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

Estimating Without Jargon

Expand Messages
  • Kevin Lawrence
    (hypothetically) I work for a large Wall Street bank. I am the project manager of a software team in the Global Derivatives Credit Management department. My
    Message 1 of 5 , Nov 1, 2002
    • 0 Attachment
      (hypothetically)

      I work for a large Wall Street bank. I am the project manager of a software
      team in the Global Derivatives Credit Management department. My boss (a
      technical person) reports to the head of Global Derivatives (a business
      person). They (my boss and the big boss) call me into the office and tell me
      that they want a system to provide real time credit information to 500
      hundred traders in 17 offices worldwide. They also want to provide daily
      reports to the credit managers and notify them in real time if any trader
      exceeds a credit limit. They want to know how long it will take and how much
      it will cost.

      I tell them (based on a similar project that we just finished) that it will
      take about a year (give or take a year) and will cost about $3m (give or
      take $3m).

      "Great !" they say. "That's well within our budget. How long will it take to
      get a better estimate ?"

      I ask if I can get access to any of the users. They tell me that Maria, one
      of the business analysts has been assigned to the project. I say well that's
      great but we really need access to real users. The Big Boss says I can get
      as much time as I need from the assistant credit manager for Latin America
      and I can have free access to the trading floor. Derivatives trading is a
      fairly leisurely business (relative to other kinds of trading), so I should
      be able to corner some of the senior traders for some quality time.

      I tell them I can get a rough estimate in about two weeks and I can get
      fairly accurate estimates two weeks later.

      "Go for it." They tell me.

      For the next two weeks I go down to trading floor with Maria and 3 of my
      programmers and we talk to the people that would be using the system.
      Everytime time someone describes something that the system should do, I
      write it on an index card. After a week we have a stack of 93 cards that say
      things like

      - Calculate the settlement risk for 30 year interest-rate swaps.

      - Listen in on ReuterNet conversations with interbank customers.
      - Beep if a proposed deal exceeds the credit limit.
      - Send a daily report of deals to the credit manager of each region.
      - Highlight trades that exceed $150m
      - Send a summary to the mainframe

      At the end of the first week, I sit down with Maria and my programmers and
      we start estimating. We get through most of them in a couple of days. Some
      are too big and fuzzy to so we go back down to the trading floor and get
      some more information. We manage to split them up into bite-size pieces and
      write them on new cards.

      We are not sure what to do about that ReuterNet thingie so I call the guy
      who installed that system. He says its just TCP with a really simple
      protocol. He sends me the manual and a test version. One of my programmers
      plays with it for a few hours and announces it's pretty straightforward. We
      go back to the traders and ask exactly what it is they want and write a few
      more cards.

      When we get through all the cards we call a meeting with the senior trader
      and the credit manager and we ask them to order the cards by how important
      they are. After a little horse trading they manage to agree on the priority.
      By the time they get to the low priority cards, they tear a few up ("what
      were we thinking?") and add a few more too.

      After two weeks I am able to go back to my boss (and his boss) and tell them
      that, with our current resources, we can be done in three and a half years.
      He almost chokes on his coffee.

      "We need something by July ! What are they asking for ?"

      I show him a printout of our release plan and say "We might be able to get
      the first 50 stories done by July. We'll know for sure in two weeks."

      "Excellent !", they exclaim in chorus. "When can we start coding ?"

      "Monday", I reply.

      The next day, my boss calls me and says that he was talking to his buddy in
      Commodities. His buddy says that there is this manager in his department who
      says that he thinks that two weeks is a long time to build a plan for a
      project like this and that we are reckless to start such an important
      project without more accurate estimates. I confess to my boss that I don't
      know of a faster way to get better estimates.

      He calls a meeting - he got hauled over the coals because his last project
      was late by a year and was cancelled the week it went live because it didn't
      do what the users wanted. He wants to make sure this time.

      The meeting comes around and I ask,

      "Bill, how do you get more accurate estimates in less than two weeks ?"




      Kevin Lawrence
      Diamond Sky Software
      www.diamond-sky.com
    • Laurent Bossavit
      Applause, please. Wonderful piece of work.
      Message 2 of 5 , Nov 1, 2002
      • 0 Attachment
        Applause, please. Wonderful piece of work.
      • Luke Burton
        ... By applying XP as a methodology on your own project. There is no absolute truth to be divined in XP s practises. There is no correct , only most
        Message 3 of 5 , Nov 1, 2002
        • 0 Attachment
          Bill Walton wrote:

          > I'm currently working on a
          > hypothesis that say's XP's assumptions are correct. How can we
          > demonstrate that?

          By applying XP as a methodology on your own project. There is no
          absolute truth to be divined in XP's practises. There is no 'correct',
          only 'most appropriate given the circumstances'. Think relativity and
          absolute frames of reference if that helps.

          > What do you think the players' underlying
          > assumptions are and how would you address those assumptions rather
          > than the words that come out as a result of them.

          Can you answer the original question posed? Nobody has yet.

          Regards,

          Luke.
        • Andrey Khavryuchenko
          Luke, LB == Luke Burton wrote: LB By applying XP as a methodology on your own project. There is no LB absolute truth to be divined in XP s practises. There
          Message 4 of 5 , Nov 2, 2002
          • 0 Attachment
            Luke,

            "LB" == Luke Burton wrote:

            LB> By applying XP as a methodology on your own project. There is no
            LB> absolute truth to be divined in XP's practises. There is no 'correct',
            LB> only 'most appropriate given the circumstances'.

            I doubt that there's even 'most' appropriate practice. I'd bet that there
            is only 'more' or 'less' appropriate ones given the circumstances.

            --
            Andrey V Khavryuchenko http://www.kds.com.ua/
            Software Solutions
          • Dossy
            ... [... snip ...] *Applause*. The only thing I have to say is I wish this were a true story ... only so I could use it as-is as testimonial. -- Dossy -- Dossy
            Message 5 of 5 , Nov 3, 2002
            • 0 Attachment
              On 2002.11.01, Kevin Lawrence <kevin@...> wrote:
              > (hypothetically)
              [... snip ...]

              *Applause*.

              The only thing I have to say is I wish this were a true story ...
              only so I could use it as-is as testimonial.

              -- Dossy

              --
              Dossy Shiobara mail: dossy@...
              Panoptic Computer Network web: http://www.panoptic.com/
              "He realized the fastest way to change is to laugh at your own
              folly -- then you can let go and quickly move on." (p. 70)
            Your message has been successfully submitted and would be delivered to recipients shortly.