               

                               

                               

                               

                               


                              From: Jeff Patton [mailto:jpatton@...]
                              Sent: Thursday, February 10, 2005 7:54 AM
                              To: agile-usability@yahoogroups.com
                              Subject: [agile-usability] Re: user centered design overlaps with traditional analysis?

                               


                              I'm just catching up with this list [and the rest of my life for that
                              matter...]

                              --- In agile-usability@yahoogroups.com, Josh Seiden <joshseiden@y...>
                              wrote:
                              > This team has been very detail oriented, so the specs
                              > represent a documentation of every possible thing that
                              > could ever happen. (What if a fire alarm sounds while
                              > the user is eating a low-fat tuna sandwich while
                              > entering this data? Is that different from the case in
                              > which the tuna sandwich is made with regular tuna
                              > salad?")

                              These are the failure conditions I hear Alistair Cockburn harp about
                              frequently... and he's winning me over.  Doing design work often has
                              me focusing on the most probable "happy path" and paying a bit less
                              attention to failure conditions.  Alistair asserts that use cases
                              force consideration of those failure conditions sooner.  And,
                              supporting what he's saying I've seen lots of nasty stuff that causes
                              delays in development bubble out of the failure conditions.  But,
                              often that nasty stuff only slightly affects UI design or user
                              interactions.

                              Is it safe to assume that fire-alarm-with-low-fat-tuna-salad was a
                              failure condition on a use case?  Did you have use cases to work
                              from, and did you find that the usecases documented basic user
                              interaction reasonably - in a way useful for your design work?

                              Thanks,

                              -Jeff






                            • Jamie Nettles
                              Absolutely. I use software to accomplish tasks. I want the software to work. I want to get the job done efficiently and move on. Makes me want to reread
                              Message 14 of 26 , Feb 24, 2005
                                Absolutely. I use software to accomplish tasks. I want the software to work. I want to get the job done efficiently and move on. Makes me want to reread "Software for Use".

