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

Re: [XP] testing web applications

Expand Messages
  • Phlip
    ... One minor problem with the /WebXP/ book is it makes only a few technical recommendations, then neglects to discuss their challenges. The authors describe
    Message 1 of 16 , Oct 8, 2004
    • 0 Attachment
      Florian Weber wrote:

      > i didnt really mean a performance overhead. sorry. i
      > meant that
      > it will be a development time overhead. why should i
      > serialize
      > all the stuff to xml and write all the xslt, if i
      > can just use straight
      > objects and a template system in only 1/10 of the
      > time?

      One minor problem with the /WebXP/ book is it makes
      only a few technical recommendations, then neglects to
      discuss their challenges.

      The authors describe their own successes rescuing
      their Web site development careers from code-and-fix,
      but they do it by essentially setting each Web site up
      the same. "Iteration 1, you build the server", for
      example.

      Breaking the XML->XSLT recommendation up into
      components, we get...

      - all data is in only one format
      - all HTML comes from only one kind of template
      - the format and template are both _standard_
      - the format and template deal natively with &
      escapes
      - user's Web browsers might offload the final -> part

      The authors have bonded with XML->XSLT at the same
      time as they learned something else about Agile GUI
      development:

      Agile GUIs are Continuously Usable.

      At all times, each integration can produce a complete,
      round-trip application, with a GUI control for nearly
      every feature and a feature for every GUI control.

      /WebXP/ emphasizes that even loosey-goosey Web
      development must show this effect. Because the first
      XSLT effort is the site map, and because more XSLT can
      eat the site map to produce blanks for the site pages,
      at all times you can navigate the site:

      --> without broken links <--

      That sounds trivial. So what if pages without real
      content have no broken links? A broken link is not a
      bug you must hunt for days. However, enforcing that
      rule is _crucial_ for Agility.

      Agile projects have zero-tolerance for bugs of any
      severity.

      I personally go further. I call browser forgiveness,
      such as accepting <br> for <br/>, a bug. That's why I
      write tests that are more hyperactive than my target
      browsers. So I say XML->XHTML.

      =====
      Phlip
      http://industrialxp.org/community/bin/view/Main/TestFirstUserInterfaces



      __________________________________
      Do you Yahoo!?
      Y! Messenger - Communicate in real time. Download now.
      http://messenger.yahoo.com
    Your message has been successfully submitted and would be delivered to recipients shortly.