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

Re: [agile-usability] Re: BAs

Expand Messages
  • Robert Biddle
    ... Actually I do not think this is quite true. In UI work we often get *better* formative feedback with a low-fi prototype. Moreover, even shipping does not
    Message 1 of 248 , May 26, 2008
    • 0 Attachment
      So we seem to have two related disagreements here:

      >But testing with a paper prototype or a wizard-of-oz interface still
      >only tests a small fraction of the product, and only partially. They
      >increase you chance of success in the real world, but do not guarantee
      >it. No matter how much testing you do, you can always learn more by
      >shipping.

      Actually I do not think this is quite true. In UI work we often get
      *better* formative feedback with a low-fi prototype.
      Moreover, even shipping does not guarantee success in the real world.

      >...
      > Yes. I'd call that part of shipping. If you don't, that's fine. I mean
      > the whole ball of wax. I'm interested in producing whole products, not code.

      This is not an of terminology. My point is that there is that the value chain does
      not have clear boundaries. There is no clear "whole ball of wax". So, for example,
      if you produce a whole product, there is the issue of what people do with the product,
      and how that affects things downstream. And that's a product; I think it's the same or moreso
      for services. But let me make it clear that I think it's fine for you to focus on a product:
      it's just not the only choice, nor is it the whole ball of wax.

      ---

      So, have I described the nature of our disagreements reasonably?

      Robt




      On 5/25/08, William Pietri <william@...> wrote:

      Robert Biddle wrote:
      > > Some of the hypotheses in a design are about the working medium. You
      > > test those as you build. But the hypotheses about what will happen in
      > > the real world -- the most important ones -- are only tested when you
      > > ship.
      >
      > I'm still not sure we're together on this.
      > With UI designs, you can do impressive testing using the "Wizard-of-Oz"
      > technique, with paper prototypes with real users in their real workplaces.
      > So that is not just testing the medium.
      >

      Yes. Sorry I was unclear. One tests some hypotheses, ones about the
      working medium, through building. One tests other hypotheses differently.

      But testing with a paper prototype or a wizard-of-oz interface still
      only tests a small fraction of the product, and only partially. They
      increase you chance of success in the real world, but do not guarantee
      it. No matter how much testing you do, you can always learn more by
      shipping.

      > And as for when you ship, well, that isn't the whole story either: all
      > kind of
      > deployment and support factors also play a role. There are many kinds of
      > agiliy.
      > Those that produce code are not alone.
      >

      > Yes. I'd call that part of shipping. If you don't, that's fine. I mean
      > the whole ball of wax. I'm interested in producing whole products, not code.




       

      My point is that in creating products for people to use, we should not
      mistake our hypotheses about what happens for what actually happens.
      Better hypotheses are still hypotheses. Partial tests cannot validate
      the whole design. The ultimate test is turning things loose in the real
      world.

      If we can test more cheaply or more safely than releasing into the wild,
      I am in favor of that. I'm even more in favor of making releases cheap
      and safe enough that people do it all the time.

      William


    • Desilets, Alain
      ... in ... That s such a great idea! Every developer needs a license to hack every once in a while. It s good for mental hygiene ;-). Alain
      Message 248 of 248 , Jun 4, 2008
      • 0 Attachment
        > Another is to encourage people to have wild ideas. One client has a
        > monthly hack day, where people are encouraged to try out crazy stuff.
        > Another has beers once a week, with bold notions jotted down and put
        in
        > the backlog for later examination.

        That's such a great idea! Every developer needs a license to hack every
        once in a while. It's good for mental hygiene ;-).

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