                                > -----Original Message-----
                                > From: Larry Constantine [mailto:lconstantine@...]
                                > Sent: Thursday, February 24, 2005 11:00 AM
                                > To: agile-usability@yahoogroups.com
                                > Subject: RE: [agile-usability] Naming the thing (was "user
                                > centered design overlaps with traditional analysis?)
                                >
                                >
                                > Josh et al.,
                                >
                                > Ux (user experience) design does seem to be winning the
                                > branding race, in part because it is so wonderfully fuzzy and
                                > capable of covering almost anything and everything--the
                                > entire "experience" of the user. Aside from the philosophical
                                > and semantic issue of whether anyone can ever design
                                > another's experience (and this from someone who wrote a
                                > classic article on psychotherapy training titled "Designed
                                > Experience"), it is precisely the enormous potential purview
                                > of UxD that is worrisome.
                                >
                                > Ultimately what gives users a good experience is their
                                > ability to do what they need to do efficiently and
                                > dependably. Supporting effective user performance is 90% of
                                > the story, which is why we make that the centerpiece of
                                > USAGE-centered design. The problem with so much of
                                > user-experience design and user-centered approaches in
                                > practice, is that they spend so much time and attention on
                                > matters that may be fun and interesting and help make users
                                > feel appreciated and attended to while promoting full
                                > employment for Ux designers but that have little if anything
                                > to do with the really important stuff in visual and
                                > interaction design.
                                >
                                > (Would you believe, for example, a forthcoming book on
                                > personas argues that clipart headshots or stock photos are
                                > not "real" enough and should not be used to illustrate
                                > personas. Instead, each team should go out and do its own
                                > photo shoots of real people based on careful research on how
                                > best to embody each persona. Somebody, methinks, has too much
                                > time on their hands.)
                                >
                                > We use small, simple, scaled-down abstract models to focus on
                                > the bare essentials with the greatest impact because we want
                                > to use our limited time for designing a user interface that
                                > really works for users, rather than for shooting and editing
                                > compelling videotapes, compiling and categorizing great heaps
                                > of stick-notes, or concocting believable biographies for
                                > fictitious users whose profiles can be turned into
                                > large-format color posters to grace the walls of the
                                > development facility.
                                >
                                > Okay, I know I am a minority voice here (and when or where
                                > has it been otherwise?), but I have seen too many loftily
                                > user-centered projects miss the boat completely because they
                                > missed the central point of good design. It is not about
                                > users, it is about uses (or usage, if you prefer).
                                >
                                > --Larry Constantine, IDSA [mailto:lconstantine@...]
                                >   Chief Scientist | Constantine & Lockwood, Ltd. | www.foruse.com
                                >
                                > > -----Original Message-----
                                > > From: Josh Seiden [mailto:joshseiden@...]
                                > > Sent: Tuesday, 01 February 2005 6:10 PM
                                > > To: agile-usability@yahoogroups.com
                                > > Subject: [agile-usability] Naming the thing (was "user
                                > centered design
                                > overlaps with
                                > > traditional analysis?)
                                > >
                                > >
                                > > Since this has come up a lot recently, I thought I'd chime in.
                                > >
                                > > There is a growing movement to call the larger practice of
                                > > user-centered-design by the name "User Experience Design",
                                > "Experience
                                > > Design" or simply, "UX", and to consider the word
                                > "usability" in it's
                                > > narrow meaning.
                                > >
                                > > This naming convention is based on a couple of ideas:
                                > >
                                > > 1. That as products and services become more complex, they require
                                > > people who practice a variety of disciplines to get the
                                > > product/service built.
                                > >
                                > > An e-commerce site might need interaction design, information
                                > > architecture, usability evaluation graphic design, service design,
                                > > corporate identity design, etc.
                                > >
                                > > 2. Some of these people have titles that identify the single
                                > > discipline they practice (graphic design) and other do not.
                                > (As noted,
                                > > "usability people" might practice a variety of disciplines.)
                                > >
                                > > 3. That there is value in identifying the common thread and
                                > methods,
                                > > and value in developing the various disciplines with greater rigor.
                                > >
                                > > Thus, we're starting to see some umbrella organizations emerge to
                                > > represent the larger practice of User Experience Design. These
                                > > organizations include AIGA-ED and UXnet.net.
                                > >
                                > > We're also starting to see some new discipline-focused
                                > organizations
                                > > like AIFIA and IxDG that advocate for more narrow
                                > definitions of their
                                > > respective disciplines.
                                > >
                                > > And of course, as with all naming exercises, this one had plenty of
                                > > politics and controversy too. ;-)
                                > >
                                > > JS
                                > >
                                > >
                                > >
                                > > --- Kay Burnett <kayburnett2002@...> wrote:
                                > >
                                > > > I have
                                > > > thought for some time that the title or topic of
                                > "usability" needs
                                > > > to either be defined narrowly or be expanded
                                > >
                                > >
                                > >
                                > > Yahoo! Groups Links
                                > >
                                > >
                                > >
                                > >
                                > >
                                > >
                                >
                                >
                                >
                                >
                                >
                                >
                                > Yahoo! Groups Links
                                >
                                >
                                >
                                >
                                >
                                >
                                >
                                >
                              • Larry Constantine
                                Jamie wrote: Absolutely. I use software to accomplish tasks. I want the software to work. I want to get the job done efficiently and move on. Makes me
                                Message 15 of 26 , Feb 25, 2005
                                  Jamie wrote:

                                  "Absolutely. I use software to accomplish tasks. I want the software to
                                  work. I want to get the job done efficiently and move on. Makes me want to
                                  reread "Software for Use"."

                                  Go for it! :-) I'll even autograph your copy.

                                  I am increasingly convinced that mainstream user-centered design has drifted
                                  off target, although I get brickbats every time I dare to suggest this in
                                  public. I did stick my neck out recently in the Cutter IT Journal in a
                                  provocative article called "Beyond User-Centered Design"
                                  (http://www.cutter.com/offers/userdesign.html).

                                  --Larry Constantine, IDSA
                                  Chief Scientist | Constantine & Lockwood Ltd | www.foruse.com
                                Your message has been successfully submitted and would be delivered to recipients shortly.