<html>
  <head>

    <meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    All,<br>
    <br>
    I took minutes during the TB QA meeting at MozCamp.<br>
    <br>
    Minutes can be found here starting on line 18:<br>
    <a href="https://etherpad.mozilla.org/tb-qa"><br>
      https://etherpad.mozilla.org/tb-qa</a><br>
    <br>
    Here's a summary:<br>
    <br>
    <meta http-equiv="content-type" content="text/html;
      charset=ISO-8859-1">
    <meta http-equiv="content-type" content="text/html;
      charset=ISO-8859-1">
    <div class="" id="magicdomid306"><span
        class="author-g-fp2xkcspz122zccid7qs b"><b>Summary</b></span></div>
    <div class="ace-line" id="magicdomid4217">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">The rapid
            release has really changed the way we do testing. Community
            involvement has dropped off.</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid4219">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">Ludo has no
            chance of watching all of Bugzilla, and sometimes somebody
            comes in with a patch, and we don't see it because they
            don't set the right flags, and then it gets bitrotted and we
            lose a potential contributor.</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid4221">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">We need more
            people to be active on Bugzilla, to take action on bugs, to
            respond when somebody files a bug to get the information we
            need to fix it (if you reply within 1 day, you have a higher
            degree of probability of getting useful information for
            fixing it).</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid4223">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">The "bubbling
            up" of bugs from Bugzilla to devs to being fixed is what
            needs work. People are getting frustrated when they file
            bugs and *nothing happens*. Letting people know that when
            they file bugs properly, that things happen...that's
            important.</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid4226">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">The notion of
            what NEW means on Bugzilla is different from person to 
            person. Some people think NEW means, "this bug is
            reproducable". Others  think it means "this bug is
            reproducable, and we know where the fix  needs to happen."</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid4228">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">IDing of bugs,
            of prioritizing them for being fixed - we need to look into
            new and better ways of doing this.</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid4240">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">There are 4
            dimensions for categorizing bugs:</span></li>
      </ul>
    </div>
    <blockquote>
      <div class="ace-line" id="magicdomid4246">
        <ul class="list-bullet2">
          <li><span class="author-g-fp2xkcspz122zccid7qs">How many
              people are likely to be affected</span></li>
        </ul>
      </div>
      <div class="ace-line" id="magicdomid4254">
        <ul class="list-bullet2">
          <li><span class="author-g-fp2xkcspz122zccid7qs">How serious is
              the effect when it occurs (workaround? consequences)</span></li>
        </ul>
      </div>
      <div class="ace-line" id="magicdomid4255">
        <ul class="list-bullet2">
          <li><span class="author-g-fp2xkcspz122zccid7qs">How much
              effort to fix this bug?</span></li>
        </ul>
      </div>
      <div class="ace-line" id="magicdomid4258">
        <ul class="list-bullet2">
          <li><span class="author-g-fp2xkcspz122zccid7qs">How
              reproducible it is</span></li>
        </ul>
      </div>
    </blockquote>
    <div class="ace-line" id="magicdomid4260">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">We need some way
            of visualizing bugs along these dimensions, and then a way
            of consolidating that visualization into a prioritized list.</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid4262">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">But generating
            lists is easy. We've been generating lists for YEARS. The
            problem is that we create a list, fix a few things on it,
            and then create a new list, and repeat.</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid4264">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">Fixing old bugs
            has taken a backseat in favour of feature development over
            the past few releases</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid4266">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">Ludo is
            concerned about performance problems in TB - his TB on his
            beefy machine is just as performant as TB 1.x was on his old
            crappy machine. That's not a good sign.</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid4268">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">Irving advocates
            a light-weight, "agile" list that is designed to be easily
            changed iteration after iteration, based on what is
            happening at the time. The list should make it easy to
            visualize the "life-cycle" of a bug, and allow us to measure
            how long it takes for a bug to go from being identified, to
            being fixed. That lets us know how many things to put on the
            list, and we can adjust accordingly.</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid4270">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">We need a way
            for developers to examine a bug, and deprioritize it if it
            is way too complicated - ie, the rewards are too slim based
            on how much work is likely involved. (Spending time thinking
            about the hard / impossible bugs is less valuable than
            fixing the bugs that we're likely to tackle and complete)</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid4272">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">We're going to
            tackle the Papercuts list, and see what happens. If we can
            show that we can fix the items on a list, and move the ball
            forward, then we'll worry about creating the list.</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid4274">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">QA people think,
            "We make lists and you devs ignore them!", devs think, "You
            QA people make lists, and we can't do anything with them!".
            We need to resolve this.</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid4276">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">The list should
            be mutable. We need to be able to change our minds. We don't
            want to thrash, but we want to be able to make adjustments.</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid4278">
      <ul class="list-bullet1">
        <li><span class="author-g-fp2xkcspz122zccid7qs">We need criteria
            to create the initial list.</span></li>
      </ul>
    </div>
    <div class="ace-line" id="magicdomid986"><br>
    </div>
    <div class="ace-line" id="magicdomid4191"><span
        class="author-g-fp2xkcspz122zccid7qs b"><b>Action items</b></span></div>
    <ol>
      <li><span class="author-g-fp2xkcspz122zccid7qs">Get a list of
          people watching each TB component, and make sure we've got
          somebody with eyes on each one.</span></li>
      <li><span class="author-g-fp2xkcspz122zccid7qs">We're going to
          tackle the Papercuts list (10 bugs) to prove that we can deal
          with a list.</span></li>
      <li><span class="author-g-fp2xkcspz122zccid7qs">rkent(?) is going
          to send out a weekly status report on the Papercuts list, and
          what needs fixing, on what got fixed.</span></li>
      <li><span class="author-g-fp2xkcspz122zccid7qs">When the list is
          empty, or close to being empty, QA is going to develop a
          process for creating the list (which includes the facility for
          devs to indicate that a bug is too difficult).</span></li>
      <li><span class="author-g-fp2xkcspz122zccid7qs">We need to come up
          with some critiera for choosing bugs like the ones on
          Papercuts. Then QA / Support will build the list.</span></li>
    </ol>
    <br>
  </body>
</html>