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

Re: [XP] If coach is a programmer, where do I fit in?

Expand Messages
  • Steve Freeman
    ... From: To: Sent: Monday, May 01, 2000 4:39 PM ... to repeat the Party Line... what s your
    Message 1 of 6 , May 1, 2000
      ----- Original Message -----
      From: <jkinghorn@...>
      To: <extremeprogramming@egroups.com>
      Sent: Monday, May 01, 2000 4:39 PM
      > side so I can better prove the worth of XP with examples? How should
      > I introduce XP? Start with Refactoring and Unit Tests and go from
      > there?

      to repeat the Party Line... what's your biggest problem right now?

      Steve
    • jkinghorn@installinc.com
      ... Our biggest problems right now are: managing requirements (internal customer keeps changing them, but wants them all in first release) lack of sufficient
      Message 2 of 6 , May 1, 2000
        >
        > to repeat the Party Line... what's your biggest problem right now?
        >
        > Steve

        Our biggest problems right now are:
        managing requirements (internal customer keeps changing them, but
        wants them all in first release)
        lack of sufficient testing
        very little project visibility (no progress measurements)

        All of these problems mean we are consistently counting on the
        heroics of the team.

        So, I guess my answer is to start with User Stories and Unit Tests?
      • Randy Coulman
        ... For the first and third problems, I think you actually want the Planning Game, for which you will need the User Stories as game tokens . User Stories
        Message 3 of 6 , May 1, 2000
          jkinghorn@... wrote:
          >
          > Our biggest problems right now are:
          > managing requirements (internal customer keeps changing them, but
          > wants them all in first release)
          > lack of sufficient testing
          > very little project visibility (no progress measurements)
          >
          > All of these problems mean we are consistently counting on the
          > heroics of the team.
          >
          > So, I guess my answer is to start with User Stories and Unit Tests?
          >

          For the first and third problems, I think you actually want the Planning
          Game, for which you will need the User Stories as "game tokens". User
          Stories will help with "managing requirements", but you need the full
          Planning Game to help the customer understand what it would cost (in
          terms of time, resources, and risk) to have all of the requirements in
          the first release.

          For the second problem, Unit Tests is definitely the answer.

          Randy
          --
          Randy Coulman, M.Sc. Settler Agricultural Software
          RandyC@... 200 - 1911 Park St. Regina, SK S4N 2G5
          Senior Developer Phone: (306) 721-7949 Ext. 224
          Special Projects Fax: (306) 721-1981 http://www.settler.com
        • Donald F. McLean
          ... Not to be contrary, but Unit Tests are only half of the answer - Functional Test are the other. It isn t enough that the program works as designed, it has
          Message 4 of 6 , May 1, 2000
            Randy Coulman wrote:
            >
            > jkinghorn@... wrote:
            > >
            > > Our biggest problems right now are:
            > > managing requirements (internal customer keeps changing them, but
            > > wants them all in first release)
            > > lack of sufficient testing
            > > very little project visibility (no progress measurements)
            > >
            > > All of these problems mean we are consistently counting on the
            > > heroics of the team.
            > >
            > > So, I guess my answer is to start with User Stories and Unit Tests?
            > >
            >
            > For the first and third problems, I think you actually want the Planning
            > Game, for which you will need the User Stories as "game tokens". User
            > Stories will help with "managing requirements", but you need the full
            > Planning Game to help the customer understand what it would cost (in
            > terms of time, resources, and risk) to have all of the requirements in
            > the first release.
            >
            > For the second problem, Unit Tests is definitely the answer.

            Not to be contrary, but Unit Tests are only half of the answer -
            Functional Test are the other. It isn't enough that the program
            works as designed, it has to be designed to do the right thing.
            --
            Donald McLean dmclean@...
            Space Telescope Science Institute http://www.stsci.edu
            AOL Instant Messaging: DonaldFMcLean
            ------------------------------------------------------------
            There are only four questions of value in life, Don Octavio.
            What is sacred? Of what is the spirit made?
            What is worth living for? And what is worth dying for?
            The answer to each is the same: only love.
            ----From "Don Juan DeMarco" (1995)
          • Randy Coulman
            ... You re right, of course. I guess I was thinking more in terms of Fix your worst problems first , which is the context of the thread. In that context, if
            Message 5 of 6 , May 1, 2000
              "Donald F. McLean" wrote:
              >
              > Randy Coulman wrote:
              > >
              > > For the second problem, Unit Tests is definitely the answer.
              >
              > Not to be contrary, but Unit Tests are only half of the answer -
              > Functional Test are the other. It isn't enough that the program
              > works as designed, it has to be designed to do the right thing.

              You're right, of course.

              I guess I was thinking more in terms of "Fix your worst problems first",
              which is the context of the thread.

              In that context, if you're not doing much testing, it's probably easier
              to start with Unit Tests than Functional Tests (at least in my limited
              experience).

              Also, Unit Tests should ensure that "lack of sufficient testing" is no
              longer your worst problem. When "the system doesn't do exactly what the
              customer wants" becomes the biggest problem, that's when to start on
              Functional Tests (at least when incrementally adopting XP by solving
              your worst problem(s) first).

              Randy
              --
              Randy Coulman, M.Sc. Settler Agricultural Software
              RandyC@... 200 - 1911 Park St. Regina, SK S4N 2G5
              Senior Developer Phone: (306) 721-7949 Ext. 224
              Special Projects Fax: (306) 721-1981 http://www.settler.com
            • Steve Freeman
              ... From: To: Sent: Monday, May 01, 2000 9:34 PM Subject: Re: [XP] If coach is a programmer, where
              Message 6 of 6 , May 1, 2000
                ----- Original Message -----
                From: <jkinghorn@...>
                To: <extremeprogramming@egroups.com>
                Sent: Monday, May 01, 2000 9:34 PM
                Subject: Re: [XP] If coach is a programmer, where do I fit in?
                > Our biggest problems right now are:
                > managing requirements (internal customer keeps changing them, but
                > wants them all in first release)
                > lack of sufficient testing
                > very little project visibility (no progress measurements)
                >
                > All of these problems mean we are consistently counting on the
                > heroics of the team.
                >
                > So, I guess my answer is to start with User Stories and Unit Tests?

                That sounds good, if you can get your users to join in a planning game. We had good experiences with this because it exposes the users to some of the gritty detail they usually ignore. It can make them much more carefully about what they really want.

                Depending on your codebase, you might want to start with Functional Tests, so you can do full regression testing and use new tests as a measure of progress.It depends on how hard it would be to set up something automated that talks to the model (i.e. don't go in through one of those UI tools).

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